Plant Genetics

TIGHT BINDING BOOK

ro

164199 >m

= CD

OUP 2273 19-1 1-7-10,000 Copio. i,**, 1

OSMANIA UNIVERSITY LIBRARY
Call No ^ J* r Accesion No.

Autho>

Thii book should be returned on or before the date last marked belom,

PLANT GENETICS

BY

JOHN M. COULTER

Head of the Department of Botany
in the Untvemty of Chuaga

AND

MERLE C. COULTER

Instructor in Plant -Gtitetu*

in the University of Chicago

THE UNIVERSITY OF CHICAGO PRESS
CHICAGO, ILLINOIS

COPYRIGHT 1918 BY
THE UNIVERSITY OF CHICAGO

All Rights Reserved

Published July igi8

Composed and Printed By

The University of Chicago Press

Chicago, Illinois, U.S.A.

PREFACE

There is developing an increasing demand for courses
in genetics adapted to the needs of botanical students,
and teachers of botany have asked repeatedly for a
suitable text. A number of texts on heredity and evo-
lution have appeared, but these have contained far too
little genetics. On the other hand, there are several
very thorough texts on certain ^phases of genetics, but
these have been more in the nature of monographs than
texts. They have focused on too limited a field to give
the reader any general appreciation of genetics as a
whole and frequently have been too difficult for the
uninitiated to appreciate at all. WALTER'S Genetics
is an admirable elementary text, but is not exactly
adapted to the needs of the young botanist. All texts, in
fact, have emphasized animal genetics more strongly
than the student of botany needs, except certain books
on plant breeding, which, however, are general in nature
and give little attention to theoretical genetics. In
short, the subject has been changing so rapidly that no
one has ventured to write a general authoritative text. 1

The group that the present book is intended to serve
primarily comprises those who intend to make botany
their profession and who, although not as yet specialists,

1 After the manuscript of this text had gone to the printer, there
appeared a book by BABCOCK and CLAUSEN, entitled Genetics in relation
to agriculture. This is "a general authoritative text/' for it contains a
treatment of both plant and animal genetics, well suited to a thorough
year's course in genetics.

vi Preface

have had general training in the fundamentals of botany.
Such students, for example, are commonly found in the
last undergraduate year or first graduate year of their
work with no distinct purpose to become geneticists, but
wishing to be able to appreciate the important current
work in plant genetics.

Such a group of young botanists became evident at
the University of Chicago. In their attack upon current
botanical literature they frequently encountered papers
dealing with plant genetics, but through lack of prepa-
ration were unable to grasp their significance. This
type of literature seemed too important to be relegated
entirely to the specialist, and therefore the authors of the
present text organized a course of lectures to meet the
need. The purpose of the lectures was not to develop
professional geneticists, but merely to initiate students
of botany into the point of view of working geneticists,
so that they could appreciate an important phase of
botanical literature. With such a purpose there was
no attempt to give a complete presentation of modern
genetics, but rather to introduce the student to genetics
in the simplest way. As a consequence, for pedagogical
reasons, certain perplexing facts were omitted, while
others were slightly adjusted so as to convey the funda-
mental ideas without confusion. The result of the
course was so successful as to suggest the desirability of
putting the lectures into text form for the benefit of
teachers who wish to profit by this experience. As a
reference book it is entirely inadequate, much represen-
tative material having been omitted and only enough
bibliography given to put the student upon the trail.
As a textbook also it has two disadvantages: (i) it

Preface vii

is avowedly not exact in some of the details; (2) it is
adapted definitely to young botanists with fairly thor-
ough elementary training. The excuse for the inexact-
ness of certain details is the pedagogical necessity. The
preparation of a text for the students referred to is
explained by the fact that it represents a very important
group which has not been provided for. In brief, the
book is neither a technical presentation of genetics nor a
general text, but a course of general lectures adapted to
a special purpose.

In connection with the laboratory phase of this course
the authors wish to make grateful acknowledgment to
E. M. EAST, ALBERT BLAKESLEE, R. A. EMERSON, and
G. H. SHULL for very useful illustrative material.

J. M. C.

M. C. C.

CONTENTS

CHAPTER

I. INTRODUCTION

II. EARLIER THEORIES OF HEREDITY ..... 4

III. THE INHERITANCE OF ACQUIRED CHARACTERS . 16

IV. MENDEL'S LAW .......... 28

V. NEO-MENDELISM .......... 4

VI. NEO-MENDELISM (Continued) ...... 52

VII. NEO-MENDELISM (Continued) ...... 63

VIII. NEO-MENDELISM (Continued) ...... 76

IX. NEO-MENDELISM (Concluded) ...... 86

X. NON-MENDELIAN INHERITANCE ...... 97

XI. MODIFICATION OF UNIT CHARACTERS .... 104

XII. PARTHENOGENESIS AND VEGETATIVE APOGAMY . 115

XIII. INHERITANCE IN GAMETOPHYTES ..... 123

XIV. SELF-STERILITY ........... 133

XV. THE ENDOSPERM IN INHERITANCE ..... 144

XVI. HYBRID VIGOR ........... 158

XVII. SEX PL TERMINATION ......... 177

XVIII. THE BEARERS OF HEREDITARY CHARACTERS . .198

INDEX ................ 209

ix

CHAPTER I

INTRODUCTION

It should be realized that genetics is a natural out-
growth from the study of organic evolution. The con-
ception of evolution began as a speculation, but became
scientific in connection with the work of LAMARCK and
DARWTN. In such work the method used was that of
observation and inference. Facts were observed and
an explanation was devised that would relate them.
Resemblances and differences among species were
noted, and it was inferred that these indicated degrees
of relationship. It was assumed that closely related
species must have had a comparatively recent common
ancestry and that more distantly related species must
have had a more remote common ancestry. 'Through
comparisons of structure and of geographical distribution
systems of phylogcny have been inferred, and an outline
of the evolution of the plant and animal kingdoms has
been the result. All of these conclusions are based upon
comparison and inference.

This method reached its extreme application in the
work of DARWIN, whose observations included a range
of forms and an extent of time unequaled by any pre-
ceding student of evolution. It may be said that in
DARWIN'S work the method of comparison and infer-
ence reached the limit of its possibilities. The students
of evolution were chiefly concerned with explaining
the changes that resulted in phylogeny. In other words,

2 Plant Genetics

by comparison and inference they related certain forms
and then tried to explain why these forms had become
different. For example, LAMARCK'S explanation of the
changes was that they are the results of "use and dis-
use "; DARWIN'S explanation was that they are due to
"natural selection.^

It was in 1900 that a new method for the study of
evolution was announced with the appearance of The
mutationtheory by DE VRIES. The new method was
experimentation. Instead of comparing two species
and inferring that one of them had produced the
other, species were bred through successive generations,
under rigid control, and were observed to produce new
species. The old method inferred that a certain thing
occurred; the new method saw it actually occurring.

In developing this experimental method the facts
of inheritance began to accumulate. Presently the
facts accumulated sufficiently to be organized into
theories of inheritance, and the special field of genetics
was the result. Jn brief lt _the.rcfQre J gengtics is the
experimental study of inheritance. Aside from the
interest of genetics itself, its possible applications are
even more important. For example, it will probably
eventually explain organic evolution; and, most impor-
tant of all considerations, it will probably enable us to
control plant and animal breeding in a way that will
be of the greatest practical importance in agriculture.

It should be realized that, in so complex a subject as
heredity, increasing experimental work must greatly
increase the range of known facts and make explanations
increasingly difficult. The facts of inheritance obtained
from experimental work with a few simple forms

Introduction 3

may permit satisfactory conclusions; but when greater
numbers and more complex forms are included, con-
clusions become not only difficult but sometimes impos-
sible. It is necessary, however, to organize facts,
as they are obtained, into some consistent theory that
will relate them to one another and that may be the
basis of further investigation. Progress in any complex
subject is marked by a series of explanations, each in
turn proving inadequate as facts multiply, but each in
turn helping to further progress. It should be under-
stood, therefore, that in the subsequent chapters the
theories presented are not to be regarded as final, but
rather as suggested explanations of the known facts.
It is certain that new facts will continue to be discovered
and that explanations will have to be modified to fit
them; but the present explanations are necessary for co-
ordinating the facts we have. In other words, a sharp ;
distinction must be drawn between established facts
and proposed explanations. The former are permanent,
the latter are temporary.

CHAPTER II
EARLIER THEORIES OF HEREDITY

It is probable that men have thought of heredity from
the earliest times, but so far as is known there was
no formulation of any definite theory of heredity before
the time of j\RiSTQJtE. Since that time there are
records of numerous theories of inheritance that may
well be regarded as wild speculations. They were
evolved with little or no basis of fact and of course are
not to be regarded as scientific. Some of these early
theories are interesting but not profitable for our pur-
pose. They developed from superstition rather than
from observation.

It is evident also that inheritance in man first
attracted attention; later, animals were taken into
consideration. The study of inheritance in plants is of
comparatively recent development, due chiefly to the
fact that ,sX was not observed in plants until late in the
seventeenth -century, and heredity was thought of only
in connection with sex. Even after sex in plants was
announced by CAMERARIUS JJ^L 1694 tjiere was no general
belief in the~cHmi until much later. For this reason
plant inheritance became a subject of observation and
theory, and the history of plant genetics began long
after the discovery of sex in plants.

The earliest distinct theory of heredity which falls
within the period of modern biology was formulated
by DARWIN (i). This was published in 1868 in his

~~~~~ 4

Earlier Theories of Heredity 5

book entitled Animals and plants under domestication
and was called pangenesis. (In a certain sense DARWIN
apologized for pangenesis, realizing that knowledge had
not advanced far enough for the construction of any
adequate theory.) The theory of pangenesis, therefore,
is to be regarded more as a suggestion than as a formula-
tion of belief.

The theory may be summarized as involving twc
theses which should be kept distinct. The first thesis,
in DARWIN'S words, is as follows: "The individual
cells and organs of the whole organism are represented in
every germ cell and bud by definite material particles.
These multiply by division, and at cell division pass on
from the mother cells to the daughter cells." This is the
essential feature of pangenesis; that is, every germ cell
(egg or sperm) contains dividing particles .that in cell
division pass on to every cell produced. This statement
agrees in general with the facts recognized today under
a somewhat different terminology, it being only neces-
sary to call these particle^ chromosomes. This is re-
markable when it is remembered that the statement was
made fifty years ago.

The second thesis is as follows: "All the cells of the
body .throw off particles at various periods of their
development. These reach the germ cells and hand
over to them any characters of the organism which they
may lack// This is known as the transportation hy-
pothesis and was an attempt to account for certain facts
which seemed to indicate^ the ^ inheritance of .acquired
^characters. It is somewhat surprising to find DARWIN
explaining heredity on the basis of the inheritance of
acquired characters, for his theory of the origin of

6 Plant Genetics

species by natural selection does not call for the inherit-
ance of acquired characters.

The theory of pangenesis proper is quite in keeping
with the present point of view; but the weakness oi
the transportation hypothesis was so evident that it
was soon set aside by biologists. In discarding the
transportation hypothesis, however, biologists in general
rejected also the doctrine of pangenesis, a common result
when a truth is found in combination with the obviously
false.

The theory of heredity which was chiefly responsible
for replacing pangenesis was proposed by WEISMANN^S),
whose publications appeared during the decade 1880^
1890. WEISMANN developed two companion theories:
one called gerjninal ^c^muity^ which has to do with

heredity; and the other called germinal selection,
which is a very imaginative explanation of individual
variation. Our concern is chiefly with the theory of
germinal continuity.

According to DARWIN'S transportation hypothesis,
any change arising in the organism at any time during
its life would become- represented by gemmules in the
reproductive cells. In this way it would be possible for
acquired characters to be inherited. WEISMANN could
discover no mechanism in plants or animals which could
justify such a conception.

It had been known for some time that germ cells and
body cells in animals remain separate during their later
development, but WEISMANN seems to haye been the
first to point out the significance of this fact. He makes
Ehe following statement: "The difficulty or the impos-
sibility of rendering the transmission of acquired char-

Earlier Theories of Heredity 7

acters intelligible by an appeal to any known force has
often been felt, but no one has hitherto cast doubt upon
the very existence of such a form of heredity." The
general belief of the time that the inheritance of acquired
characters was a fact came from two sources: (i) many
examples of the inheritance of acquired characters
were being reported; and (2) inheritance of acquired
characters was thought necessary to explain evolution.
WEISMANN therefore faced two problems: (i) to explain
away the reported cases of the inheritance of acquired
characters; and (2) to provide a theory which would
make evolution possible without the inheritance of
acquired characters.

He took up individually the many reported cases
of such inheritance and discredited them one after
another, showing convincingly how they could be
explained in some other way. He also cited from his
own experience numerous cases in which the inheritance
of acquired characters was distinctly absent. As a
result of his investigation of these cases, he developed
his theory of germinal continuity, commonly spoken of
as continuity of the germ plasm, a theory which is in good
standing today. In the attempt to provide a theory
which would make evolution possible without the
inheritance of acquired characters WEISMANN dis-
tinctly failed. To explain variation, which is the
basis of evolution, he proposed the theory of germinal
selection, which is even more imaginative than DAR-
WIN'S transportation hypothesis. As a consequence
WEISMAN^S experience was much like that of DARWIN.
His theory of germinal continuity has fairly well stood
the test of later investigation and is still current among

& Plant Genetics

biologists, while his theory of germinal selection has
been practically discarded. A brief explanation of these
two theses is as follows.

GERMINAL CONTINUITY. A statement of this theory
in the words of the author is as follows :

I believe that heredity depends upon the fact that a small
portion of the effective substance of the germ plasm remains
unchanged during the development of the egg into an organism,
and that this part of the germ plasm serves as a foundation
from which the germ cells of the new organism are produced.
There is a continuity of germ plasm from one generation to
another .... hence it follows that the transmission of acquired
characters is an impossibility, for if the germ plasm is not formed
anew in each individual, but is derived from that which preceded
it, its structure, and above all its molecular constitution, cannot
depend upon the individual in which it happens to occur; but
such an individual only forms as it were the nutritive soil, at
the expense of which the germ plasm grows, while the latter
possessed its characteristic structure from the beginning, namely,
before the commencement of growth, but the tendencies of
heredity, of which the germ plasm is the bearer, depend upon its
molecular structure, and hence only those characters can be
transmitted through successive generations which have previ-
ously been inherited, namely, those characters which were poten-
tially contained in the structure of the germ plasm. It also follows
that those other characters which have been acquired by the
influence of special external conditions during the lifetime of the
parent cannot be transmitted at all.

This is the theory of germinal continuity and it is in
general agreement with the results of biological work
today (see fig. i).

GERMINAL SELECTION. The purpose of the con-
ception of germinal selection was to construct a theory
of variation, and therefore of evolution, which does not
involve the inheritance of acquired characters. When

FIG. i. Diagram illustrating WEISM ANN'S theory of germinal
continuity. Three generations are represented, with cells of germ
plasm shaded, and those of body plasm unshaded; germ plasm con-
tinuous from generation to generation, carried over from parent to
offspring by zygote (Z) ; impossible for body plasm to perpetuate itself
into a second generation.

io Plant Genetics

the theory was being constructed the theory of natural
selection was widely accepted. It should be understood
that WEISMANN did not criticize natural selection, but
he did not believe that the individual variations upon
which natural selection was based were to be explained as
the inheritance of acquired characters. The theory of
germinal selection, therefore, was intended to explain the
origin of individual variations in some other way. The
explanation of the theory begins as follows :

I believe that it is possible to suggest that the origin of heredi-
tary individual characters takes place in a manner quite different

from any which has as yet been brought forward In the

first place, it may be argued that external influences may not
only act on the mature individual, or during its development,
but that they may also act at a still earlier period upon the germ
cell from which it arises. It may be imagined that such influ-
ence of different kinds might produce corresponding minute
alterations in the molecular structure of the germ plasm; and as
the latter is, according to our supposition, transmitted from one
generation to another, it follows that such changes would be
hereditary Without altogether denying that such influ-
ences may directly modify the germ cells, I nevertheless believe
that they have no share in the production of hereditary individual
characters Hereditary individual differences must there-
fore be derived from some other source. I believe that such a
source is to be looked for in the form of reproduction by which the
great majority of existing organisms are propagated, sexual

reproduction The object of sexual reproduction is to

create those individual differences which form the material out of
which natural selection produces new species.

WEISMANN forestalls a criticism which was sure to
come by saying that the opinion has already been
expressed that deviations from the specific type are
rapidly destroyed by the operation of sexual reproduc-

Earlier Theories of Heredity n

tion. Such an opinion may be true with regard to
specific characters, because deviations from the specific
type occur in such rare cases that they cannot hold their
ground against the large number of normal individuals,
but the case is different, as he claims, with those minute
differences which are characteristic of individuals,
because every individual possesses them, although
of different kind and degree. He states that cross-
breeding between all the individuals of a species is
impossible, hence the obliterating of individual differ-
ences is impossible.

According to WEISMANN, therefore, individual varia-
tions, and therefore the possibility of natural selection,
originate in the sex process. Sex reproduction originates
no absolutely new characters, but merely makes possible
innumerable combinations of characters. It is evident
that the pairing individuals must differ to some extent,
if the result is to be a new combination of characters.

This raises' the question of the origin of differences
in the first sexual individuals. Sexual reproduction
was derived from asexual reproduction, and, according
to the theory, must have begun with variable material.
How did these individual differences arise ? WEISMANN
offers the following explanation. He had stated that
environment has no modifying influence upon the germ
plasm in general, but says that "in the lowest one-
celled organisms the case is entirely different/ 7 In
this case, as he puts it, "parent and offspring are in a
certain sense one and the same thing/ 7 and germ plasm
and body plasm have not been differentiated. In this
case, therefore, variations induced by external conditions
are hereditary because the germ plasm is the body plasm,

12 Plant Genetics

In this way individual differences arose in one-celled
organisms; and when the sexual method appeared it
found varying individuals ready to work upon. The
mission of sex, therefore, according to WEISMANN, was
to lay hold of these variations and multiply their com-
binations, thus providing a wider range of choice for
natural selection. In this way he explains the origin of
individual variation without the inheritance of acquired
characters.

WEISMANN also devised a mechanism by which
sexual reproduction results in multiplying combinations
of individual differences. This mechanism is the basis
of his theory of germinal selection and is as follows.

In WEISM ANN'S time chromosomes as such had not
been recognized, but they were represented in his termi-
nology by the term idants. He conceived of each idant
as composed of one or more units called ids. An id
stood for a complete individual, each id being able to
control the complete development of an individual. A
child received both paternal and maternal ids in equal
numbers; that is, father and mother were each repre-
sented by a 50 per cent influence on the offspring;
each grandparent would be represented by 25 per cent
of the ids in an individual; and so on, indefinitely. As a
consequence, all existing individuals must now contain
as many different kinds of ancestral ids as they are
capable of containing. Sooner or later, owing to the
accumulation of ids, "sex reproduction can proceed
only by reduction in the number of ancestral ids, a
reduction which is represented in every generation."

This is WEISMANN'S explanation of the grosser
mechanism of heredity, but it was only preliminary to

Earlier Theories of Heredity 13

the theory of germinal selection. In the first place
he imagined .that each id is composed of determinants,
which are ultra-microscopic, each of them representing,
not the whole body structure as does an id, but all the
characters that belong to a particular kind of cell.
Each determinant in turn was conceived of as composed
of still smaller units called biophores, each of which
represented a single character of a cell. Biophores
have the capacity to assimilate food, to grow, and to
reproduce themselves in the germ plasm. This imagi-
nary structure of the cell is spoken of as WEISMANN'S
cell architecture, in which biophores are built into deter-
minants, determinants into ids, ids into idants. In
fact, this is a remarkable theoretical analysis of the
structure of a chromosome.

Selection is conceived of as occurring in the following
way. After fertilization the biophores are too numerous
for all of them to obtain expression, and the result is
a struggle for existence, a sort of ultra-microscopical
warfare. As a consequence some biophores perish and
others survive and are perpetuated. This struggle of the
biophores for nutrition, and the almost infinite combina-
tions that might enter into the structure of the resulting
determinants, would open up a wide field for variation.
This is the theory of germinal selection, a struggle and
a survival of fortunate biophores, and it may well
be regarded as fantastic. It has been discarded, not
only because it is without foundation in fact, but
chiefly because it is beyond the reach of experimental
testing.

What is called Weismannism had a very wide influ-
ence for a number of years, and the continuity of the

14 Plant Genetics

germ plasm, with its result on the doctrine of the inherit-
ance of acquired characters, is still current. The first
result of Weismannism was to overshadow DARWIN'S
theory of pangenesis, which had unjustly shared the
deserved fate of its companion theory, the transporta-
tion hypothesis. DE VRIES (2) was the first to recog- 1
nize this injustice, and in 1889 published his Intracellular
pangenesis. Without attempting to defend the trans-
portation hypothesis, DE VRIES showed the real value
of the doctrine of pangenesis. He claimed that his
theory of intracellular pangenesis is little more than a
restatement of the fundamental ideas of DARWIN'S
pangenesis, but in fact it contains enough new material
ty justify a definition.

The material particles which DE VRIES conceived
ot as the carriers of heredity he called pangenSj the
equivalent of DARWIN'S gemmules. He claims that in
the nucleus every kind of pangen of the individual is
represented, while the cytoplasm in each cell contains
only those pangens that become active in the cell. He
concludes that, with the exception of the pangens con-
cerned in nuclear activities, such as nuclear division,
all pangens have to leave the nucleus in order to become
active. But most of the pangens of almost every kind
are represented in the nucleus, where they multiply,
partly for the purpose of nuclear division, partly to pass
into the cytoplasm to engage in other activities. These
pangens not only pass out into the adjacent cytoplasm,
but are also carried by the protoplasmic currents into*
the various organs of the protoplast, where they multiply
and become active. All protoplasm consists of such
pangens derived at different times from the nucleus,

Earlier Theories of Heredity 15

together with their descendants. Hence the theory
is called intracellular pangenesis, as contrasted with
DARWIN'S body pangenesis, that is, distribution of the
heredity particles in the cell as contrasted with distri-
bution through the body as a whole.

DE VRIES formulated this theory before he had
conducted the breeding experiments that led to his
theory of mutation, and afterward he applied the theory
to his experimental results in breeding.

1. DARWIN, CHARLES, The variation of animals and plants under
domestication. London. 1890.

2. DE VRIES, HUGO, Intracellular pangenesis. Chicago. 1910.

3. POULTON, SHONLAND, SHIPLEY, WEISMANN, on heredity.
Clarendon Press. 1891.

CHAPTER III
THE INHERITANCE OF ACQUIRED CHARACTERS

The inheritance of acquired characters has been
mentioned in its relation to WEISMANN'S theory of
germinal continuity, but it deserves a somewhat fuller
consideration. The idea was first developed by LAMARCK
in connection with his theory of evolution, the so-called
theory of appetency, or the effect of use and disuse.
FRANCIS GALTON, in 1875, was one of the first to express
skepticism in regard to such inheritance, but it was
WEISMANN who was most influential in combating the
idea. After WEISMANN'S presentation of the situation,
biologists were divided into two camps in reference
to the inheritance of acquired characters: (i) neo-
Lamarckians, who affirm belief in the inheritance of
acquired characters, and. (2) neo-Darwinians, who deny
it. Geneticists and embryologists, however, whose work
brings them most in contact with the problem, seem to
be fairly well agreed that acquired characters are not
inherited.

Much of the lack of agreement in this controversy
is due to the definition of an acquired character. It
should be kept in mind that actual characters are not
inherited, but only the determiners, which regulate the
way in which the organism reacts to its environment.
For example, when it is said that a child inherits its
father's nose, the statement is not meant to be literally
true; it is meant that just as there was something in

16

Inheritance of Acquired Characters 17

the body of the father that was responsible for the
development of a particular type of nose, so there was a
similar something in the child's body that produced a
similar result. It is merely a matter of convenience to
speak of the inheritance of characters.

WEISMANN defined an acquired character as "any
somatic modification that does not have its origin in
the germ plasm." This definition is not always easy
to apply. Examples of acquired characters in the
Weismann sense are mutilations, effects of environment,
results of function (as in the use or disuse of certain
organs), and many diseases that affect the bodily mech-
anism. WEISMANN gave three reasons for rejecting the
belief in the inheritance of such characters: (i) there is
no known mechanism by which somatic characters may
be transferred to the germ plasm; (2) the evidence that
such a transfer does occur is inconclusive and unsatis-
factory; and (3) the theory of. the continuity of the
germ plasm is sufficient to account for the facts of
heredity. This last reason has been discussed, but the
other two should be considered.

When WEISMANN says that there is no known mech-
anism by which somatic characters can be transferred
to the germ plasm, to him it is equivalent to saying that
it is hard to see how the water that has gone over the
dam can return and affect the flow of the water upstream.
He assumes, of course, that the germ plasm is isolated
from the somatoplasm very early in the development
of the fertilized egg into an individual, and that when
it is isolated it takes no active part in the history of the
body (see fig. i). The somatoplasm is thus merely a
carrier of the germ plasm and is unable to affect the

1 8 Plant Genetics

character of it any more than a rubber hot-water bag,
although capable of assuming a variety of shapes, can
affect the character of the water it contains (WALTER 6).

Opponents of WEISMANN object to any such view
of the complete isolation of germ plasm. Studies of
cell lineage in animals have shown that germ cells are
set apart very early in the development of the individual,
and that they certainly are not derived from distinctly
somatic cells. No such statement, however, can be
made in reference to plants. Germ cells in plants are
derived from epidermal or hypodermal cells, which
previously were distinctly somatic; and furthermore
germ cells have been induced by experiment to form from
almost any living tissue which is as distinctly somatic
as any plant tissue can be.

Another objection to WEISMANN'S view is that every
organism is a physiological as well as a morphological
unity, and that cells completely insulated in such a
unity would be impossible. Cytologists also have come
to believe that there are protoplasmic connections
between adjacent cells in practically all plant tissues, and
in general physiology tends to confirm this. All of this
means a growing belief that the somatoplasm can affect
the germ plasm.

The reply of the Weismannians is that even though
somatoplasm can affect germ plasm in this general phys-
iological way, this is a very different thing from the in-
inheritance of some definite acquired character. To be
inherited such a character would have to be exactly re-
developed in the germ plasm, and the influence referred
to cannot be so specific as that. This of course is a
theoretical answer and the question can only be decided

Inheritance of Acquired Characters 19

by experimental work. A theoretical rejoinder to this
answer may be suggested. It is like the voice in a tele-
phone transmitter, which starts vibrations that make
the receiver repeat the voice.

WEISMANN'S disposition of the claimed cases of the
inheritance of acquired characters should be considered.
The supposed evidence for such characters falls chiefly
into four categories.

i. MUTILATIONS. Most of the evidence under this
head is in relation to animals. It must be the common
experience that mutilations are not inherited in man and
the domesticated animals. A . few quotations from
WALTER (6) suggest the situation:

It is fortunate that the sons of warriors do not inherit their
fathers' honorable scars of battle, else we would now be a race of

cripples The feet of Chinese women of certain classes

have for centuries been mutilated into deformity by bandages
without the mutilation in any way becoming an inherited char-
acter The progressive degeneration or crippling of the

little toe in man has been explained as the inheritance of the
cramping effect of shoes upon generations of shoe-wearers; but
WIEDERSHEIM has pointed out that Egyptian mummies show the
same crippling of the little toe, and no ancient Egyptian could be
accused of wearing shoes, or of having had shoe-wearing ancestors.

Sheep and horses with docked tails, as well as dogs
with cropped ears, never produce young having the
parental deformity. WEISMANN'S experiments with
mice, later verified by other investigators, give addi-
tional evidence that mutilations are not inherited. He
bred mice whose tails had been cut off short at birth, and
continued this performance through twenty-two genera-
tions, with absolutely no effect on tail length.

2o Plant Genetics

Mutilations in plants have received no serious
consideration; in fact, no one seems to have con-
sidered the possibility of such inheritance. Cuttings
for propagation, for example, are usually trimmed to
prevent excessive transpiration, but no one ever expects
to find this mutilation perpetuated, even in the plant
developed from the cutting, much less in the next
generation developed from seed. In fact, since we
have begun to learn of the remarkable powers of regenera-
tion possessed by plants and animals^ we would not
expect the inheritance of mutilations.

2. EFFECTS OF ENVIRONMENT. This has long been
a topic in botany. Trees deformed by prevailing winds,
like the willows that line the canals in Belgium and
Holland, or storm-crippled trees along exposed seacoasts,
are not known to produce progeny showing these char-
acters when the adverse environmental conditions are
removed. ZEDERBAUER, on the other hand, found that
Capsella, which in the course of many years had gradu-
ally crept along the roadside up into an alpine habitat
and there acquired alpine characters, retained these
characters when transplanted to the lowlands. This has
been accepted as an authentic instance of the inheritance
of acquired characters; but it is possible that this con-
quest of an alpine habitat by Capsella can better be
explained by the gradual natural selection of just those
germinal variations that best fitted individuals to cope
with alpine conditions. This would result in the gradual
establishment of a strain of germ plasm that would
produce body structures fitted to alpine conditions.
In other words, this is just the way in which natural
selection would develop a new elementary species

Inheritance of Acquired Characters 21

from the original type. If such a type were established,
of course its germ plasm would produce alpine plants,
even in lowland conditions. They might not survive
long- and natural selection might eliminate them, but
their structure would be due, not to the inheritance of
somatic structures, but to the inheritance of an alpine
germ plasm.

If corn is planted in poor soil weak individuals
result. Seed from these weak individuals, when planted
in good soil, will develop again somewhat weakened
individuals, and this suggests the inheritance of acquired
characters. In fact, however, it is merely the direct
effect of environment continuing through the second
generation. The weak individuals in the poor soil
develop small seeds with low nutritive capacity, and
plants developed from abnormally small seeds are always
weak, whether the individual that produced the seed grew
in poor soil or not.

In 1909 M AYR (4) wrote a notable work on silvicul-
ture in which he claimed that only species characters are
inherited in trees, and that the effects of climate are not
inherited, and therefore that the source of the seed
makes no difference. In other words, seeds of Scotch
pine would always produce Scotch pine progeny,
no matter at what latitude or altitude the ancestors
had been growing. According to MAYR, therefore,
there is no inheritance of acquired characters in
trees.

In 1912; however, the United States Committee on
Breeding Nut and Forest Trees (5) came to the con-
clusion that the source of seed is of great importance.
This conclusion was based chiefly upon the testimony

22 Plant Genetics

of numerous nurserymen, the only significant experi-
mental work being that of ENGLER of Zurich. ENGLER
(3) found that in the seedlings in his nursery growth
in height distinctly decreased as the altitude or latitude
from which the seed came increased. He also found
that seeds from pines which had been crippled by
growing in poor soil conditions gave rise to crippled
plants when grown in good soil. In many cases trees
of the third generation still showed the habit " acquired"
by their grandparents in different habitats.

These are striking results, but they may be explained
in either of the two ways mentioned. ENGLER might
have been dealing with slightly different strains of
trees, differing in germinal constitution; or it may have
been another case of the " false inheritance of acquired
characters " that was explained in connection with corn.
Seeds from higher latitudes and altitudes might well have
been smaller, so that we should have expected smaller
progeny, even when grown in the lowlands.

A completely satisfactory investigation of the inherit-
ance of acquired characters in plants still remains to be
made. Seed from a single parent from a pure strain
should be planted in diverse conditions and the various
responses noted. Seed of the second and later genera-
tions of each lot should then be planted back in the
original conditions. If there should appear in this
generation even slight deviations from the original
type it would be significant, provided the deviations are
similar to those shown by the immediate parents in
each case and characteristic of the conditions under
which the parents grew. The peculiarities of progeny
due to light-weight seed should be tested by controls

Inheritance of Acquired Characters 23

and eliminated from consideration. Certain parts of
this investigation have been carried on frequently and
satisfactorily and the whole investigation has been tried,
but under poorly controlled conditions. It remains,
therefore, to conduct the entire investigation under
proper conditions before one can reach any reliable
conclusion in reference to the inheritance of acquired
characters by plants.

Another possible illustration of the inheritance of
responses to environmental conditions may be obtained
from the work of BOLLEY (i) on flax. The resistance of
a plant to a given disease is regarded as a character
peculiar to certain strains and transmitted as a very
definite factor in Mendelian inheritance. BOLLEY claims
that he can get a resistant strain of flax from almost any
known variety. According to him the resisting ability
increases from generation to generation, if the crop
is constantly subjected to disease attack. He took a
pure-pedigreed strain of flax which had come originally
from a single non-resisting seed. This was planted in
slightly "sick" soil, that is, soil infected with the wilt-
producing organism. Most of the individuals died, but
"a few scrubs 7 ' survived. He then planted seeds from
these in slightly " sicker" soil than before, and thus,
by gradually working his crop into sicker and sicker soil
in the later generations, he finally obtained a fully
resistant strain from the pure non-resistant strain with
which he started. Such a strain he says will not lose its
resistance if planted progressively in more infected soils.

BOLLEY also found that if he added to the soil manure,
alkalies, etc., known to increase the disease by stimulat-
ing the fungus, he obtained a very few poorly developed

24 Plant Genetics

resisting plants, whose seeds, however, the next year
developed a race with enormous resistance. He gives
the following theoretical explanation of his results:

Either (i) the so-called unit character of resistance was present
in undeveloped form and becomes stronger from year to year
under conditions of disease; or (2) there never was any character
present which is entitled to be called a unit character, but it began
to develop the first year the parent plant came in contact with the
disease, and the protoplasmic nature of the ancestors of the plants
which we now have has been such that they accumulated more
and more the resisting power from year to year, just as they had
opportunity to develop resistance against a constantly acting
factor of disease, which, when too powerful, acts as an eliminating
factor.

BOLLEY inclines to the second alternative. This
general conception seems to explain why homegrown
seed is regularly more resistant than seed from the
same variety which has had a vacation away from
home for several years. It has kept in training like a
football player. BOLLEY says that if these conclusions
are correct there are probably no unit characters which
are not fluctuating and there are no fluctuating characters
which may not readily be fixed.

These results are striking enough, but their signifi-
cance depends entirely upon the purity of the strains
which were used originally and also upon the preserva-
tion of purity during the experiment. BOLLEY'S phrase
" elimination factor/' which he uses repeatedly, suggests
selection from an impufe strain. If his conception is
true it could be demonstrated by developing a large
majority of resistant individuals among the non-resistant
plants which were first subjected to disease attack,
rather than merely "a few scrubs." In other words, the

Inheritance of Acquired Characters 25

whole result resembles the selection of a few resistant
individuals from an impure strain.

3. EFFECT OF USE AND DISUSE.- --WEISMANN dis-
credited this belief, which was the foundation of LA-
MARCK'S theory of evolution. He was successful in
practically all the cases that had been presented for
animals. In plants, of course, it would be hard to find
anything exactly analogous to the use and disuse of parts
in animals. One fact, however, may be mentioned which
is a common experience of botanists. Functionless
organs gradually become aborted, becoming mere ves-
tiges, or even suppressed. A study of the organogeny
of a flower shows that when a floral member is belated
in its development it is destined sooner or later not to
appear at all. The Weismannian explanation of this
situation would probably be as follows. A given species
has given a nutritive capacity; the less it draws upon
its nutritive capital for the development of one organ
the more it can afford to expend upon the development
of other organs. When an organ becomes functionless
it no longer has any survival value; survival is then
dependent upon the relative development of the other
organs. Certain variations develop the functionless
organ less than usual and therefore develop the other
organs more than usual, and under the new conditions
these variations will survive and the others be eliminated.

This is a description of natural selection by which
functionless organs become more and more aborted and
vigorous organs survive, but such an explanation is
rather imaginative.

4. DISEASE TRANSMISSION. Roughly speaking this
may be grouped as (i) infection by bacteria or fungi, and

26 Plant Genetics

(2) some inherent organic weakness. Since the latter
condition is chiefly serious only in inviting attacks
by bacteria or fungi, we are concerned chiefly with
diseases caused by these pathological forms. Realizing
this, true inhentan^^disease seems to be an impossi-
bility, for if the p^asite^enters the germ cell it is prac-
tically sure to destroy it, and there will be no progeny.
Tn many cases, however, progeny are born diseased,
but this is due to reinfection of the young embryo from
the body of the mother. This is not an inheritance but
i reinfection. In smuts, for example, there is much
.reinfection by means of the transmission of spores upon
seeds. This can in no sense be spoken of as inheritance.

In one respect, however, one may speak of disease
inheritance. Breeding experiments have shown that
predisposition to disease and disease resistance, com-
monly called susceptibility and immunity, are inherited.
This means that it is of germinal origin, so that it does
not involve the inheritance of an acquired character.
On the other hand, as was stated, BOLLEY maintains
that disease resistance is built up under the influence of
disease attack, increasing gradually through the genera-
tions. This would make it an acquired character and one
that is inherited.

These are a few of the examples of so-called acquired
characters and the claims for and against them. Investi-
gation is not yet in a position to come to any definite
conclusion in reference to them. The bulk of available
evidence, however, seems to be against the Jnheritance
of acquired characters; but there are a number of
biological facts that seem difficult to explain in any other
way, In animals the mechanism may seem to make the

Inheritance of Acquired Characters 27

inheritance of acquired characters impossible; but the
situation in plants is distinctly different.

In closing, we quote the opinion of EAST (2) :

My confession of faith is, the environment has been an
immense factor in organic evolution, but its effects are shown
either so infrequently or after the elapse of so great a time, that
for the practical purposes of plant breeding we can neglect it as we
would neglect an infinitesimal in a calculation.

1. BOLLEY, H. L., The importance of maintaining a constant
elimination factor in plant breeding. Ann. Rep. Amer.
Breeders Assoc. 8:508-514. 1912.

2. EAST, E. M., The bearing of some general biological facts on
bud-variation. Amer. Nat. 51:129-143. 1917.

3. ENGLER, ARNOLD, Influence of source of seed. Jour. Heredity
5:185-186. 1914.

4. MAYR, H., Waldbau auf naturgesetzJicher Grundlage. Berlin.
1909.

5. SUD WORTH, GEO. B., Report of committee on breeding nut and
forest trees. Ann. Rep. Amer. Breeders Assoc. 8:515-522.
1912.

6. WALTER, HERBERT E., Genetics. New York. 1913.

CHAPTER IV
MENDEL'S LAW

Mendel's law is the basis of all work in genetics and
should be understood from its original statement to its
somewhat complex development. Before passing to the
more complex features it is well to recall the points of
the original thesis.

In 1865 GREGOR MENDEL (2) published in the pro-
ceedings of a local scientific society the result of eight
years of breeding experiments. The publication was so
obscure that scientific men in general did not see it, and,
in addition to this, Darwinism was at that time absorbing
the attention of biologists. For these two reasons
MENDEL'S work remained unnoticed, and of course
unappreciated, until it was discovered in 1900 and
became the great classic of genetics. Its influence,
therefore, dates from 1900 rather than from the year of
its publication.

The substance of MENDEL'S experiments is as fol-
lows. Wishing to discover the contributions of each
parent to the make-up of their progeny he chose for his
work the simple garden pea, which would breed rapidly,
and exhibited well marked varieties. To magnify his
results he secured hybrids by crossing distinctly different
types of peas, and to avoid confusion he considered
only one character in each experiment. For example,
he crossed peas which contrasted in character of height,
of flower color, and of seeds. In every case he obtained

28

MendeVs Law

29

the same result, so that a single example will suffice.
Furthermore, he discovered that it made no difference
whether the staminate parent was a dwarf and the pistil-
late tall, or vice versa, and so for all the characters used.
In other words, what are called reciprocal crosses gave

the same results.

The progeny of a tall
parent and a dwarf parent
were all tall. This genera-
tion is known as the first
hybrid generation or the
F r generation. When this
generation was inbred the
progeny was made up of
tall and dwarf individuals
in a ratio of 3:1. This
generation is known as the
second hybrid generation,
or the F 2 generation. The
dwarf forms of the F 2
generation subsequently
bred true, producing only

T X D

T
1
T

T

I

T T T D T T 1 D

D
I
D

FIG. 2. Diagram illustrating
visible results of MENDEL'S experi-
ments. Cross between tall parent
(T) and dwarf parent (D) gives
hybrid progeny, which are all tall;
hybrid progeny inbred gives 3 : i
ratio in second hybrid generation;
inbreeding each of these four indi-
viduals separately gives for third
hybrid generation results indicated

in bottom line.

dwarfs. Of the tall forms

one-third bred true and two-thirds split up in just such
a 3 : i ratio as did their immediate parents of the F x
generation. This may be expressed diagrammatically
(fig. 2).

MENDEL'S explanation of this behavior involved
three theses which at that time were new to biology.
These theses must be kept distinct from one another.

i. INDEPENDENT UNIT CHARACTERS. This means
that an organism, although representing a morphological

30 Plant Genetics

and physiological unity, from the standpoint of heredity
is a complex of a large number of independent heritable
units. Thus if one pea plant is tall and another one is
dwarf the behavior of the hybrid produced from them
with reference to this character will be the same, no
matter what other characters the parent plants may
have had. In other words, the characters are independ-
ent units, unaffected by other characters or units. The
character of tallness from a tall plant with wrinkled
seeds or purple flowers will act just the same as from a
tall plant with smooth seeds or white flowers. Tallness
is a unit and its behavior in inheritance is independent
of all other units.

2. DOMINANCE.- In the germ plasm there are certain
determiners of unit characters which dominate during
the development of the body, causing these characters
to dominate over others and thus become visible. The
characters dominated over and thus not allowed to
express themselves are called recessive characters. These
recessive characters are present in the germ plasm,
but cannot express themselves and become 'visible as
long as the dominant characters are present. When a
dominant character is absent, however, its recessive alter-
nate is free to express itself and become visible.

For example, in the case of tall and dwarf peas, tall-
ness is a dominant character and dwarfness is its alter-
native recessive. When a dwarf appears, therefore,
there is present no dominant tallness to suppress it. In
the Fx generation all the individuals were tall because,
although they had all received the recessive character
of dwarfness from one of the parents, they had received
the dominant character of tallness from the other parent,

Mendel's Law 31

and so dwarf ness did not appear in any of them. Such
pairs of alternative characters are now commonly called

I ' L '" ~""" ir "'' lff """ r " -^"^''nr.i.mui.tu ^^^^^ ,.*.,. **.Avi^, WwiMw|M , .

allelomorphs. Thus tallness and dwarfness are allelo-
morphs in the pea, one dominant over the other, which
is therefore recessive.

3. PURITY OF GAMETES. A gamete can contain
only one of two alternative characters. For example,
it may contain the character for tallness or for dwarfness,
but not both. In other words, allelomorphs cannot be
represented in the same gamete. If the gamete having
the character for tallness unites with one having the
character for dwarfness, the resulting zygote will con-
tain both, but will produce a tall individual because
tallness is dominant over dwarfness. When this tall
hybrid produces gametes, however, one-half of them
will contain the character for tallness and one-half of
them the character for dwarfness. Thus the alternative
characters are " segregated " in gamete formation and
no gamete will have both characters.

These three theses, independent unit characters,
dominance, and purity of gametes (better called segre-
gation) , make up the theoretical explanation of Mendel's
law. Independent unit characters was of course a
necessary conception. It was original with MENDEL,
and has also been original with other investigators, but
this conception does not represent the essential fea-
ture of Mendel's law. The idea of dominance had been
somewhat vaguely proposed before MENDEL'S time.
In the old literature on animal breeding one meets
theories of prepotency p , which were proposed again and
again before the discovery of MENDEL'S work in 1900.
In any event MENDEL was the first to formulate definitely

32 Plant Genetics

the theory of dominance among unit characters. It
should be realized also that dominance is not an
essential feature of MENDEL'S theory. Many cases
are known in which dominance fails, but in other
regards the Mendelian inheritance is strictly followed.

The essential feature of MENDEL'S theory is his
conception of the [ purity ^ gametes , brought about by
the segregation of alternative characters. The striking
fact is that this conception, purely theoretical with
MENDEL, has since been confirmed by cytology. In the
mechanism of cell division each chromosome is divided
into two equal parts and each daughter-cell receives
one of these parts. It is a reasonable inference that
chromosomes are bearers of hereditary characters u . In
the production of gametes the number of chromosomes
characteristic of the organism is reduced one-half. As
a consequence each gamete carries only one-half the
characters of the individual that produced it. An
application of these statements to an explanation of
MENDEL'S 3 : i ratio will illustrate the situation.

For convenience we will assume that the nuclei of
MENDEL'S peas have four chromosomes each (fig. 3).
In the case of a tall plant two of the four chromosomes
carry the character for tallness, that is, something that
determines the production of the tall character in the
somatoplasm, which is practically the body builder.
This unknown something is called by various names
in the literature of genetics, the commonest one
being determiner. In our illustration, therefore, two
of the four chromosomes carry the determiner for tall-
ness (p. 33). At this point two questions may be
asked.

Mendel's Law

33

Why do just two of the four chromosomes carry
the determiner for tallness rather than all of them or
only one of them? Just here it would be difficult to
explain why no more than two of the four chromosomes
are represented as carrying the same determiner. This
will be explained later. It is easy to answer, however,
why the determiner is being carried by more than one
chromosome. When gametes are formed the chromo-

F, Hybrid

Dwarf Parent

Gametes

FIG. 3. Diagram illustrating behavior of chromosomes in MENDEL'S
cross of tall and dwarf peas. Large rectangular figures, nuclei of
zygotes or mature individuals; large circles, gametes; small circles
within zygotes and gametes, chromosomes; letters on chromosomes,
determiners (T, tallness; D, dwarfness).

some number is reduced one-half. Since every gamete
from a pure tall plant carries the determiner for tallness
there must have been at least two chromosomes carrying
the determiner before the gametes were formed.

2. Do these two chromosomes carry any other deter-
miner than that for tallness? In a tentative way this
question may be answered in the affirmative, but a
fuller discussion of the situation must be deferred.
There is much experimental evidence that indicates
that more than one determiner is carried on a single

34 Plant Genetics

chromosome. In some cases also there are more Men-
delian determiners than there are chromosomes.

The situation is represented in fig. 3. This shows
a somatic cell with the diploid or 2x number of chromo-
somes. In the formation of gametes this number is
reduced to the haploid, or x number, which in this case
is two. The diagram shows that the reduction separates
(segregates) the two chromosomes carrying the char-
acter for tallness, so that each gamete contains one.
This occurs for the other characters as well as for that of
tallness. From the tall plant, therefore, all the gametes
will contain the character for tallness, and from a dwarf
plant all of the gametes would contain the character for
dwarfness. When these two individuals are crossed the
zygote will contain both characters, and these two char-
acters will be transmitted together in the succeeding
cell generations, "pie individual from such a zygote of
course would be-all, but at the same time it would be
carrying a recessive determiner for dwarfness, and this
fact would be shown by its behavior in breeding. The
result of inbreeding such hybrids is indicated in the ac-
companying diagram (fig. 4), which represents the chance
matings of two kinds of gametes. The obvious results
are three tall individuals and one dwarf. This is the
so-called monohybrid ratio, which means the ratio when
a single pair of allelomorphs is considered.

Before discussing the further development of Men-
del's law it will be necessary to explain some of the
terminology of genetics. When each gamete carries
the same kind of determiner the zygote is said to
receive a double dose; when a zygote receives only a
single such determiner it is said to receive a single dose.

Mendel's Law

35

In fig. 4 one zygote receives a double dose of tallness
and two others a single dose. These phrases are more
or less common in the literature of the subject, but the
more frequent terminology is as follows. When two
similar gametes unite to form a zygote it is called a homo-
zygote; when the two pairing gametes are different the
zygote is called a heterozygotc. Using this terminology

FIG. 4. Diagram illustrating behavior of first hybrid generation
(Fi) when inbred. Illustrates meaning of " segregation" and " purity of
gametes," and how chance matings of Fi gametes result in 3:1 ratio
in F 2 generation; dwarf individual produced only by zygote in lower
right-hand comer.

it is evident that the 3 : i ratio of the F 2 generation is
really a 1:2:1 ratio, as follows: i homozygote for the
dominant character, 2 heterozygotes, and i homozygote
for the recessive character. The 1:2:1 ratio therefore
is the significant one and appears as a 3 : i ratio only
because of dominance.

In the experiment represented in fig. 4 three tall
individuals appear in the F 2 generation. Superficially
the individuals look alike, but it is realized that i differs

36 Plant Genetics

from the other 2 in germinal constitution, for i will
produce only one kind of gamete, while the other 2 will
produce two kinds. . To indicate this situation JOHANN-
SEN (i) has introduced some appropriate terminology.
Organisms which seem to be alike, regardless of their
germinal constitution, are said to be phenotypically
alike, or to belong to the same phenotype. On the other
hand, organisms having identical germinal constitution
are said to be genoty pically alike, or to belong to the same
genotype. From the standpoint of phenotypes only,
MENDEL'S F 2 generation shows the 3:1 ratio; but if
genotypes are considered, it shows the 1:2:1 ratio. In
other words, this group of forms contains two phenotypes
but three genotypes.

Referring again to fig. 4 several things may be
inferred. It can be seen what will happen in the F 3
generation when the F 2 individuals are inbred. The
dominant homozygote will produce only dominant
homozygotes in the F 3 generation and will continue to
produce them as long as it is inbred. The two hetero-
zygotes will split up in the F 3 generation in the same
1:2:1 ratio as did their hybrid parents of the F x genera-
tion. The recessive homozygote will produce only
recessive homozygotes as long as it is kept pure by
being inbred.

It is interesting to consider what will happen if a
heterozygote form is crossed with a homozygous reces-
sive. It should be obvious that one-half of the progeny
would be pure recessives, while the other half would be
heterozygotes, that is, there would be a i : i ratio. A
similar result would be obtained by crossing a hetero-
zygote with a dominant homozygote, although all the

Mendel's Law 37

immediate progeny would show the dominant character.
The real situation would be revealed, however, when this
progeny was inbred, for one-half would be homozygous
(pure breeders) and the other half would be heterozygous
(hybrid breeders).

Thus far we have considered only what is called the
monohybrid ratio, that is, the ratio obtained from one
pair of contrasting characters, such as tallness and
dwarf ness. The next step is to consider the dihybrid
ratio. MENDEL also used contrasting seed characters,
finding, for example, that smoothness in seeds is domi-
nant to a wrinkled condition. Introducing this pair of
contrasting characters into the situation we have been
considering, the dihybrid ratio will be the result. Cross-
ing a tall, smooth-seeded individual with a dwarf,
wrinkled-seeded individual it is evident that all of the Fi
or first hybrid generation will be tall, smooth-seeded
individuals, since both of these characters are dominant.
In the F 2 generation, however, the following ratio will
appear: 9 tall smooth, 3 dwarf smooth, 3 tall wrinkled,
i dwarf wrinkled ; which is a 9 : 3 : 3 : i ratio. This is the
dihybrid ratio, the explanation of which may be .indi-
cated in fig. 5. The question may be raised why the
characters for tallness and smoothness are not repre-
sented on the same chromosome. If they were, the
result would be a simple monohybrid ratio, except that
the tall individuals would always be smooth-seeded as
well, and dwarfs would be always wrinkled-seeded.
The possibility of one chromosome carrying two differ-
ent determiners will be considered later, but at present
we shall assume that these determiners are on different
chromosomes.

38 Plant Genetics

Fig. 5 shows that we are dealing with two homo-
zygotes, each producing only one kind of gamete, so
that all the hybrid progeny will be similar, both geno-

FIG. 5. Diagram illustrating dihybrid ratio. Upper part shows
how original parents were crossed to give Fj hybrid; lower part shows
F x hybrid producing four kinds of gametes; chance ma tings among these
gametes, when F z is inbred, results as indicated in the large set of squares
and explains the 9:3:3:1 ratio in the F^ generation.

MendeVs Law 39

typically and phenotypically, that is, with the same
germinal constitution and the same appearance. By in-
breeding these F x individuals, it will be seen that four
kinds of gametes are involved. Crossing these four kinds
of gametes the resulting combinations are indicated in
fig. 5. The result is four phenotypes, as follows: nos. i, 2,
3> \> 5? 7> 9? I0 > X 3 are tall smooth individuals; n, 12,
15 are dwarf smooth; nos. 6, 8, 14 are tall wrinkled;
no. 1 6 is dwarf wrinkled. This is the 9:3:3:1 ratio.

It will be noticed that nos. i, 6, n, 16 are homo-
zygotes and therefore will breed true; but the rest are
heterozygotes, either for one pair of characters or for
both, and these would split into various types upon
further breeding.

The next step is the trihybrid ratio. MENDEL found
yellow seeds dominant over green seeds, and if this pair
of characters is included with those used above the tri-
hybrid result can be observed. The experiment would
consist in crossing tall, smooth, yellow individuals
with dwarf, wrinkled, green individuals; and it is
obvious that the hybrid progeny would all be tall,
smooth, yellow, since these three characters are domi-
nant. Inbreeding the hybrids gives the following
result in the F 2 generation: 27 tall smooth yellow, 9 tall
smooth green, 9 tall wrinkled yellow, 9 dwarf smooth
yellow, 3 tall wrinkled green, 3 dwarf smooth green,
3 dwarf wrinkled yellow, i dwarf wrinkled green.
The trihybrid ratio therefore is 27:9:9:9:3:3:3:1.
This involves 64 individuals and 8 phenotypes.

1. JOHANNSEN, W., Eiemente der exakten Erblichkeitslehre.
Jena. 1909.

2. MENDEL, G., Versuche liber Pflanzen-Hybridcn. Verb. Naturf .
Vereins in Briinn 4: 1865.

CHAPTER V
NEO-MENDELISM

Thus far we have been considering Mendel's law in
its simple form and have enlarged but little upon
MENDEL'S original statement. The value of the law is
apparent. Upon its republication in 1900 it was taken
up by biologists and numerous breeders set to work to
test it. As a consequence data for and against it began
to accumulate. As might be expected, there was much
apparent evidence against the law, but as geneticists
developed a better conception of the mechanism the
contradictory evidence was explained away. Almost
every type of inheritance has now been explained
according to Mendel's law. Some of the explanations
are very complicated and cannot be included in this
presentation. A few of the more important cases, how-
ever, will be^ presented.

(l) PRESENCE AND ABSENCE HYPOTHESIS

This may be regarded as a new method of Menclelian
thought. It was first suggested by CORRENS (3), but
later was worked out in detail by other geneticists,
especially HURST, BATESON, SHULL, and EAST. It is
merely a modification of the mechanism involved. For
example, in the case of a Hybrid obtained by crossing tall
and dwarf parents the result had been explained as due
to the fact that one chromosome bears a determiner for
tallness and the other one of the pair carries the deter-

40

Nco-Menddism 41

miner for dwarfness. In other words, each one of a pair
of allelomorphs is represented by a determiner, two
determiners thus being present. Dwarfness in this
case would be the result of the interaction of that deter-
miner and its environment during the development of
the body; and the same for tall ness. When both were
present, however, the conception of the situation was as
follows. The determiner for dwarfness, setting up its
usual series of reactions, early became paralyzed by the
determiner for tallness or its products. This result
was called the dominance of the character for tallness.
It was as if the determiner for tallness completely pre-
vented the activity of the determiner for dwarfness.
This conception was apparently borne out by the facts
and was the explanation of the mechanism generally
accepted.

According to the presence and absence hypothesis,
however, the situation is looked at from an entirely dif-
ferent point of view. Tallness is the result of a deter-
miner, but dwarfness is merely the result of the absence
of the determiner for tallness. The dominant character
is produced by an inheritable determiner, but the reces-
sive character appears only when the dominant deter-
miner is lacking. This conception has some evident
advantages and may modify the previous Mendelian
diagram, as shown in fig. 6. This appears to be a simpler
mechanism to account for the phenomenon called domi-
nance. In the case of the dwarf form there is a normal
course of development; in the case of the tall parent or
hybrid, however, an additional determiner stimulates cell
growth, or cell division, or both. It is a simpler and
more useful conception, so long as it fits the facts. Some

42 Plant Genetics

investigators, however, claim that it cannot be applied
to all the situations that have been discovered.

This hypothesis introduces some additional termi-
nology suggested by BATESON. In our illustration the
tall parent has two determiners for tallness and therefore
BATESON calls it^dupleg, having a double dose. For the
same reason the F t individuals, having only one deter-
miner for tallness, he calls simplex. According to the

F, Hybrid

Dwarf Parent

Gametes

FIG. 6. Diagram showing how the original scheme must be modi-
fied to satisfy the presence and absence hypothesis.

same terminology the dwarf parent is nulliplex with
respect to its character of tallness.

Additional advantages of the presence and absence
hypothesis will appear in connection with a considera-
tion of blgnding inheritance and of cumulative factors
in inheritance. Attention, however, should be called
to the fact that those who accept the presence and ab-
sence hypothesis do not use the form of notation thus far
used in explaining Mendelian inheritance. Assume that
T is used to express the determiner for tallness, the same
letter (/) is used to express its absence. For example,
instead of using D for dwarfness, t is used for "lack

Neo-Mendelism

43

of tallness" (fig. 7). It is a matter of convenience
to have a symbol to represent the recessive, the absence
of something that is present in another individual.

In summary, the essential difference between the
presence and absence hypothesis and that of dominant
and recessive is that in the former case the recessive

Dwarf Parent Gametes

FIG. 7. Diagram showing how presence and absence scheme is
actually used, with small letter representing " absence."

determiner has no existence at all, while in the latter
case it exists, but is in a latent condition when associ-
ated with the dominant.

m^ BLENDS

This type of inheritance when first discovered was
thought to be in direct conflict with Mendel's law.
It is a case in which dominance seems to fail, for the
two alternative characters both express themselves and
ithe result is an average between them. It is easy to
explain this situation in accordance with the presence and
absence hypothesis without any violation of Mendel's
law.

The classic example of blending inheritance was
presented by CORRENS (3) in breeding work upon

44

Plant Genetics

Mirabilis Jalapa, the common four-o'clock. CORRENS
crossed red and white varieties, and all the hybrid
progeny had rose pink flowers. This was a color blend,
distinctly intermediate between the colors of the two
parents. The F x generation, therefore, seemed to con-
tradict Mendel's law in that one color character was
not completely dominant over the other. The real situa-
tion, however, appeared in the F 2 generation obtained by

Sperms

P . k (S

.

FIG. 8. Diagram illustrating blending inheritance, discovered
by CORRENS in Mirabilis Jala pa.

inbreeding individuals of the F! generation which showed
the blend. By inbreeding the pink hybrids CORRENS,
obtained the perfect 1:2:1 ratio, that is, i red like one
grandparent, 2 pink like the hybrid parent, and i white
like the other grandparent.) Segregation was evidently
taking place, the only unusual thing being the appear-
ance of the Fi individuals, and that was explained im-
mediately as failure of dominance (see fig. 8).

Neo-Mendelisnt 45

The question this introduces, therefore, is that of a
mechanism which could account for such a result. The
easiest explanation offered is that the red parent was a
homozygote for redness (double dose) and the hybrid a
heterozygote (single dose) ; the inference is that a single
dose produces pink while a double dose produces red.
A theoretical explanation of this occasional difference
in the result of double and single doses is as follows.
Imagine that the body cells of a plant have a certain
capacity for expressing hereditary characters. In such a
case, just as a given quantity of solvent can dissolve
only a given amount of solute, so the body cells can
express hereditary characters only to a definite limited
extent. In the four-o'clock a single dose of redness
may be thought of as half saturating the body cells,
while a double close completely saturates them. In
cases showing complete dominance, however, a single
dose completely saturates the cells and a double dose
can do nothing more. This analogy assists in visualizing
on the one hand the necessary mechanism of blends
(apparent failure of dominance) and on the other hand
that for cases of complete dominance.

Another example of simple blending inheritance is
the case of Adzuki beans, described by BLAKESLEE
(i). In this bean the mottling of the seed coat is
dominant to the lack of mottling. In the hybrid
condition, however, the mottling is lighter than in
the * pure or homozygous condition. Heterozygous
plants, therefore, can be easily distinguished from
homozygous plants, so that the 1:2:1 ratio is evi-
dent on external inspection rather than the usual 3 : i
ratio.

46 Plant Genetics

These two cases are illustrations of what may be
called simple blending inheritance. There are other
cases of blending which are more complex and cannot
be explained so easily.

EAST (4), in crossing starchy and sweet corn, obtained
all starchy in the F x generation, followed by the usual
3:1 ratio in the F 2 . He says that starchiness is due to
the presence of a determiner which enables the corn to
mature starch grains. When this determiner is absent
starch grains cannot be formed, so that the carbo-
hydrates of the endosperm are left in the condition of
sugar, and the result is a sweet corn. This is a simple
case of presence and absence, usually accompanied by
complete dominance. Occasionally, however, the cross
between pure sweet and pure starchy races gave semi-
starchy progeny. The hybrid therefore was inter-
mediate in this regard and would seem to represent a
case of blending inheritance, which of course means
.incomplete dominance. This case, however, cannot be
explained easily by the presence and absence hypothesis
or any other hypothesis. The difficulty is that inherit-
ance, which usually shows complete dominance, occasion-
ally gives way to blending -inheritance. The blend in
the four-o'clock was a constant behavior; in this
corn case it is the occasional behavior; so that the
mechanism proposed to explain the four-o'clock blend
cannot explain this inconsistent behavior of corn.
The best that can be said, apparently, is that some-
thing is interfering with the mechanism; and this
is about all that EAST concludes; he says " some-
thing is interfering with dominance; it is only
partial."

Neo-Mendelism 47

There are two other types of blending inheritance that
may be mentioned, the mechanism involved in which will
be discussed in another connection.

After the first announcement by CORRENS of simple
blending inheritance further investigation revealed
many similar cases; in fact it was soon regarded to be a
common phenomenon that the F x should be an inter-
mediate.' At the same time it was fully expected in such
cases that the hybrid would split in the F 2 generation
in a 1:2:1 ratio. Practical breeders were advised that
Mendel's law was invariable; that a hybrid must split;
and therefore that the only way to preserve a valuable
hybrid was by vegetative propagation. Many breeders
realized the value of this advice, but some reported cer-
tain surprising results. They claimed that they had
made numerous crosses and that in many cases the
intermediate F r individuals continued to breed true to
the intermediate hybrid characters, even when propa-
gated from seed. Gardeners today continue to tell of
certain Begonia hybrids that continue to breed true from
seed in intermediate condition. BURBANK (2) claims
many such cases, in which hybrid blends breed true to
seed, as did their parents. All such claims contradicted
the experience of scientific breeders and seem to destroy
their theories. Accordingly, the Mendelians sought to
explain these claims of pure-breeding hybrids and suc-
ceeded in doing so remarkably soon. There has been
much discussion of this situation and there is no need to
go into the details of it. It will be sufficient to mention
briefly the principal methods by which pure-breeding
hybrids have been explained as not contradicting
Mendel's law.

48 Plant Genetics

Inthe first place botanists suggested parthenogenesis
or vegetative apogamy in the F x , which of course would
mean vegetative propagation and would not involve the
segregation of characters and recombinations in fertiliza-
tion. In such a case, of course, the hybrid characters
would be continued. This is true, but why should the
FI generation exhibit apogamy any more than any other
generation ? An ingenious theoretical answer to this
pertinent question has been suggested by EAST (5) as
follows. "May not the difficulty of maturing sexual
cells in a wide cross sometimes cause apogamous develop-
ment, and therefore a continued propagation of a con-
stant and uniform race?' 7 This suggestion seems
reasonable, for it conforms to two other recognized
phenomena.

JEFFREY (7) and his followers have shown that in
wide crosses the resulting hybrid produces more or less
sterile pollen. An investigation of wild plants has indi-
cated that much hybridizing occurs in nature, especially
in certain families, as indicated by the amount of sterile
pollen, which is sometimes as much as 100 per cent.
Such plants, which reveal their hybrid character only
by the presence of sterile pollen, are called by JEFFREY
" crypthybrids." This fact has also been brought out by
many experimental breeders.

A second significant fact in this connection was
announced by EAST, who says " apogamy is evidently
induced by the extraordinary irritation of foreign pol-
len." The mechanism involved should be considered.
A study of the development of the ovule and female
gametophyte shows that stages are passed through before
pollination, bringing the ovule to a certain stage of

Neo-Mendelism 49

maturity. At that point, however, it was claimed that
development would stop were it not for the stimulus of
pollination and pollen tube development. This stimulus
apparently causes the female gametophyte to continue
its development to the final stage of preparation for
the reception of the sperm. Sterile pollen might furnish
this stimulus as well as fertile. If then pollination by
sterile pollen has taken place and the female gametophyte
has been stimulated to its complete ante-fertilization
development, this development may continue in spite of
the failure of an effective sperm to arrive, and the result
might be an apogamous embryo. In some such way the
relation of apogamy to sterile pollen, resulting from
wide crosses, might be explained.

This conception meets with many objections, so many
that they cannot all be discussed. The chief difficulty,
however, may be stated as follows. Is there any mecha-
nism for the transmission of a pollen stimulus to the
developing female gametophyte ? We know of none,
but neither can we explain the mechanism of the sensitive
plant in which there is obviously the transmission of a
stimulus. Furthermore we know from facts in connec-
tion with parthenocarpy that pollination stimulates
development as deep within the tissue as the ovule
itself, and it would be easy to extend this to the develop-
ment of the female gametophyte as well.

All this is still a field of speculation, but in any event
the reason for EAST'S contention that wide crosseg
may give hybrids that reproduce apogamously is evi-
dent. Wide crosses give sterile pollen; sterile pollen
stimulates the development of the female gametophyte
but cannot effect fertilization; the result is apogamy

50 Plant Genetics

so that the hybrid breeds true to its intermediate
character.

Another explanation of pure-breeding hybrids devel-
ops from the mathematical possibilities of chance mating
when the character in question is due to numerous
separate factors. Without going into the details it
may be said that in such a case the conclusion is reached
that the intermediate hybrids must of mathematical
necessity go on producing intermediates with but a
remote chance of giving back either of the extreme parent
types.

In these various ways the Mendelians explain away
pure-breeding hybrids. In doing so they discard a
great deal of recorded data which seem to them un-
substantiated. A definite statement of the situation may
be obtained from a paper by EAST (5), published in 1910.

It is believed by many that there are kinds of inheritances
other than Mendclian; that is, inheritance in which no segrega-
tion occurs. Far be it from me to deny this. / simply slate
the fact that there are no exact data extant proving other kinds of
inheritance. Such data may be found, but it is useless to specu-
late upon other laws without such evidence. There are several
cases in which either new characters that breed true or blended char-
acters that breed true appear to have been formed, but they have
not been studied with sufficient care for analysis of their mode of
inheritance to be accurate and final.

In 1913 EMERSON and EAST (6), in a joint paper,
stated that there are on record only two indisputable
cases of non-Mendelian characters. These will be
considered later in their proper connection. Neither of
them, however, is a case of pure-breeding intermediate
hybrids.

Neo-Mendelism 5 1

Before leaving the general topic of blending inherit-
ance a statement should be made concerning "garticu-
late inheritance/' which means that in certain crosses
something happens causing unblended fragments of
both parental characters to appear in the hybrid progeny,
resulting in that patchwork of parental characters called
a " mosaic." This is not true blending inheritance, but
has been named by GALTON "participate inheritance."
A common illustration is that of a variegated Amaran-
thus, whose leaves show an irregular mosaic of white and
green. This is produced by crossing pure white and
pure green strains of Amaranthus.

1. BLAKESLEE, A. F., and AVERY, B. T., Adzuki beans and Jimson
weeds. Jour. Heredity 8:125-131. Jigs. 5. 1917.

2. BURBANK, LUTHER, Another mode of species forming. Ann.
Report Amer. Breeders Assoc. 5:40-43. 1909.

3. CORRENS, C., Die neuen Vererbungsgesetze. Berlin. 1912.

4. COULTER, CASTLE, DAVENPORT, EAST, and TOWER, Heredity
and Eugenics. Chicago. 1912.

5. EAST, E. M., The role of hybridization in plant breeding. Pop.
Sci. Monthly 77^342-355- figs. 12. 1910.

6. EMERSON, R. A., and EAST, E. M., Inheritance of quantitative
characters in maize. Bull. Agric. Exper. Sta. Nebr. no. 2.
pp. 120. figs. 21. 1913.

7. JEFFREY, E. C., Spore conditions in hybrids and the mutation
hypothesis of DE VRIES. Bot. Gaz. 58:322-336. pis. 22-25.
1914.

CHAPTER VI

NEO-MENDELISM (Continued)
(uj) THE FACTOR HYPOTHESIS

MENDEL concluded that each plant character de-
pends upon a single determiner. Inheritance, however,
has proved to be a much more complex phenomenon
than indicated by MENDEL'S peas. Ratios have ap-
peared that were puzzling, and geneticists were forced
to the conclusion that there may be a compound
cleterminer for a single character. This conception is
called the factor hypothesis, and the growing complexity
of genetics has developed in connection with this
hypothesis. With the consideration of factors instead
of determiners one passes from elementary to advanced
genetics. Previously we have used the word determiner,
implying MENDEL'S idea that a single determiner is
responsible for the development of a plant character, and
this has been true of the examples of inheritance previ-
ously considered. It is understood now, however,
that a character is frequently determined by the inter-
action of two or more separately heritable factors, and
hence the factor hypothesis. The distinction between
factors and determiners should be made clear. In case
only one factor is involved in determining a character,
there is no distinction between factor and determiner;
and in such a case the term factor should not be used.

'(D COMPLEMENTARY FACTORS. This is the simplest
expression of the factor hypothesis and it may be
illustrated by some of EAST'S work. Crossing red-

5 2

Neo-Mendelism 53

grained and white-grained corn he obtained all red in
the F| generation. This would suggest that the F 2
generation would show 3 red to i white; but it showed
9 reds to 7 whites, which did not suggest Mendelian
inheritance. It is in accord with Mendel's law, how-
ever, if we consider that two complementary factors are
necessary to produce the red character, and that each of
these factors is inherited separately. Such a situation
would give a dihybrid ratio, as indicated in fig. 9. It
will be seen that out of 16 progeny 9 will be red, for they
alone contain the complementary factors; the other 7 will
be white. The situation is thus explained by the dihybrid
ratio, but although only one character is involved that
character depends upon two complementary factors.

Another situation is worth noting. No. 6 of the
diagram is white because it contains only one of the
necessary factors; no. n is white for the same reason,
but its germinal constitution is just the opposite. What
would happen if these two are crossed ? There is only
one possibility, since each is a homozygotc producing only
one kind of gamete. The result would be red, and
thus a cross between two whites would produce only
reds. What would happen from crossing nos. 6 and 15,
the former being a homozygote and the latter a hetero-
zygote ? It is obvious that the resulting progeny would
be one-half white and one-half red, although both parents
are white. The same result would be secured in cross-
ing nos. ii and 14. A cross between nos. 14 and 15,
both of which are heterozygotes, would result in 3
whites and i red, the ordinary 3:1 ratio. These illus-
trations show how differently the same phenotype may
behave in inheritance. In each case two whites were

54

Plant Genetics

crossed, that is, the same phenotypes, but three different
ratios were obtained because the genotypes were different.

FIG. 9. Diagram illustrating behavior of complementary factors
in cross between red-grained and white-grained corn. R and C must
both be present to produce red-grained corn.

The striking feature of this situation is that one can
cross two whites and get a red. This gives an insight
into the so-called phenomenon of reversion. For ex-

Neo-Mendelism 55

ample, in the course of numerous breeding experiments
BATESON (i) obtained two strains of white sweet peas,
each of which when normally "selfed" bred true to the
white color; but when these two were artificially crossed
all the progeny had purple flowers, like the wild Sicilian
ancestors of all cultivated varieties of the sweet pea.
This appeared to be a typical case of reversion. Further
breeding, however, showed that this was just such a
case of complementary factors as we have been con-
sidering. One of BATESON'S white strains had one of
the factors for purple and the other strain had the other
factor.

Complementary factors have been defined and the
method of their inheritance described, but is there
any mechanism to explain the situation ? A suggestion
may be obtained from plant chemistry (2). The most
prominent group of pigments in plants is the group of
anthocyanins, which are produced as follows. Plants
contain compounds called chromogens, which are color-
less themselves but which produce pigments when acted
upon by certain oxidizing enzymes or oxidases. This is
a sufficient mechanism for the behavior of comple-
mentary factors. If one of EAST'S white strains of corn
contained a chromogen capable of producing red but
lacked the necessary oxidase it would remain colorless.
If the other white strain contained the oxidase but no
chromogen it would remain colorless. In crossing them,
however, chromogen and oxidase would be brought
together and a red-grained hybrid would be the result.
Inbreeding such red-grained individuals of course would
give red and white progeny in a ratio of 9; 7, as explained
in connection with EAST'S corn. This seems to be the

56 Plant Genetics

explanation of the behavior of complementary factors
in many cases of color inheritance.

Where other characters are involved the mechanism
must be somewhat different. In some cases the two fac-
tors may be the enzyme and the compound the enzyme
attacks, as in the oxidase and chromogen situation just
described. Qn_the other hand, we might be dealing
with two chemical compounds that are inert when occur-
ring separately but active when brought together,
active in such a way as to produce a distinctly new
character. Also two active substances might neutralize
one another when brought together in a hybrid, and the
failure in their activity might result either in a new char-
acter or the failure of some parental character to develop.
Such are some of the possible mechanisms to explain the
behavior of complementary factors.

Hybridizing, therefore, is much like mixing chemicals
in a test tube. We know that very wide crosses cannot
be made successfully; but the surprising thing is that
certain very close crosses are constantly unsuccessful,
even though both parents may cross freely with closely
related types. We obtain a glimpse of the possibility
of such apparently inconsistent behavior when we
consider the chemical possibilities suggested by the
behavior of complementary factors.

The origin of complementary factors is an interesting
field of speculation. Did they originate together or
separately? A natural inference would be that they
originated together, for neither would be of any use with-
out the other. It should be remembered, however, that
the idea of use as explaining the occurrence of every-
thing in a plant is being abandoned; one must think

Neo-Mendelism 57

rather of a plant as a complex physico-chemical labora-
tory. No one claims that all chemical reactions are
useful; they are simply inevitable ; and plant characters
are the result of chemical reactions and physical necessi-
ties. Even though we assume the simultaneous origin
of two complementary factors, they would have to be put
on separate chromosomes, for the factors are separately
inherited.

The other alternative is to suppose that these factors
originated independently in the history of a plant. In
this case, of course, the first one to be produced would
remain functionless until finally its complement came
into existence. This might be an explanation of what
are called latent characters. Also they might have not
only originated independently but in different varieties
or species. In this case if natural hybridizing should
bring them together the result would be the appearance
of a new character, and this may have been a very im-
portant factor in the origin of species.

This may serve as an introduction to the factor
hypothesis, with complementary factors as an illus-
tation, simply because it is the simplest situation. There
are many other kinds of factors recognized, but we shall
not attempt to list all of the proposed types. A simple
illustration of the better known types is as follows:

a) A complementary factor is added to a dissimilar
factor to produce a particular character.

b) An inhibitory factor prevents the action of some
other factor.

c) A supplementary factor is added to a dissimilar
factor with the result that the character is modified in
some way.

58 Plant Genetics

d) A cumulative factor, when added to another similar
factor, affects the degree of development of the character.

Some examples of these types will make them clear,
those for complementary factors having been given
previously.

INHIBITORY FACTORS. Recalling EAST'S experi-
ment with red-grained corn it will be remembered that
when both factors for red were present the grain was reel,
but when either factor was absent the grain was white.
Later he crossed these strains with a new white strain,
and the result was surprising (3). The pure red strain
produced gametes carrying both the red factors, and
it would be expected that whatever such a gamete
mated with would result in red progeny; but when this
pure red was crossed with the new strain of white the
progeny were all white, although the hybrids certainly
contained both factors for red. The explanation which
first occurred to EAST, and which later experiments
confirmed, was that the new white strain contained
an inhibitory factor, which prevented the development
of red even though both the complementary factors for
red were present. Fig. 10 illustrates the situation and
shows why all the individuals of the Fj generation are
white. It is interesting to note further the possibilities
of white and red in the F 2 generation. They would be
numerous, since we are dealing with trihybrid ratios
(see fig. n). This does not exhaust the possibilities, for
the cases given were homozygotes, each producing a
single kind of gamete. There remains for consideration
the heterozygote situation (see fig. 1 2) .

The possible mechanism of the inhibitory factor
is as follows. We have assumed that red is produced

Neo-Mendelism

59

Pure Red Parent

White Parent with
Red Inhibitor

F, White
FIG. 10. Diagram illustrating behavior of inhibitory factor

White

FIG. ii. Diagram showing some possible combinations in F 2
when F x of fig. 10 is inbred. Individual on left end of upper set red-
grained, because R and C both present and / absent; other individuals
in upper set white, because lacking C or R or both; individuals in lower
set with inhibitory factor and therefore white, whatever other combina-
tions of factors they may contain.

60 Plant Genetics

only when the enzyme is present to oxidize the chromo-
gen. Enzymes are very sensitive; their activities may
be affected or completely checked by various agents.
Suppose that 7 of the diagram be such an agent and the
necessary mechanism is apparent. When / is present
R is paralyzed, so that it cannot oxidize C.

@ SUPPLEMENTARY FACTORS. A supplementary fac-
tor is one that is added to a dissimilar factor, with the
result that a character is modified in some way.

In his work upon red-grained races of corn EAST
found occasionally a few purple grains. His conception
of the situation is as follows (3) . The
pure red plant contains two comple-
mentary factors, one (C) a chromo-
gen, and the other (R) an enzyme,
which when brought together pro-
(i\ (l) duced the red color. The purple
^-^ vlx grains, however, must be explained

F IG . I2 by the presence of still another factor

(P), the resulting situation being
represented in fig. 13. Of course when C is absent no
pigment whatsoever can be produced. As a consequence
we will assume that the presence of C is constant, and
that P and R are variables. For a similar reason we
will assume that the absence of I is constant. The
figure shows three possibilities, from which the following
conclusions may be drawn: (i) when P and R are both
present the result is purple grains; (2) red appears only
in the absence of P; (3) P although present will not
develop any color in the absence of R.

This is a typical case of a supplementary factor, that
is, one which is added to a dissimilar factor, with the

Neo-Mendelism

61

result that the color character is modified. The mech-
anism of this situation will make clearer the behavior
of the supplementary factor. If C is the chromogen
and R the enzyme, what is P? The suggested answer
can be obtained from plant chemistry. It is found
that the purple pigment is produced by the same sub-
stance as the red, but represents a higher state of oxida-
tion. The conclusion is obvious. C is oxidized by R
up to a certain point, where red is produced; but if P

Purple Red WKite

FIG. 13. Diagram illustrating action of supplementary factor

is also present it represents an additional enzyme, which
attacks the red pigment and oxidizes it still further into
purple. P is incapable of attacking the original chromo-
gen, but when R carries the attack to a certain point,
P can function and carry the oxidation further. As
a consequence P without R gives white grains, while R
gives red grains only in the absence of P.

(Q CUMULATIVE FACTORS. These will be considered
under the next heading, " Inheritance of quantitative
characters."

In addition to the four types of factors given,
the literature of genetics also contains discussions on

62 Plant Genetics

intensifying factors, diluting factors, distribution fac-
tors, etc. These, however, do not introduce any new
mechanisms.

1. BATESON, W., Mendel's laws of heredity. Cambridge. 1909.

2. CZAPEK, P. and M. E., Biochemie der Pflanzen. Jena. 1913.

3. EAST, E. M., and HAYES, H. K., Inheritance in maize. Conn.
Agric. Exper. Sta. Bull. no. 167. pp. 142. pis. 25. 1911.

CHAPTER VII

NEO-MENDELISM (Continued)

(jfr INHERITANCE OF QUANTITATIVE CHARACTERS.
This phase of the factor hypothesis, if true, is of funda-
mental importance not only to genetics but to general
biology. It is based upon the conception of cumula-
tive factors, and as it is presented it will be realized
that it throws light not only upon numerous breeding
experiments but also upon variation in general, which
means evolution also. A cumulative factor was defined
as one which, when added to another similar factor,
affects the degree of development of the character.

It will be recalled that CORRENS crossed reel and
white strains of Mirabilis and obtained pink hybrids.
The suggested explanation of this result was that a
single dose of the red determiner gives pink while a
double dose gives red. When CORRENS inbred these
pink hybrids, he obtained the result presented in fig. 8,
that js, i red, 2 pink, i white. This result is obvious
and the mechanism is plain.

With this diagram in mind we shall consider some of
the experiments of NILSSON-EHLE (2) at the Swedish
Experiment Station. He crossed two strains of wheat
with red and white kernels. The Fj individuals had
light red kernels, which of course suggests a repetition of
the situation shown by Mirabilis in the experiment of
CORRENS. The F 2 generation, however, showed a very
different result. The reds and. whites appeared in the

63

64 Plant Genetics

ratio of 15:1; but in addition to this, among the 15 reds
there could be distinguished varying degrees of redness.
NILSSON-EHLE suspected that 15:1 meant a dihybrid

FIG. 14. Diagram illustrating NILSSON-EHLE'S explanation of 15: i
ratio obtained in F 2 generation from cross between red-grained and
white-grained wheat.

ratio, 1 6 individuals being necessary to give the ratio,
so that he constructed the tentative scheme shown in
fig. 14.

Neo-Mendelism 65

This shows a regular dihybrid ratio, except that the
two factors involved are similar. Applying the single
dose and double dose conception, as used in the case of
CORREN'S pink Mirabilis, we reach the following con-
clusions: no. i only has four doses and therefore it only
is deep red; nos. 2 7 3, 5, 9 have three doses and are some-
what lighter red; nos. 4, 6, 7, 10, n, 13 have two doses
and are still lighter red; nos. 8, 12, 14, 15 have one dose
and are very light reel; while no. 16 alone has no dose
and is the only pure white. This accounts for the 15: i
ratio, and the different shades of red. This is entirely
in accord with the conceptions that have been presented,
and only two assumptions are necessary: (i) that domi-
nance is absent, and two doses have twice the effect of
one; (2) that the independent similar factors are cumu-
lative in their operation, and are paired with their
absence in the hybrid. This was NILSSON-EHLE'S
conception, and of course he tested it by further experi-
mental work, the results consistently confirming the
conception.

Since it is important to fix this conception clearly in
mind, another type of diagram may represent the facts
even more clearly. The proportion of individuals
showing the various degrees of redness in the F 2 is
graphically recorded in fig. 15, each dot representing
one dose of the factors in question.

Continuing these investigations, NILSSON-EHLE next
discovered a new strain of red-grained wheat, which,
when crossed with the pure white strain, yielded Fi
hybrids of intermediate intensity of red as before.
The F 2 generation, however, showed a different situation.
Reds and whites were obtained in the proportion of

66

Plant Genetics

63 : i ; the 63 reds as before falling naturally into differ-
ent groups on the basis of degree of redness. Applying
the same conception as before NILSSON-EHLE discovered
that in this case he was dealing with a trihybrid situation.

Pure Red

Grades of Pink

White

FIG. 15. Another method of visualizing NILSSON-EHLE'S 15:1
ratio (see fig. 14).

Without constructing the usual Mendelian diagram,
which would have to be extensive enough for 64 indi-
viduals, the situation as it appeared in the F 2 generation
may be represented by fig. 16. If the graph is sur-
mounted by a curve we recognize the regular " proba-
bility curve/' exactly the kind of curve biometrigians

Neo-Mendelism

6 7

While

Pure Rcd Intermediate Grades

FIG. 16. Diagram illustrating NILSSON-EHLE'S 63: i ratio

68 Plant Genetics

use to represent the fluctuating individuals about a
specific type.

This conception of cumulative factors, therefore,
has far-reaching significance. For a long time biologists
have recognized individual variation within the species.
DARWIN depended upon it as the basis of his theory of
natural selection as the origin of species; in fact, ever
since DARWIN'S Origin of species, individual variation
has been fundamental in our conceptions. To account
for this universally recognized phenomenon, DARWIN
proposed his transportation hypothesis as a possible
explanation, which, as will be recalled, did not long
survive. WEISMANN offered in explanation his germinal
selection, which was soon discarded because it was beyond
the possibility of experimental testing. Aside from
these two attempts to explain individual variation no
other comprehensive scheme had been presented. Bi-
ologists had simply recognized the fact of individual
variation without any conception of the mechanism.
They knew that individual variation existed but had
even stopped asking why it existed.

The importance of this new theory, therefore, is
obvious. It is an ingenious explanation of the inherit-
ance of quantitative characters and of the existence of
individual variations. Furthermore, the theory has
not been developed through meditation, but has its
basis in scientific experiments. It is imaginative to a
certain extent, of course, as is every other valuable
theory, but unlike most such theories it has a substantial
foundation, namely, Mendel's law.

The importance of the possible role of cumulative
factors in explaining individual variation, which in

Neo-Mendelism 69

turn is the basis of evolution, has been emphasized
because its importance has perhaps not yet been suffi-
ciently appreciated. It promises to be one of the most
important theories of biology, which of course must be
tested by future investigation.

The doctrine of cumulative factors was further
developed by EAST (i) in his work with corn. He was
able to explain some of the ratios obtained by assuming
three or four separately inherited cumulative factors,
just as NILSSON-EHLE had done. He obtained other
ratios, however, which required more independent
cumulative factors to explain. Some idea of the extent
of these investigations by EAST and his associates may
be obtained by noting the list of the plant characters
whose inheritance they explained on the basis of cumu-
lative factors: number of rows, length of ear, diameter
of ear, weight of seed, breadth of seed, height of plant,
number of stalks per plant, earliness of flowering. In
all of these cases breeding gave the same characteristic
results. A cross between extreme parents gave hybrid
progeny intermediate as to the character in question;
and in the Y 2 generation the two extremes reappeared,
along with all gradations of intermediates.

NILSSON-EHLE had been able to put his F 2 inter-
mediates into definite classes, corresponding to the
number of doses of the determiner each had received.
EAST, however, could not do this with such exactness.
His results showed all gradations, but he could not
distinguish any definite groups; that is, gradation was
continuous and complete. In other words, he could
not tell with certainty from outward appearance just
how many doses of a given determiner an individual

70 Plant Genetics

contained. His results, therefore, do not seem so clear
and striking as those of NILSSON-EHLE, but they are by
no means vague and uncertain. For example, even if he
could not say definitely that a certain individual had
exactly three doses, he could always say approximately
how many doses it had; and the breeding results always
confirmed the idea of a number of cumulative factors
at work. For example, a plant with three doses may
vary with respect to the character in question. It may
approach the condition of the plant with four doses or
it may vary toward the two-dose condition. Such
variation may be explained by outside influences. Any
classification of the F 2 individuals on the basis of the
number of closes is more or less obscured by the influ-
ence of outside factors which are uncontrollable, or at
least uncontrolled as yet in breeding work.

EAST has visualized these outside factors and dis-
cussed them. In order to explain this discussion,
however, we must recall a feature of genetics which
has previously been mentioned. Plant variations in
the largest sense fall under two categories, due to
(i) differences in gametic constitution, and (2) responses
to environment. The first category is the basis of all
Mendelian conceptions, while the second category in-
cludes such variations -as are usually thought not to be
inherited, being, acquired characters. This category
is now commonly called fluctuating variations.

An illustration will make these two categories clear.
Assume that a plant with a determiner for tallness
usually becomes 6 feet, while one without this determiner
becomes 3 feet. The 6-foot plant, however, grown in
good soil becomes 6 . 5 feet, while in poor soil it is 5.5

Neo-Mendelism 71

feet. In inheritance of course the 6.5- and 5.5-foot
plants behave exactly alike; the same is true of 6-foot
plants. It must be evident, therefore, that a classi-
fication of F 2 individuals on the basis of the number of
doses might well be slightly obscured. If outside influ-
ences were lacking, the F 2 situation could be represented

FIG. 17. Diagram illustrating distribution of phenotype classes
in an F 2 population from cross involving cumulative factors. Practically
same diagram as fig. 16, and interpreted in same way; short rectangle
at left indicates that very few plants of population contain maximum
number of doses; short rectangle at right indicates that very few plants
contain minimum number of doses; plants with intermediate number of
doses most numerous, as indicated by tall rectangle in middle (see also
%. 1 8).

by fig. 17, or better, fig. 18; but when outside influences
are active, it may be represented by fig. 19. It will be
seen from this last diagram that not all individuals
belonging to a particular size class may have the same
number of doses; that is, conditions for a smaller-dosed
individual may be so much better than those for a larger-
dosed individual that they may exchange size classes in

72 Plant Genetics

the result. In this way the results of germinal consti-
tution may be somewhat obscured by the varying
external conditions of growth.

FIG. 18. Graphic illustration of situation represented in fig. 17. In
both diagrams it is evident that two plants appearing in same quanti-
tative class must have same number of doses; this should always occur
if environmental influences did not obscure the result.

FIG. 19. Diagram illustrating how environmental influences may
obscure phenotype classes of F 2 . Overlapping of phenotype classes
makes possible that two apparently identical plants might actually have
a different number of doses; diagram also shows that while breeder
could not recognize whether a plant had two or three doses, he could
distinguish between plants of two and four doses, etc.; thus intelligent
selection could be effective.

Neo-Mendelism 73

Another factor that may obscure these results is what
is called physical correlation. For example, a corn
plant of small size but with the hereditary capacity for
producing large ears could not fully express this capacity.
It could not produce as large ears as if it had been a
large-sized plant.

Still another factor that might obscure the result
may be called gametic correlation, which is a conception
presented by EAST. The idea is that factors which are
inherited quite independently and which affect different
plant characters might still be conceived of as having an
influence upon one another. As EAST puts it, "A gamete
may be a mosaic of independent factors, but the plant
(produced from the gametes) will not be a mosaic of all
the characters and factors produced, for these factors
act and react upon one another in complex ways during
their development."

Such are some of the conditions or factors that tend
to obscure results in the F 2 generation and give rise to
ratios hard to interpret. The weaker the influence of
these factors the more clearly do the phenomena of
cumulative factors come out. The total result of this
phase of EAST'S work, in spite of obscuring conditions
that have arisen, has been to strengthen greatly the
conception of cumulative factors. A summary of his
conclusions may be stated as follows:

When one is dealing with quantitative characters, that is,
those produced by cumulative factors, he is confronted by exactly
the same principles of Mendel ian inheritance as have long been
known to apply to qualitative characters. With quantitative
characters, however, the problem is more complex, due chiefly to

74 Plant Genetics

factors cumulative in their operation; (2) the significance of the
breeding results is usually somewhat obscured by the natural
fluctuations due to response to uncontrollable factors in the
environment.

Others have investigated the problem of cumulative
factors, and many of the results are favorable to the con-
ception; but much more investigation should be made,
for the conception deals with exceedingly fundamental
situations. If it is true it is extremely important; if it
is not true a knowledge of this fact is just as important.
It is like a fork of the roads in our biological progress; it
is important to know which is the right road for future
progress.

In concluding the general topic of cumulative factors
a modifying statement should be made as to the mecha-
nism involved. Heretofore it has been assumed that we
were dealing with numerous, separately inherited factors,
absolutely identical in their nature, cumulative in their
effect. No doubt one might regard with suspicion such
a seemingly artificial mechanism. Probably it would
be easier to believe if it were modified in the follow-
ing manner. Instead of assuming numerous factors,
identical in function, we may assume that each of these
factors has its own peculiar function, but that that
function plays a part, directly or indirectly, in develop-
ing the quantitative character in question. For example,
suppose height is the character. One of the factors
determines the development of long internodes; another
results in numerous nodes; another increases the amount
of chlorophyll; another determines the size and vigor
of the root system; another brings early germination
and long growing season. Such factors, then, although

Neo-Mendelism 7 5

not identical, will be cumulative in increasing the height
of the plant. Of course a single dose of one type of
factor will not bring the same increase in height as would
a single dose of one of the other types, and therefore the
mathematics of the situation would be slightly modified.
The fundamental mathematical system, however, would
remain the same, and we would have the satisfaction of
dealing with a natural mechanism rather than an
artificial one.

1. EMERSON, R. A., and EAST, E. M., The inheritance of quanti-
tative characters in maize. Bull. Agric. Exper. Sta. Nebr.
no. 2. pp. 120. figs. 21. 1013.

2. NILSSON-EHLE, H., Einige Ergebnisse von Kreuzunger bei
Hafer und Weizen. Bot. Notiser 1908: 257-294.

, Kreuzungsuntcrsuchungen an Hafer und Weizen.

Lunds Univ. Arsskr. N.S. II. 5:1-122. 1909.

CHAPTER VIII

NEO-MKNDELISM (Continued)

() TRUE-BREEDING HYBRIDS. This situation was
referred to in chapter v, but comes up again in the
present context. The statement was made that a num-
ber of practical breeders had reported that by crossing
distinct races they had not only obtained hybrids inter-
mediate with regard to one or more characters, but also
that these hybrids when inbred continued to breed true
to their intermediate condition and that the original
parent races never reappeared.

One explanation of this situation offered by geneti-
cists was tHat where a great many factors are involved
it is practically a mathematical impossibility for either
of the parent races to reappear, for the chances would be
overwhelmingly against the exact coming together of the
exact combination of factors. This explanation should
be considered a little more fully in the light of what has
been stated concerning cumulative factors.
_ , JSuppose that CORRENS in crossing the red and white
races of Mirabilis had been dealing, not with one factor
determining redness, but with six, the red color being
determined by six different factors, cumulative in their
effect but separately inherited. With only two cumu-
lative factors the F 2 situation is represented by fig. 15,
where it is evident that one individual of each parent
type appears in 1 6 of the F 2 individuals. This ratio

76

Neo-Mendelism 77

may be referred to as 1:4:6:4:1. When three cumula-
tive factors are considered the F 2 ratio would become
1:6:15:20:15:6:1 (see fig. 16), that is, one of each
of the original parents out of 64 individuals. In the
same way the six cumulative factors assumed, which
is really a conservative number, would give the following
F 2 ratio : i : 1 2 : 66 : 220 : 495 : 792 : 924 : 792 : 495 : 220 : 66 :
12:1, that is, one of each of the original parent types
out of 4096 individuals of the F 2 generation.

With these facts it is easy to explain the experience of
breeders and horticulturists. In the first place, practi-
cal breeders are not as carefully discriminating as are
geneticists, so that they would put their individuals of
Mirabilis into three classes, that is, red, pink, and white,
without considering the various degrees of redness or
pinkness. Therefore if they were dealing with six
cumulative factors and grew less than 4096 individuals
of the F 2 generation, the chances are that they would
obtain no individuals representing either the red parent
or the white parent. Naturally they would conclude
that the original cross produced an intermediate hybrid,
which when inbred without exception bred true to the
intermediate hybrid character. The question might be
raised in reference to the later generations. Would not
the original parent type reappear in some generation
later than F 2 ? The mathematics of the situation is too
complex for presentation, but the result can be visualized
in a less exact way. It has been seen that the vast
majority of the F 2 are intermediates of varying degrees,
that is, when the F x intermediates were inbred the result-
ing F 2 was made up of practically all intermediates and
no extremes. Inbreeding F 2 , therefore, would continue

78 Plant Genetics

the same situation, so that the chances of a reappear-
ance of the original parent types would be very slight,
so slight that if they did appear they would be described
as " reversions." In this way many of the cases of true-
breeding hybrids can be explained. Of course we have
been dealing only with intermediate hybrids which are
intermediate quantitatively, due to the action of cumu-
lative factors. The gardeners and horticulturists might
claim that their hybrid is not merely different in degree
from either parent but that it is different also in kind;
that it exhibits an entirely new character, that is, it
differs qualitatively. If this is true of course the
explanation suggested does not apply. The explana-
tion can be modified, however, leaving the mathematical
possibilities the same. Suppose the original parent
strains differed, not by* six cumulative factors, but by
six factors of different kinds. Under these conditions
it would be possible to bring together in- a cross an
entirely new combination of factors which might result
in characters that would seem entirely new. If we are
dealing with complementary or supplementary factors
we might get a character in a hybrid that is entirely new.
If these factors were sufficiently numerous, the chances
of separating them again and recombining them exactly
as in the original parents would be exceedingly slight.

These possibilities have been presented for two rea-
sons: (i) as explanations of true-breeding hybrids,
whether quantitatively or qualitatively different from
their parents; (2) to illustrate the method of using the
conceptions of the factor hypothesis.

(^ A PRACTICAL ASPECT OF THE CUMULATIVE FACTOR

HYPOTHESIS. Assume that a practical breeder crosses

Neo-Mendelism 79

two extreme parent types in the hope of obtaining -a
hybrid combining the desirable characters of the two
parents. If the material is corn, he might use one
parent with large grains but few in number, while the
other parent has many grains but small ones. Such
quantitative characters as these would be determined
by cumulative factors, and the hybrid would be inter-
mediate in respect to both of these characters, that is,
the grains would be of medium size and medium number.
No matter how many crosses he made, he would always
get this result and not the desired combination of large
grains and many of them.

Suppose now that these intermediate hybrids are
inbred in the hope of obtaining the desired combination
in the F 2 generation. It will be realized that the chances
of obtaining a plant combining the two extreme char-
acters of large grains and numerous grains would depend
upon the number of factors that enter into the make-up
of these quantitative characters. Assume that there are
five factors in each case. The mathematics of the situa-
tion would show that in order to get the desired pure
type from a cross between two parents, each having
their desirable character determined by five cumulative
factors, it would require 100 acres of corn to have an
even chance of getting one such individual in the F 2
generation. It is altogether unlikely that any farmer
would use 100 acres and a corresponding amount of labor
on such an extreme chance. Even if he did, it would be
very problematical whether he would be able to select
the proper solitary individual on his 100 acres. Even
an agricultural experiment station would not feel justified
in conducting such an experiment.

80 Plant Genetics

The question arises whether there is any way of
avoiding this impossible situation. The escape is sug-
gested by the fact that time can take the place of
numbers. EAST (2) has shown that by growing 1000
individuals in the F 2 generation, 100 in the F 3 , and
50 in the F 4 , one stands as much chance of getting the
desired combination as by growing 250,000 in the F 2 ,
provided an intelligent selection is made in each genera-
tion. In other words, one who understands the mecha-
nism of the inheritance of quantitative characters will
grow only 1000 individuals in his F 2 generation and will
select for seed only those individuals with the right
number of factors. In this way, by intelligent selection,
[actors are piled up in the right direction from year to
^ear. In a few years the desired result will be reached
without the necessity of growing a very large number
Df individuals. Such work is practicable at experiment
stations, and it is the kind of work that a number of them
have been doing recently. Even the ordinary farmer
should be able to do such work. Although his selection
of individuals should not be quite as intelligent as that
of a scientific breeder, he would probably be selecting in
the right direction and making some advance. A little
more time and a little more acreage would bring him
to the desired result.

AJuj^her^application of the factor hypothesis may be
considered. The practice we have been discussing under
the name of the inheritance of quantitative characters seems
to be little more than what has already been called
artificial selection, which is the oldest of all methods of
plant breeding. It is a method that was thought to
be discredited entirely by the work of DE VRIES (i),

Neo-Mendelism 8 1

JOHANNSEN, and others when they discovered what
they called elementary species or pure lines and demon-
strated that artificial selection could never result in any
large or permanent improvement. In consequence of
this, artificial selection, as the most important method
of securing desirable races, gave place to pedigree
culture at the most important experiment stations. It
was not abandoned, for it had its uses, but it seemed to
many to be a mediaeval method of breeding. The
artificial selection, however, which we have been de-
scribing is different from that so long practiced. In
brief statement the difference is as follows.

In the first place, the selection proposed is preceded
by an intelligent hybridizing, and after that genotypes
rather than phenotypes are selected; that is, the selection
is made on the basis of germ plasm rather than of body
plasm. This would be a sufficient reason for the superi-
ority of the new method of artificial selection as com-
pared with the old. A little further analysis, however,
will make the difference clearer.

In the old method of artificial selection the breeder,
in the first place, is dealing with such germinal variations
as happen to appear in his crop; and in the second place
he is dealing with those fluctuations which appear as
responses to the environment. When he selects a large
plant to use for seed that plant may be large on account
of its germinal constitution; but on the other hand it may
be large because it is growing in a less crowded place
or is more heavily fertilized than the others. In that
case the large plant might not furnish good seed. The
plant breeder of the old method undoubtedly made such
unfortunate selections frequently; that is, he selected

82 Plant Genetics

on the basis of external appearance, and external appear-
ance is often a^yery poor index of hereditary capacity.
Intelligent selection is based on germinal constitu-
tion only, and therefore its results are quicker and
surer.

Another phase of the subject should be considered.
When a plant breeder is trying to improve his crops by
selection for quantitative characters, although he uses the
old method of selection, he is likely to be making some
gain, as the experience of hundreds of years has shown.
The germinal constitution of his crop plants is masked
by fluctuations of course, but this mask is not complete.
Most of the plants he selects are bound to possess high
numbers of factors of the right kind, and he probably
rejects most of the plants with few factors. In any
event, he has generally succeeded in the long run in
getting a somewhat improved race.

A summarized statement of this situation may be
helpful. Our recently developed knowledge of the in-
heritance of quantitative characters seems to justify
artificial selection, but it does not justify the old blind
method of selection. It emphasizes the need of intelli-
gent, trained selection and shows how such selection can
be made. In order to do this one must understand
the mechanism of the inheritance involved and must
understand the make-up (race or pedigree) of the plants
dealt with, being sure that they are of pure race or
strain, for selection from the ordinary mixed races of crop
plants is not only tedious at best but often leads to
chaotic results.

The situations just considered enable one to under-
stand two phenomena which have been baffling scientific

Neo-Mendelism 83

plant "breeders for some years. The races of plants
improved by artificial selection have usually reverted
to type when selection ceases. This fact was recognized
for a long time, but was first pointed out clearly by DE
VRIES (i). Since then we have always expected this
result, that no improvement will maintain itself but
will run back unless the selection is continuous. When
a practical breeder announces that he has developed
by selection a new race which continues to breed true
without further selection we are inclined to disbelieve
him, for we know that only elementary species breed
true. We explain that the practical breeder bases
his selection on fluctuations, and therefore his new
race is bound to revert to type. It is obvious now
that there is a flaw in this argument. The practical
breeder may be basing his selection on fluctuations, but
at the same time he may be piling up cumulative fac-
tors in the right direction. Thus he might eventually
secure a race containing all the cumulative factors.
Such a race would be a homozygote and could not help
breeding true. Most of the claims of artificially im-
proved races that breed true may be false, but it should
be remembered that such a thing is possible and may
be stumbled upon accidentally, even with unscientific
breeding.

There is another phenomenon which has been
much discussed, and which can now be explained in
the same way. This is the so- called fixation of hybrids.
For years breeders have made promiscuous crosses and
then begun artificial selection with the F 2 generation.
Eventually they have secured a pure-breeding hybrid.
It will be remembered that it was in this way EAST

84 Plant Genetics

worked with the quantitative characters in corn, and
the explanation is the same.

In addition to the practical value of the conception
of cumulative factors, the theoretical value is worth
considering, for it explains things which have been very
vaguely understood. The understanding of the fixation
of hybrids just mentioned, and races produced by artifi-
cial selection, clears up our practical breeding methods,
and this is valuable iiif ormation ; but the conception also
shows that the origin of species by natural selection as
announced by DARWIN is possible, a method which
for some time has been thought to be impossible.

Of course natural selection in a certain sense has
always been accepted, almost as generally as the
fact of evolution. The point in dispute is as follows:
DARWIN used as the basis of natural selection those
small individual variations which we have come to call
fluctuations, the same kind of variations the old plant
breeder used in his artificial selection. DARWIN claimed
that such variations could be piled up until the result
would be a new species. It was in 1900 that DE VRIES
(i) showed in a convincing way that this kind of varia-
tion never resulted in a new species; at best it could only
develop a race which approached the boundary of the
species and never crossed it. Moreover, such a race
would revert to type rapidly as soon as some slight
change in conditions set up a new standard for selection.
This argument, confirmed by experiment, has been
generally accepted.

We now know that individual variations are not
always mere fluctuations or responses, but may be due
to varying doses of cumulative factors. A selection on

Neo-Mendelism 85

this basis may well result in a new race that breeds
true; and a race that breeds true is DE VRIES' definition
of a new species. To reestablish DARWIN'S theory of
natural selection is certainly an important consideration,
and the situation illustrates how genetics and evolution
are tied up together, so that neither one of them can
be appreciated fully without some knowledge of the
other.

A few words may be said in reference to the reversion
of an old race to its original specific type. DE VRIES
outlined the situation clearly, and his conclusions are
generally accepted. It is doubtful, however, whether
it has ever been understood, since no one has ever
devised a reasonable mechanism for such a reversion.
The conception of cumulative factors, however, supplies
this mechanism. A new race, developed by natural or
artificial selection among individual differences, means
the piling up of cumulative factors in a given direction.
Stop the selection and the old plants with the small
number of factors are allowed to survive; reproduce,
cross with the new race, and eventually bring back the
old species to the original average condition.

1. DE VRIES, H., Species and varieties, their origin by mutation.
Chicago. 1905.

2. EMERSON, R. A., and EAST, E. M., The inheritance of quantita-
tive characters in maize. Bull. Agric. Exper. Sta. Ncbr. no. 2.
pp. 1 20. figs. 21. 1913.

CHAPTER IX

NEO-MENDELISM (Concluded)

Thus far we have based our considerations on the
rather simple theoretical mechanism of inheritance which
was explained in connection with Mendel's law. We
shall now consider some well established facts of inherit-
ance which will oblige us to enlarge somewhat our
theoretical mechanism.

(I) COUPLED AND ANTAGONISTIC CHARACTERS. The
classic illustration of coupled characters was brought
to light by EMERSON (i) during breeding experiments
with corn. His material included a strain with red
grains and red cob. This strain, when crossed with
another having white grains and white cob, gave an Fj
generation with red grains and red cob. This indicated
that red was dominant over white in both grain and cob;
and since the F 2 generation gave the orthodox 3 : i
ratio (3 reds and i white) the obvious conclusion was
that redness in both grain and cob is due to a single
determiner.

Other breeding experiments, however, gave a differ-
ent aspect to the situation. It was found that there
are races of corn with red grains and white cobs, and
others with white grains and red cobs. It is evident,
therefore, that redness in grain and cob is due to two
separate determiners.

Let us consider a possible theoretical explanation of
these two apparently contradictory situations. The red

86

Neo-Mendelism

grain and red cob of the first mentioned experiment
were not produced by a single determiner, but by
two determiners coupled together in inheritance. It is
as if the two chromosomes concerned had a bond between
them which kept them together during the reduction
division, and made them pass side by side into the
same gamete. A diagram will illustrate this situation
(fig. 20). The only unusual thing in this diagram is
the bond which holds together the chromosome carrying

Dominant Parent

with Red Seeds

and Cobs

Recessive Parent

with White Seeds

and Cobs

F, (Red Seeds and Cobs)

FIG. 20. Diagram representing, in a somewhat obvious way, how
the behavior of "coupled characters" might be visualized.

the determiner for red seeds and the chromosome
carrying the determiner for reel cob. A peculiar function
of the bond, however, appears in the reduction division
in connection with gamete formation by the F x genera-
tion.

In the case of a dihybrid in which two characters
are involved the situation developed by the reduction
division will be recalled (fig. 21). It will be noted
that there are four chances of pairing off by the four
scattered chromosomes. The chances of getting a
gamete with the two determiners is no greater than

88

Plant Genetics

getting one with only one of the determiners. There
would be just 25 per cent of each of the type represented
in the diagram, the resulting ratio in the F 2 being
9:3:3:1.

F, No Coupling 4 Possible Gametes

FIG. 21. Diagram showing how normal dihybrid segregation, in
absence of coupling, produces four different gametes.

In the case we are considering, however, the mecha-
nism is different. At the reduction division of the Fj
generation two chromosomes with the dominant de-
terminers are linked together by a bond. In conse-
quence of this there are only two types of gamete

F, Coupling

Only 2 Gametes
Possible

These 2 Gametes
Impossible, Due to
Presence of Bond

FIG. 22. Diagram showing how coupling would limit possible
kinds of gametes.

possible, as indicated in the diagram (fig. 22). In the
terminology of cytology the two chromosomes with
the determiners are held so firmly together that they
pass to the same pole of the spindle. This is the phe-
nomenon of c^j)ledinhjritance and a theoretical
mechanism that was devised to explain it.

Neo-Mendelism 89

EMERSON then crossed strains having red grains and
white cob with strains having white grains and red cob.
The F x generation showed all red grains and cobs, as
would be expected; but in the F 2 generation three differ-
ent types of individuals appeared as follows: i with red
grains and white cob; 2 with red grains and red cob;
i with white grains and red cob. This result in the F 2
generation differs from any that has been cited, and
EMERSON explains it as follows. In this case the two
determiners are not coupled; in fact they are antago-
nistic, so that no true-breeding types are produced
having red grains and red cobs. It is as if the two
chromosomes carrying the determiners are mutually
repulsive, so that they always pass to different poles of
the spindle at the reduction division. EMERSON inter-
prets the results of these experiments, therefore,, as due
to the existence of two coupled chromosome^ in the
first case and two antagonistic chromosomes in the
second case.

But is there not another way of looking at this
situation ? Granted that there is such a phenomenon
as the coupling of chromosomes, the question arises why
the second situation may not also be explained by coup-
ling, the difference being that in this case the coupling
is between different chromosomes. According to this
suggestion the two situations would be represented by
the diagram given in fig. 23. Such a scheme explains
both cases by using the same kind of force, that is,
coupling, and does not call for attraction to explain
one situation and repulsion to explain the other.

This last conception, however, raises the question,
why should a chromosome which in some strains of corn

Plant Genetics

is coupled with a similar one, in other strains become
coupled with its opposite (its allelomorph) ? This is a
theoretical difficulty which will be cleared up when we
consider the next topic. It will be realized that genetics
as yet represents an accumulating mass of data in

COUPLED CHARACTERS"

Only 2 Gametes

Possible
F 2 Gives 3 : 1 Ratio

"ANTAGONISTIC CHARACTERS"

Parents

Only 2 Gametes

Possible
F 2 Gives 1:2:1 Ratio

FIG. 23. Diagram showing how coupled and antagonistic char-
acters might be explained by the same mechanism.

reference to the facts of inheritance, and also an ac-
cumulating mass of speculations as to the explanations
of the facts. The facts are undoubted, while the specu-
lations suggest further experimental work.

(g) LlNKAGE_AND CROSSING OVER. EMERSON COn-

cluded his original work on coupled and antagonistic

Neo-Mendelism 9 1

characters with the following statement: "This is an
example of a feature which is probably very wide-
spread in the plant world, but of which at present we
know little. " Long before any further important work
was done along this line in the plant world, however,
MORGAN (2) published the results of his very careful
and intensive breeding experiments with the fruit fly.
His ideas have had a very profound influence upon
subsequent work in genetics. He has given us a more
accurate picture of the hereditary mechanism and one
that fits the facts better than any previously proposed.
In simplest terms the picture is this. Each chromosome
is a rodlike structure and numerous determiners are
arranged in a line along this rod.

We cannot discuss here the many ways in which this
fundamental conception has cast light upon work in
genetics. Suffice it to say that it has resulted in a
new "school." of geneticists whose experiments have
been more intensive, more exact, and perhaps more
"fundamental" than those of any other school. The
work so far has been done mainly with the fruit fly
and is of a rather complicated nature. We will present
here some of the simplest underlying ideas merely to
show the nature of MORGAN'S hereditary mechanism.
It will be seen how such a mechanism may explain such
phenomena as the coupled and antagonistic characters
in corn. This general situation, however, will now be
referred to by the term linkage, which is in more com-
mon use.

When first considering Mendel's law the statement
was made that mpre than one determiner might be
located on a given chromosome. As yet we have

92 Plant Genetics

considered no such case, but linkage involves this situa-
tion. In connection with some of MENDEL'S original
crosses fig. 24 will be recalled. In this case a double
dominant mates with a double recessive, and the result
is a dihybrid ratio in the F 2 generation. Suppose,
however, that determiner T and determiner S are on the
same chromosome, the situation would be as represented

Dwarf Wrinkled
Parent

4 Possible Gametes
/.F 2 Shows 9:3:3:1 Ratio

FIG. 24. Diagram showing normal dihybrid behavior when no.
linkage is involved.

in fig. 25. This is linkage; that is, with two determiners
a monohybrid ratio instead of a dihybrid ratio is
obtained. MORGAN'S definition is as follows: "When
factors lie in different chromosomes they give the
Mendelian expectations; but when factors lie in the
same chromosome they may be said to be linked, and
they give departures from the Mendelian ratios." It
will be noted that this conception rests upon the belief
that the chromosome is the indivisible unit in inheritance,
a conception that seems to have been justified by most
of the breeding results and which conforms to Mendelian
inheritance.

Neo-Mendelism

93

On the other hand, facts have begun to appear which
seem contrary to this view. When the idea of linkage
began to be developed the breeding expectations were
modified in accordance with it. For example, when it
is discovered that tall individuals always have smooth
seeds this fact is explained as linkage. The inference is
that there never can appear a tall plant with wrinkled

LINKAGE"

Only 2 Gametes

Possible

.'. F 2 Shows 3 : 1 Ratio
Dwarf Wrinkled (3 Tall Smooth : 1 Dwarf Wrinkled)

Parent

FIG. 25. Diagram showing how linkage would limit the possible
kinds of gametes. This mechanism will explain simply and accurately
the "coupling " and "antagonism " of the previous diagrams; this mental
picture of chromosomes, suggested by EMERSON and developed fully by
MORGAN, is now generally accepted by geneticists.

seeds or a dwarf plant with smooth seeds, for if there is
linkage and the chromosome is the indivisible unit ot
inheritance it would be impossible for tallness and
smoothness to become separated.

This was the situation a few years ago, when suddenly
there appeared what corresponds to tall individuals
with wrinkled seeds and dwarf individuals with smooth
seeds. These unusual types appear rather rarely as a
general rule, although in a few special cases they have
appeared as frequently as 20 in 100. The work was done
with Drosophila (the fruit fly) under such conditions

94 Plant Genetics

of control that there could have been no experimental
error. This new fact demanded explanation, for with
such chromosomes as TS and ts it would be impossible
to obtain a tall wrinkled individual as long as the
individuality of the chromosome is maintained. When
chromosomes were examined with the modern lenses
they were found to show all kinds of tangled contortions
during reduction division, and accordingly the scheme
shown in fig. 26 was devised. These five stages repre-
sent phases that allelomorph chromosomes may go
through during reduction division. Two allelomorph
chromosomes, which normally would be side by side
(i) may at times come to lie across one another (2).
In this position the middle regions of the chromosomes
are in contact and are conceived of as fusing (3). The
spindle fibers from each pole then lay hold of this com-
pound chromosome and the pull comes in the direction
of the arrows shown in the figure. This results in the
break indicated in 4. Finally, two new chromosomes
separate from the old compound chromosome, as indi-
cated in 5. Thus T is linked with s, and later, when
two such chromosomes are brought together in crossing,
the result is a tall wrinkled individual. In the same
way dwarf smooth individuals may appear.

This is known as crossing over, and the literature in
reference to it is extensive. It is on the basis of this
conception that certain geneticists claim to be able not
only to locate given determiners on their proper chro-
mosomes, but also to tell exactly on what parts of the
chromosomes these determiners are located.

This mechanism readily explains the phenomenon of
coupled and antagonistic characters without involving

Neo-Mendelism

95

such clumsy bonds and repulsions between the chromo-
somes as were presented in the previous scheme.

FIG. 26. Illustrating how crossing over may occur

g6 Plant Genetics

As stated before, most of the work on linkage and
crossing over has been done with the fruit fly. The
same technique, however, is now being applied among
plants. In a recent article WHITE (3) discusses four
linkage groups in Pisum, and considerable data on
linkage and crossing over in corn are now being assembled
at Cornell under the able direction of EMERSON.

Perhaps a warning is needed as to the term correla-
tion in this connection. Coupling and linkage are
phenomena of inheritance, involving the reduction
division, the gametes, and the zygote. Correlation is
a physiological phenomenon that appears in the develop-
ing plant. A tall corn plant may produce large ears
merely because the plant is tall and regardless of the
determiners for ear size. This is correlation, and it
appeared, not in the gametes and zygote, but in the
physiology of the individual after it started to develop.
A correlation, therefore, may sometimes deceive as to
inheritance.

1. EMERSON, R. A., see Heredity and eugenics (COULTER, CASTLE,
DAVENPORT, EAST, and TOWER). Chicago. 1912 (pp. 101-
104).

2. MORGAN, T. H., Heredity and sex. New York. 1914.

3. WHITE, ORLAND E., Inheritance studies in Pisum. IV.
Interrelation of the genetic factors in Pisum. Jour. Agric.
Research n : 167-190. 1917.

CHAPTER X
NON-MENDELIAN INHERITANCE

Up to this point Menclelism and its various modi-
fications have been considered, and the impression to be
gained thus far is as follows. Mendel's law as originally
stated is clear, reasonable, and well established. There
have appeared certain types of inheritance which have
seemed at first sight to contradict Mendel's law; but
practically all of these later have been shown to be con-
sistent with that law. The conclusion, therefore, may
be that Mendel's law is firmly established but that its
expression is not always so simple as was once sup-
posed. Illustrations have been given showing how
Mcndelian explanations must be based frequently upon
numerous factors of various types. The impression,
therefore, may be that Mendelism is so established and
its application so universal that it explains every type of
inheritance that is known or that ever will be known.
It is true that Mendel's law is dominant with most
geneticists as explaining most of the known facts of
inheritance, although there are certain types of inherit-
ance that have not been explained as yet quite satis-
factorily in this way. As a consequence, a number
of prominent geneticists protest against the universality,
of the law. These unexplained phenomena must now
be considered under what may be called non-Mendelian
inheritance. It will be recalled that EMERSON and
EAST stated that only two indisputable cases of non-
97

98 Plant Genetics

Mendelizing characters arc known, and these cases may
be used as illustrations of non-Mendelian inheritance.

In connection with his work on Mirabilis CORRENS (2)
uncovered the following situation. The ordinary race of
Mirabilis has pure green leaves, but one was discovered
with variegated leaves. The leaves were green for the
most part but showed irregular white patches, an
examination showing that in the white areas the chloro-
plasts were more or less bleached out. CORRENS
named this race the albomaculata type, a name which
appears frequently in the literature of genetics.

The behavior in this type of inheritance was as
follows. When self-fertilized it bred true, but the pollen
from albomaculata, when used in crosses, behaved as if
it had come from normal green plants. For example,
CORRENS took pollen from an albomaculata plant and
used it to pollinate a normal green plant. The indi-
viduals of the F! generation were all normal green,
and the natural Mendelian conclusion would be that
albomaculata is a recessive condition appearing when
the determiner for normal green leaves is lacking. But
when the F x generation was inbred the progeny con-
tinued to be normal green generation after generation.
If albomaculata pollen had introduced a recessive
character into the hybrid, this would have reappeared
in the later generations, but it did not reappear. ; The
obvious conclusion was that albomaculata pollen is just
like the pollen of the normal green plants and that all
pollen carried the determiner for the development of
normal green plants.

This raised the question as to the result from using
pollen of normal green plants to pollinate albomaculata.

Non-Mendelian Inheritance 99

It is obvious that the F! generation should all be normal
green and albomaculata would appear only in later
generations. CORRENS made this cross, using normal
green pollen and albomaculata ovules, and secured
results that seemed startling. All of the individuals
of the Fj generation were albomaculata and all the later
hybrid generations were albomaculata. To CORRENS
there seemed to be only one conclusion possible, and
that was that the pollen did not affect anything one
way or the other, and that inheritance was all on the
maternal side. When a possible mechanism to explain
such a phenomenon is considered it is obvious that the
situation cannot be explained on the basis of Mendel's
law. The explanation suggested by CORRENS is ingen-
ious and seems fairly reasonable.

He suggests that the albomaculata character is due
to a disease of the cytoplasm which does not affect the
nucleus. The nuclei in these albomaculata plants are
supposed to carry factors for chlorophyll formation, just
like those of the normal green plants, but the cytoplasm
about the nucleus is diseased. Thus the disease is passed
on at cell division simply because the cytoplasm is
divided and passed on. In fertilization, however, no
cytoplasm enters the egg with the male gamete, but only
a nucleus; therefore it would be impossible for the dis-
ease to be transmitted by the pollen, and an albomaculata
pollen would have no effect in inheritance. The only
chance to transmit the disease would be through an
egg with diseased cytoplasm. In inheritance, there-;
fore, the albomaculata character is handed down
only by the female parent, and this is maternal
inheritance.

ioo Plant Genetics

This would explain the peculiar breeding results
obtained by CORRENS. If this explanation is true
Mendel's law is not contradicted , for such a phenomenon
is entirely outside the field of that law. In fact, this
should perhaps not be regarded as inheritance at all
but as a case of reinfection. The albomaculata condi-
tion is not a true plant character; it is a pathological
state, not inherited as such, but the bacteria in the
cytoplasm of the egg reinfect the next generation.

The conclusion therefore is that MendePs law is not
contradicted by this phenomenon, but it introduces
entirely new possibilities of inheritance quite outside the
scope of Mendel's law. This conclusion is reached if
one accepts CORRENS' explanation, but this might well
be doubted. To state that albomaculata cytoplasm is
always infected and that the nuclei are always immune
is a bold assumption. If we assume, with some cytolo-
gists, that the nucleus consists entirely of chromatin and
that chromatin is immune to this disease this would
furnish the proper mechanism, but this is questionable.
Again, when it is claimed that no bacteria could enter
the egg with the male nucleus another rash assumption is
made. After all, the case of albomaculata may still
be regarded as a genuine case of non-Mendelian inherit-
ance of the type known as maternal inheritance.

The other example of non-Mendelian inheritance to
which EMERSON and EAST referred is much more serious,
introducing somatic segregation, which is a very far-
teaching topic. When EMERSON and EAST made their
statement somatic segregation had been rarely observed,
there being really only one authentic case on record,
that is, the case of the common geranium (Pelargonium

Non-Mendelian Inheritance 101

zonale) described by BAUER (i). Since that time,
however, many more cases have been discovered, and
the fact of somatic segregation has become fairly well
established.

An illustration of the situation may be given as
follows. If a white-leaved plant and a normal green-
leaved plant are crossed the resulting hybrid illustrates
what has been mentioned as "particulate inheritance/'
that is, the hybrid is variegated, showing irregular
patches of white and green. If one of these white
patches completely includes a bud there will probably
be produced by that bud a completely white branch.
The flowers of this branch, when self-fertilized, give rise
through their seeds to white individuals only and would
evidently continue to breed true to the white condition
if white individuals could be matured. In like manner
the variegated hybrid may give rise to a pure green
branch, which would start a line of pure green indi-
viduals.

This is an Illustration of what is called " somatic
segregation." Ordinarily, of course, factors are segre-
gated only at the reduction division, and only in that
division does the proper cytological mechanism for seg-
regation exist. Somatic segregation, however, means
that segregation takes place in the somatic tissue, quite
apart from the usual reduction division. Cytologists
assure us that cell divisions in the somatic tissue are not
reduction divisions, in this tissue chromosomes being
regularly reproduced in equal numbers. Each chromo-
some divides to form two chromosomes, and these two
are similar to the mother and to one another; therefore
they carry exactly the same quota of determiners, that is,

102 Plant Genetics

there is no segregation of determiners. This situation is
not merely shown by cytology, but is also in accord
with most of our breeding results.

In the case of somatic segregation, therefore, there
seems to be only one possible conclusion, and that is that
sometimes reduction^ division occurs in the somatic
tissue, at least so far as certain chromosomes are con-
cerned. Such a conclusion is not unreasonable. Crosses
of certain factors might cause a very unstable condition in
certain chromosomes, so that they might not conform
to the ordinary mechanism of cell division. Certain
chromosomes, or certain determiners on the chromo-
somes, instead of dividing normally might pass un-
divided to one pole of the spindle at cell division and
be entirely lacking at the other pole.

The only other alternative would be to discard the
old ideas of the mechanism of cell division and inherit-
ance, and this we are hardly ready to do. In any
event the phenomenon of somatic segregation opens
an important field for some critical cytological research.

It will be realized that somatic segregation is a
much more fundamental situation to explain than
maternal inheritance. There is no reason to suppose,
however, that it contradicts Mendelian inheritance
seriously; perhaps it enlarges the scope of Mendelism.
Mendel's law is violated only in the fact that segregation
occurs at an unusual place; and that is no reason for
discarding Mendelism. It complicates the situation
somewhat because the ordinary reduction division
furnishes such a complete mechanism for segregation;
but MENDEL formulated his law before the reduction
division was discovered in plants. Somatic segrega-

Non-Mendelian Inheritance 103

tion may lead to the discovery of an occasional unusual
kind of reduction division in the somatic tissue.

1. BAUER, ERWIN, Das Wesen und die Erblichkeitsverhaltnissc cler
"Varietates almomarginatae hort." Von Pelargonium zonale.
Zeitschr. Ind. Abstamm. Vererb. 1 1330. 1909.

2. CORRENS, CARL E., Ubcr Bastardierungsvcrsuche mit Mirabilis-
Sippen. Bcr. Dcutsch. Bot. Gesell. 20:594. 1902; ibid.,
Vererbungsversuche mit blass (gelb) griinen und bltmtblat-
terigen Sippcn bei Mirabilis jalapa, Urtica pilulifcra, u. Lunaria
annua, Zeitschr. Ind. Abstamm. Vererb. 1:291. 1909.

CHAPTER XI
MODIFICATION OF UNIT CHARACTERS

The severest blow at Mendelism is the modification
of unit characters. The three theses involved in
Mendel's law are unit characters, dominance, and segre-
gation. It has been seen how dominance may fail with-
out affecting the essential feature of Mendel's law. We
have just considered how segregation might occur else-
where than at the reduction division and yet not inter-
fere seriously with Mendel's law. Unit characters,
however, cannot be treated in this way, for they are
the very foundation of Mendel's law. In the very
nature of things they remain as units, separate in
inheritance ; if they are observed splitting up or varying
in any way they would no longer be units, and the
foundation of Mendelism would be weakened.

Before taking up the real cases of modification of
unit characters, another unusual phenomenon concern-
ing unit characters may be considered first, a phenome-
non which is quite in accord with Mendelism as explained
today. It is introduced here because at first thought
it might seem to strike at the essential nature of unit
characters; in fact, it has been so used.

CASTLE (i), working with fancy mice, found that on
rare occasions a unit character might disappear com-
pletely in cross-breeding. It might be claimed that this
result is due to the introduction of an inhibiting factor,
or to the separation of complementary factors. CASTLE

104

Modification of Unit Characters 105

realized these possibilities and tested them. Of course
if it was a case of complementary factors which had
become separated he would have been able to recombine
them in later generations by the proper crosses. Recog-
nizing the significant crosses and making them he ob-
tained no result; the character did not reappear. In
the same way breeding tests for an inhibiting factor
which could be separated out in later generations yielded
no results. The unit character was gone beyond recov-
ery; it had simply dropped out and was lost.

Such an occurrence may seem surprising at first, but
it is what should be expected as an occasional occur-
rence. Mendel's law was republished along with DE
VRIES' mutation theory, and as a consequence scientific
breeders, interested in Mendel's law, were also on the
watch for mutations. DE VRIES had shown that muta-
tions might involve either the appearance of an entirely
new character or the dropping out of an old character.
It is this situation, therefore, that appeared in CASTLE'S
experiments with mice, namely, the abrupt disappear-
ance of a unit character is simply the result of mutation,
and this involves no violation of Mendel's law.

The real attack upon Mendelism involves a different
kind of behavior on the part of unit characters. The
notable example of this unexpected behavior of unit
characters was discovered in connection with CASTLE'S
work on hooded rats (2).

He isolated from his rat populations a certain strain
which showed a particular black and white coat pattern.
This type was then inbred for a number of generations
to insure that it bred true. This fact having been
established it was next determined that this black and

io6 Plant Genetics

white coat pattern behaved as a simple unit character
in inheritance. Then, starting with a single pair of
rats of this new pure strain, the following breeding
experiments were performed. For twelve generations
selections were made from this new strain without a
single outcross, that is, every generation was inbred,
thus insuring the constant purity of the stock. In one
series selection was made for an increase in the extent
of the pigmented areas; in the other series selection
was made for a decrease in the extent of these areas.
The result was that the areas in the one series steadily
increased while in the other they steadily decreased.
Thus far nothing very unusual is involved. CASTLE
points out, however, the following important facts which
were developed: (i) with each selection the amount of
regression (" running back") grew less; that is, the
effects of selection became more permanent; in other
words, in each succeeding generation there was a
decreasing tendency to revert to the original average
type; (2) advance in the upper limit of variation was
attended by a like advance of the lower limit. The
total range of variation, therefore, was not materially
changed, but there was a progressive change in the
point about which the variation occurred. In other
words, it is like a progressive shifting of the center of a
circle ; the diameter of the circle does not change but the
position of the circle, determined of course by its center,
is gradually changing. These were the two important
facts which CASTLE brought out and they have been
stated approximately in CASTLE'S words.

Fig. 27 will help to make the situation plain. The
average amount of variation in any one generation of

Modification of Unit Characters

107

the pure stock (the diameter of the circle referred to) is

A
indicated by < > . Of course, even "pure

stock" varies somewhat, since no two individuals are

> " M> ' Pure-Bred
Generation*

The Control

FIG. 2 7. --Diagram illustrating CASTLE'S selection experiment
with hooded rats.

exactly alike, biology recognizing what is called " indi-
viduality/' The point, however, is that the compara-
tively small variation in a pure stock is not due to
germinal differences but to responses called out by
varying external conditions, such as nutrition, light, etc.

io8 Plant Genetics

These response variations, usually called fluctuations,
vary with different individuals, but the hereditary
capacity of all of them remains the same. A selection on
the basis of fluctuations within a pure line, therefore,
should not result in any permanent improvement; in
fact, it has been demonstrated many times that no such
improvement can be effected in this way. When selec-
tion is made, however, among varying doses of cumula-
tive factors, an entirely different situation is faced, for
in such a case we are not dealing with a pure line.

The significance, therefore, of CASTLE'S results may
be realized. He bred his original pure line for many
generations and found that it varied only within very
narrow limits; and these slight variations therefore he
regarded as mere fluctuations. Furthermore, he found
that the character of his pure line developed in crossing
as a simple unit character and that no complex factors
were involved. With this evidence he should not have
been able to effect any permanent changes by selection;
but this is exactly what he did. Selecting in opposite
directions he developed two new strains, the boundaries
of the new strains being distinct from one another and
distinct from the boundaries of the original strain, that
is, the non-selected pure type.

CASTLE'S next step was significant. He crossed
each of his new strains with the same wild race, the
result being that each of his new strains behaved as a
simple recessive unit, giving a 3 : i ratio among the
grandchildren.

The logical conclusion from this series of experiments
may be given in CASTLE'S words, as follows: "The
conclusion seems to me unavoidable that in this case

Modification of Unit Characters 109

selection has modified steadily and permanently a
character unmistakably behaving, as a simple Mendelian
unit." The importance of this conclusion is evident.
Mendelism has been based upon the conception that
unit characters cannot be modified. Mendelians have
granted only two possible methods for the origin of new
races: (i) by recombinations of existing characters by
hybridizing; (2) by the sudden and complete dropping
out of an existing unit or the equally sudden addition of
a new unit, both of which possibilities may arise from
mutation. No Mendelians will grant, however, the
possibility of modifying an existing unit character, the
thing which CASTLE claims to have done, and bases his
claim upon well controlled experimental breeding. If
CASTLE'S contention is true it must result in the fun-
damental modification of Mendel's law. The whole
mechanism will have to be modified or new fields of
variation not known to exist will have to be taken into
account.

The statements of the Mendelians in reference to this
situation should be considered. Their explanation is
based in effect upon the situation we have already
developed in connection with cumulative factors. The
claim is made that CASTLE started with a peculiar
character which fluctuates continually and has never
been brought to as small a variability as have most
other characters. The question is raised whether
CASTLE'S assumption that this variability is merely due
to fluctuation is altogether justified. May not the
variability be due to varying closes of cumulative
factors? Suppose for the moment that this were the
case; it would not be surprising that CASTLE could

no Plant Genetics

develop two diverse strains by selection, for selection
would result in piling up the cumulative factors in one
direction or another. CASTLE'S rejoinder would be that
if this is a cumulative factor situation why do none of
the extremes appear in the non-selected stock, which
instead breeds approximately true within very narrow
limits of variability? The answer is made that the
extremes do not appear in the pure-bred stock merely
because of mathematical possibility. If we are dealing
with six cumulative factors, and the so-called pure
stock has an intermediate number of doses, there
could not be much chance of getting out the extremes
in later generations. It will be remembered that it
would be necessary to secure over 4000 progeny to
have an even chance of getting one such extreme; or
at least 50 progeny to get anything that would visi-
bly approach the extreme. It would seem, therefore,
that CASTLE'S chances to determine this would be very
small. Rats certainly do not produce 4000 progeny
in a single generation; in fact, they produce much less
than 50; therefore CASTLE'S pure stock goes on in the
intermediate condition, and only by selection can he
pile up the factors and reach either extreme.

Thus far the explanation seems satisfactory. CASTLE
showed, however, that the coat pattern condition
behaved in crosses as a simple Mendelian unit; that is,
it did not split up into complex ratios, but came out as
a recessive in a regularly 3 : i ratio. This really involves
no difficulty. Suppose CASTLE crosses one of his pure
strain rats having the pattern character with another
strain having an inhibitory factor for the pattern or some
other character that conceals the pattern. Since the

Modification of Unit Characters in

inhibiting factor is simple, the resulting ratio may be a
monohybrid ratio; that is, in the F 2 generation from
such a cross the ratio of pattern to non-pattern will be
1:3, and this is exactly what CASTLE got and what would
be expected. At the same time the amount of pigmenta-
tion, determined by numerous cumulative factors, goes
on in the, same intermediate condition, unaffected by the
cross. The relation of pattern to non-pattern is merely
a simple monohybrid system temporarily superimposed
upon the other more complex system without perma-
nently affecting it, any more than any inhibitory factor
permanently affects the factors it inhibits, or a dominant
permanently affects a recessive.

It is in this way that Mendelians can explain away
CASTLE'S results. CASTLE does not admit the justice
of the explanation, but continues to maintain that he has
modified a unit character by selection, and some geneti-
cists agree with him. Whether CASTLE is dealing with
cumulative factors or not can never be settled by exact
demonstration, since the progeny of rats is too small.
It is rather significant, however, that in plants, where
larger progenies are involved, the cumulative factor
hypothesis is well established, and no one claims to have
modified a Mendelian unit in plants. The best that can
be said concerning CASTLE'S claim is that it cannot be
proved false, but that it is possible to explain his results
by the cumulative factor hypothesis.

The question might be raised, however, why cling
so strongly to the cumulative factor hypothesis and
force CASTLE'S results into this conception? Is there
anything sacred about a unit character that it should
not be modified just as complex chemical molecules may

H2 Plant Genetics

be modified in certain reactions? Why not admit
that Mendelian factors may be modified, and explain
CASTLE'S results in this way? The reason is that when
we begin to admit that unit characters and single
Mendelian factors may be modified, the whole conception
of inheritance becomes chaos. The great advantage of
the factor hypothesis is that it furnishes the clearest
method of describing breeding results. A statement
by EAST (3) on this point is pertinent.

Taking into consideration all the facts, no one can well deny
that they are well described by terminology which requires
hypothetical segregating units, as represented by the term
" factors." What then is the object of having the units vary at
will ? There is then no value to the unit, the unit itself being only
an assumption. It is the expressed character that is seen to vary;
and if one can describe these facts by the use of hypothetical
units, theoretically fixed, but influenced by the environment and
by other units, simplicity of description is gained. If, however,
one creates a hypothetical unit by which to describe phenomena,
and this unit varies, he really has no basis for description.

In other words, this statement means that the great
value of the factor hypothesis is to supply a terminology
or describing the facts of inheritance rather than a
means of explaining these facts. According to this,
therefore, it may be seen that although CASTLE claims
to have modified a unit character by selection, his
results can come under the factor hypothesis, and thus
our descriptions may be kept clear and uniform.

There is another piece of work, however, that cer-
tainly cannot be explained (that is, described) in terms
of the factor hypothesis; on the contrary, it almost forces
one to admit that unit characters, even though hypo-
thetical, may vary. This is the work of JENNINGS (4)

Modification of Unit Characters 113

on Protozoa. JENNINGS, for example, takes a single rhi-
zopod (Difflugia) and uses it as the basis of a new strain.
Every member of this new strain has come from this
original individual by cell division. In this case sex
is not involved, and the possibility of new combina-
tions of factors is eliminated. It is obvious that such
a strain should breed perfectly true, but JENNINGS
shows that it does not. Changes of two kinds occur,
namely, mutations, which of course are a part of the
Mendelian program, but according to JENNINGS "the
overwhelming majority of hereditary variations are
minute gradations. Variation is as continuous as can
be detected." He gradually piles up these minute
variations by selection and finally develops a new
species. There can be no claim of piling up cumulative
factors here, for no sex is involved nor can these gradual
changes be mutations. Mutations of course can be
minute, but they are not continuous and in the same
direction; mutations are jumps, even if the jumps are
small, and they occur in every direction.

The only other possibility is as follows. It will be
remembered that WEISSMAN claimed that environment
may directly affect the germ plasm in the simplest micro-
organisms, just such organisms as JENNINGS deals with;
but JENNINGS claims that the characters of his rhizopods
are not modified by environment. How he is assured
of this is a question, for the factors of environment are
very numerous and complex and organisms are probably
very sensitive. So far as circumstances permit, the
conditions of environment were kept constant, but
whether they really were constant or not is another
matter. Leaving out of consideration this question of

ii4 Plant Genetics

environment, it must be admitted that JENNINGS'
work proves that unit characters may be modified by
selection.

The evidence just given as to the modification of
unit characters seems rather conflicting. It will be
realized, however, that there are at present two leading
schools of geneticists, one believing that unit characters
may be modified by selection, and the other believing
that unit characters are invariable. This latter school
we have been referring to as Mendelians, but this is
hardly an accurate designation, since both schools are
really Mendelians. They are usually distinguished as
follows: "mutationists/' who believe in the introduction
of new hereditary units by mutation alone and believe
that unit characters cannot be modified by selection;
and " selectionists," who believe that unit - characters
may be modified by selection. In the previous dis-
cussions the views of the mutationists have been em-
phasized for two reasons: (i) mutationists are in the
majority at present; (2) their hypothesis gives a com-
prehensive and systematic basis for describing the facts
of inheritance. JENNINGS has recently tried to reconcile
the two schools by showing that their claims amount
to the same thing. Such a reconciliation is hopeful but
it may prove to be impossible.

1. CASTLE, W. E., Heredity in relation to evolution and animal
breeding. New York. 1911.

2. , The inconstancy of unit characters. Amer. Nat.

46:352-362. ipi2.

3. EAST, E. M., The Mendelian notation as descriptive of physio-
logical facts. Amer. Nat. 46:633-655. 1912.

4. JENNINGS, H. S., Observed changes in hereditary characters in
relation to evolution. Jour. Wash. Acad. Sci. 7: 281-301. 1917.

CHAPTER XII
PARTHENOGENESIS AND VEGETATIVE APOGAMY

Parthenogenesis and vegetative apogamy are very
important to keep in mind in connection with work in
genetics. Geneticists have paid too little attention to
these phenomena, and as a consequence their data are
not always reliable. The distinction between the two
terms is as follows. Parthenogenesis is the develop-
ment of an embryo by an unfertilized egg, while vege-
tative apogamy is the development of an embryo by
another cell of the female gamctophyte. The term
" apogamy" includes both, meaning the production of a
sporophyte by a gametophyte without involving an
act of fertilization. Parthenogenesis is not peculiar to
any region of the plant kingdom, and a few illustrations
will be sufficient.

In water molds (Saprolegnia) parthenogenesis is very
common. All stages in the abortion of the male organ
(antheridium) are found and still eggs apparently
germinate as freely as if fertilized. In fact, in many
species of Saprolegnia it is a question whether fertiliza-
tion ever occurs.

Ferns are notably apogamous, any cell of the game-
tophyte (including unfertilized eggs) producing an
embryo under certain conditions. This miscellaneous
production of embryos by the gametophyte is an
impressive illustration of the fact that any vigorous
protoplast, under appropriate conditions, can produce

n6 Plant Genetics

a new individual. This behavior of ferns is particularly
interesting on account of the chromosome situation
involved. Since the gametophyte is an x structure and
the sporophyte a 2x structure, when apogamy was first
discovered it was taken for granted that in some way
nuclear fusion had occurred and the 2X number obtained.
More recent work, however, notably that of YAMA-
NOUCHI (2), showed that the apogamous sporophyte is
an x structure, and such a sporophyte produces spores
without the reduction division.

Ferns are also aposporous, which means the formation
of gametophytes by vegetative cells of the sporophyte,
the gametophyte in such a case being a 2X structure.
With the phenomena of apogamy and apospory estab-
lished, the question naturally rises as to their bear-
ing upon the machinery of Mendelism, involving the
reduction division and the segregation of hereditary
factors.

Seed plants, however, are of particular interest, since
they have furnished the chief material for' work in
genetics. Many cases of parthenogenesis and vegetative
apogamy in seed plants have come to light. It is an
interesting fact that, although these cases are well dis-
tributed throughout the groups of seed plants, they seem
to be especially numerous among the Compositae. In
many cases, therefore, the embryos of seed plants are
produced by unfertilized eggs or other cells, as in Anten-
naria, Hieracium, Taraxacum, Tkalictrum, etc. Some
of the best known cases were subjected to cytological
investigation, and it was found that in most cases the
parthenogenetic egg is diploid (2^), thus containing the
sporophyte rather than the gametophyte number of

Parthenogenesis and Vegetative Apogamy 117

chromosomes. It is evident that for some reason the
reduction division had failed in connection with spore
formation, so that the spores were diploid, resulting
naturally in a diploid gametophyte. In these cases it
was found that no reduction division occurred during
the life- history, the diploid number continuing through
both sporophyte and gametophyte. The interesting
observation was made also that parthenogenetic eggs
are slower in starting to germinate than those normally
fertilized.

Polyembryony is also common among seed plants,
and this involves apogamy. Mention has been made
of the case of an onion seed containing five embryos,
only one of which could have come from a fertilized egg.
The four possible sources of embryos are found to be the
fertilized egg, the unfertilized egg, gametophyte cells,
and sporophyte cells.

Another interesting fact concerning parthenogenesis
in seed plants has been added comparatively recently (i).
A plant known to be parthenogenetic was shown to have
2x eggs, but it was also known to be pollinated and its
sperms were seen discharged into the embryo sac. These
sperms were apparently functional in every particular,
and with the x number of chromosomes, and yet,
although a sperm came into contact with an egg, there
was no fusion. The inference has been that a diploid
egg may be incapable of fertilization and is partheno-
genetic if it functions at all. It does not follow that only
diploid eggs are parthenogenetic, for parthenogenesis
in ferns contradicts this conclusion.

The bearing of these facts upon genetics should be
considered. Plant geneticists have been using the most

n8 Plant Genetics

complex material. Results have been given in experi-
mental work with peas, four-o'clocks, wheat, and, most
of all, corn, all of them seed plants. This selection of
material was natural, for such plants represent the
region of the plant kingdom in which practical breeding
occurs. It was natural to make practical breeding
scientific.

Not only have geneticists used the most complex
material but they have selected sexual reproduction as
the method of reproduction to be investigated, which is
the most complicated kind of reproduction. This selec-
tion was also natural, for Mendel's law is based upon
sexual reproduction, and it is this law that geneticists
have been testing.

The further difficulties of the situation should also
be considered in order that one may be in a position to
estimate the value of results. Students of morphology
are familiar with the fact that the sexual structures of
seed plants are peculiarly involved with other structures.
They do not stand out distinctly, as in the algae, for
example. Not only are the sexual structures (eggs and
sperms) beyond the reach of observation, and therefore
of experimental control, but there is the alternation of
generations to consider, inheritance being carried through
one generation to express itself in the next. A sporo-
phyte does not beget a sporophyte but a gametophyte,
and this in turn begets the embryo sporophyte.

Add to this complex situation the possibility of
parthenogenesis and vegetative apogamy and it is ob-
vious that the origin of embryos found in seeds is not
assured. If two embryos appear it is evident that at
least one of them holds no relation to the pollen parent ;

Parthenogenesis and Vegetative Apogamy 119

and even when only one appears, as is usually the case,
how can it be certain that it has arisen from an act
of fertilization? One can assume for the most part
that embryos are the result of fertilization, but the
increasing number of cases of apogamy introduces a
serious element of uncertainty. If two individuals
have been crossed, and the expected results in the F r
or F 2 generations do not appear, how can it be told
whether the pollen parent has entered into the result
at all? Doubtless puzzling results in genetics have
arisen more than once through apogamy; the crossing
has been controlled, but the results are taking place out
of sight and of course out of control. The crossing is
only the gross part of the performance. If instead of
applying pollen to stigma one could apply sperm to egg,
there would be far more control. The program of events
between pollination and fertilization should be kept in
mind, and also the program between fertilization and
the escape of the embryo from the seed. Not a single
stage of these performances is under observation, much
less under control. Pollination is effected, and then the
plant emerging from the seed is observed. One must
take for granted that all of the numerous events that
lie between have been completely orthodox. Usually
they are, but sometimes they are not; and then the
unexpected result is forced into the orthodox program.
It is really surprising, with so much of our experi-
mental* work out of sight and beyond control, that so
many constant results are obtained. It is an impres-
sive lesson as to the wonderful uniformity of conditions
between pollination and fertilization, and then between
fertilization and embryo escape. This gives assurance

I2o Plant Genetics

that, in spite of the complexity of material and in spite
of lack of control, the facts of inheritance are being
uncovered when the consistent results are considered.
It also assures us that an unexpected result may occur
now and then and that it is not necessary to explain it
by trying to make it consistent with other results. A
single illustration will make this clear. \ Suppose that a
species is persistently parthenogenetic, as some are
known to be, and still there is pollination. If one were
conducting breeding experiments with such a plant he
would discover that inheritance was of a sort called
maternal inheritance, that is, inheritance determined
by the ovule parent and not affected by the pollen
parent., In explaining this situation it would not be
necessary to have recourse to CORRENS' idea of diseased
cytoplasm ; one would suspect parthenogenesis and then
look for it.

Self-sterility, which is an important and growing
subject, to be developed later, is related in one way or
another to apogamy. Mention has been made of the
stimulating effect of the mere presence of a pollen tube
in inducing an unfertilized egg to germinate without
sperms being discharged or even formed. An increas-
ing number of plants are found to behave in this way,
especially in certain families. Notably in some culti-
vated races of apples and grapes self-sterility has been
developed, that is, the pollen grains are impotent upon
their own stigmas. In such cases the fruit develops
just as if fertilization had taken place; in many cases
seeds develop within the fruit and even embryos in the
seeds. Strangely enough if " foreign" pollen is used
it may not be impotent but results in fertilization. All

Parthenogenesis and Vegetative Apogamy 121

of this indicates that the origin of embryos cannot always
be assured, and apogamy is a factor that may at any
time upset any expectation based on Mendel's law
without violating that law. If Mendel's law is based
on the act of fertilization, and that act does not occur,
the result has nothing to do with the law.

In view of what has been said, the recent movement
of some plant geneticists to include the lower plants in
their work will be understood. This results in the
following opportunities. In the study of sexual repro-
duction the simpler plants (such as algae) have sexual
structures that are not involved with other structures,
and the whole performance of fertilization and embryo
development is in sight and capable of control. The
difference between a sex act and embryo development
under cover and in the open, when observation and
control are desired, is obvious. Furthermore, back of
inheritance through the sex act is inheritance through
spores. Such inheritance needs investigation in con-
nection with inheritance through sex. There are cer-
tain things that all forms of inheritance have in
common, and these things should be kept distinct from
those which are peculiar to sex inheritance. Only in
this way can any knowledge be reached of the special
role of sex in inheritance and the peculiar features it
has added.

A practical plant breeder may be interested only in
the fact that he can obtain a new individual from a seed
the pedigree of whose embryo in the nature of things
cannot be demonstrated; but a scientific plant breeder,
that is, a geneticist, must be interested in the conditions
that determine inheritance, and these will be discovered

122 Plant Genetics

by investigating and controlling all kinds of inheritance.
No more favorable material for determining the funda-
mental facts of inheritance can be found among plants
than the spores of the simpler forms. They are accessible
and therefore capable of control; a succession of spore-
produced individuals represents a line whose purity can-
not be questioned; the so-called " modification of the
germ plasm" can be accomplished with a precision that
is impossible in the case of an egg nucleus within an
ovule and an ovary. In short, free from all the entangle-
ments of sex, the possibilities of variations in pure lines
can be determined and the possibilities of inheritance
of such variations. Such observations will establish the
facts common to all inheritance and will enable us to
recognize the contribution of sex to inheritance.

This is no criticism of the work that has been done
with seed plants. That work was the natural beginning
and the results have been remarkable. It is simply
calling attention to the necessary limitations of experi-
mental work in genetics with seed plants and the desira-
bility of extending the field of genetics so as to include
the lower forms that will permit certain desirable
things to be done with greater exactness.

1. PACE, LULA, Apogamy in Atamosco. Bot. Gaz. 56:376-394.
pis. ij, 14. 1913.

2. YAMANOUCHI, SHIGEO, Apogamy in Nephrodium. Bot. Gaz,
45:289-318. pis. Q, 10. 1908.

CHAPTER XIII
INHERITANCE IN GAMETOPHYTES

Thus far the discussions have dealt with inheritance
in sporophytes; in fact, genetics practically never con-
siders gametophytes, through which inheritance must
pass from one sporophyte to the next. The reasons
for this neglect are obvious. Practically all of our land
vegetation is made up of sporophytes and therefore
practically all of our experimental material has been
sporophytes. Furthermore, gametophytes are incon-
spicuous (out of sight in seed plants), hard to get at,
hard to work with, and apparently of no economic
importance. Besides, in animals, as is well known, the
generation equivalent to the gametophyte of plants is
represented only by a few cell divisions in the matura-
tion of gametes. In other words, the gametophyte
has no significance in the animal kingdom; and since
inheritance in plants is of interest to most investigators
chiefly because it throws some light upon inheritance in
animals, there has been no demand for any knowledge
of inheritance in gametophytes.

This means that this subject has not been studied, but
it may be considered here briefly. It may be admitted
that our ideas of genetics have been fairly well developed
without including the gametophyte in inheritance, but
the question is whether they cannot be developed better
and perhaps more quickly by including the gameto-
phyte. There are two reasons that may be suggested:

123

124 Plant Genetics

(i) Such a study will test the theories of inheritance
derived from the study of the sporophyte; (2) it deals
with less confusing material.

In the first place, the genetics of gametophytes may
furnish a good test of the theories of inheritance derived
from the study of sporophytes. The chromosome
equipment of a gametophyte is exactly the same as the
chromosome equipment of a gamete, in both cases the
chromosome number being haploid. This results from
the fact that in plants the reduction division occurs in
connection with spore formation; it is only rarely that
it occurs in connection with gamete formation, as in
animals, and then only when spores are cut out of the
life-history. Fig. 28 represents the typical life-history
of one of the higher plants, so far as the chromosome
situation is concerned. At the reduction division,
whose Mendelian significance has been considered, the
chromosome number is reduced from 8 to 4; and as a
result the spore, the gametophyte, and the gamete
all contain 4 chromosomes. At fertilization two sets
of 4 chromosomes are brought together by the male and
female gametes, and therefore the zygote contains 8,
and the resulting sporophyte continues the 8 chromo-
somes. At a certain season or, better, under certain
conditions cells of the sporophyte are set apart to pro-
duce spores (spore mother-cells, each with 8 chromo-
somes), and the reduction division follows, each spore
receiving 4 chromosomes. This familiar picture of the
life-cycle may now be applied to genetics.

The fact that the chromosomes are reduced in num-
ber from 8 to 4 is not the significant fact; the significant
fact is how this reduction takes place, that is, not the

Inheritance in Gametophytes

superficial result but the process. The process may be
recalled for a moment. Suppose that each of the 8
chromosomes in a spore mother- cell contains a single
determiner ; in such a case a nucleus can be represented

SPORE

MOTHER-CELL
8 CHROMOSOMES

(REDUCTION;
1 DIVISION 1

SPOROPHYTE

INDIVIDUAL

8 CHROMOSOMES

GAMETOPHYTE

INDIVIDUAL
4 CHROMOSOMES

ZYGOTE

8
CHROMOSOMES

<- \ FERTILIZATION!

FIG. 28. Diagram representing life-history of a higher plant as
regards chromosome number.

by fig. 29. When the reduction division occurs it
means that 4 of these chromosomes will go to each of
the two resulting nuclei. When this happens, are the
chromosomes chosen indiscriminately? It has been
learned that there is a certain limitation in the choice,

126 Plant Genetics

a limitation expressed by the statement that at the
reduction division allelomorphs never go to the same
nucleus, that is, the reduction division would never
produce a spore whose nucleus would contain both A
and a, for A and a are allelomorphs. All of this is
evident in connection with the Mendelian diagrams.
It should be remembered, however, that when it is said
that a spore or a gamete contains only half as many
chromosomes as does a sporophyte the whole truth has
not been stated. In fact, such a statement does not
include the most significant fact from the Mendelian
point of view; that is, that a spore
or a gamete never contains both of a
pair of allelomorphs in the nucleus.
It should be realized that 4 and 8
chromosomes are used only as con-
venient representative numbers, for

in plants there is a wide variation as

FIG. 29 L

to these numbers, and in the vast
majority of cases the sporophyte number is consider-
ably more than 8. In angiosperms, for example, the
haploid number ranges from 3 in Crepis to 45 in
Chrysanthemum. The statement that the essential
feature of the reduction division is to separate allelo-
morphs into different nuclei is simply a statement of
the theory of segregation, which is the fundamental
feature of Mendel's law. If this statement is not
true Mendel's law is not true, and all of our ideas of
inheritance will have to be reconstructed.

With this in mind, it may be considered how a knowl-
edge of inheritance in gametophytes may be of service.
For convenience we will consider first a single pair of

Inheritance in Gametophytes 127

allelomorphs (A and a). It will be recalled that such a
pair of allelomorphs may behave in two different ways in
hybrid sporophytes. In case of complete dominance,
A alone will be expressed in the hybrid A a. In case of
lack of dominance or of only partial dominance, the
hybrid A a will partake of the nature of both A and a; or,
in other words, it will be a blend, intermediate between
the two pure parents A A and aa.

A similar situation may be considered in the game-
tophyte generation. In a gametophyte, A can be
present or a, but not both. Here at least is a case of
presence and absence about which there is no doubt.
In this case there would be invariably pure dominance,
or what corresponds to it, but there could not be a blend.
It is interesting to consider the results of this situation.
If in a study of inheritance in gametophytes cases of
pure dominance only are discovered the result is quite
consistent with our theories. If, on the other hand,
cases of blending inheritance are discovered what are to
be the conclusions ? The only possible conclusion would
be that our theoretical mechanism is wrong; that
chromosomes are not the bearers of hereditary characters,
or at least not the only bearers; or else that the behavior
of the chromosomes at the reduction division is different
from that which cytologists and breeders have long
supposed. In other words, we have here a chance to
test the doctrine of segregation through the reduction
division.

The case of complementary factors may also be
considered. Recalling the behavior of complementary
factors in the sporophyte generation the question arises
about their behavior in the gametophyte generation.

128

Plant Genetics

Of course complementary
factors can exist together in
a gametophyte. A gameto-
phyte could not contain A a
(allelomorphs), but it could
contain both A and B,
which may be comple-
mentary factors.

It is evident also that
supplementary and inhibi-
tory factors are also possible
in gametophytes, just as are
complementary factors, for
they are not segregated by
the reduction division. Of
course a gametophyte show-
ing an entirely new char-
acter produced by the
interaction of two comple-
mentary factors would have
to be a gametophyte grown
from a hybrid sporophyte
in which these two factors
had been brought together;
and the same is true in
reference to supplementary
and inhibitory factors.
With these facts in mind the
validity of our theoretical
mechanism could be tested.

Finally, the case of
cumulative factors could be

Inheritance in Gametophytes

129

considered. These could also coexist in a gametophyte
but not to the same degree as in a sporophyte. If three
cumulative factors were active in a sporophyte, it is
evident that there could be gradations as indicated
by fig. 30. If, on the other hand, three cumulative
factors were active in a gametophyte there could only
be the gradations indicated by fig. 31. Thus the
behavior of cumulative factors in sporophytes and
gametophytes would differ quantitatively; there would

A

* s

S N

A'

2 Doses

I Dose

Dose

GAMETOPHYTE

FIG. 31. Diagram showing how the cumulative factors of fig. 30
could produce only four different quantitative classes in the gameto-
phyte generation.

not be so many gradations in gametophytes for the
same number of factors involved.

As a general conclusion, therefore, it may be said that
a knowledge of inheritance in gametophytes would test
our theoretical mechanism for inheritance in sporo-
phytes.

There is another important value to be secured from
a knowledge of genetics in gametophytes. In the study
of genetics in the sporophytes of the higher plants we
are dealing with very complex organisms. Attention
has been called to the possibilities of parthenogenesis

130 Plant Genetics

and vegetative apogamy, and to the fact that in such a
case any theoretical scheme of factors constructed to
explain breeding results would be without any signifi-
cance.

Again, in such a complex organism as the sporophyte
it is hard to say at all times just what things are being
inherited and what things are merely responses. Finally,
there always remains the confusing problem connected
with germ plasm and body plasm. Are these two plasms
as distinct in the higher plants as they seem to be in
animals ? In any event, can the body plasm affect the
germ plasm in any significant way, and is there any
mechanism to justify the inheritance of acquired char-
acters? These questions are simply samples of the
many complexities that confront us when dealing with
such organisms as the higher plants. The suggestion
has been made already that simpler plants be studied
which show fewer complexities. Algae, and even bry-
ophytes, have much to teach in the field of genetics,
for in such plants germ plasm and body plasm are not
clearly differentiated; frequently they are almost one
and the same thing.. Also such a phenomenon as
parthenogenesis is more easily studied in such plants, for
the structures involved are superficial and therefore
easily observed and controlled. This suggests another
advantage, therefore, in the study of inheritance in
gametophytes, for in algae practically everything is
gametophyte, while in bryophytes the gametophyte
generation is the dominant one and therefore the one
providing the best experimental material.

There is one disadvantage in breeding work among
these lower forms that should be realized, and that is

Inheritance in Gametophytes 131

the question of technique. It might be asked how one
could breed such a form as Ulothrix, in which the pro-
ductive cells are microscopic and are practically impos-
sible to identify or even to isolate successfully. One
could begin with such a form as Spirogyra, however,
in which the zygotes are rather large cells, so large
that they can easily be detected without a microscope.
Two biologists, BARBER (i) and KITE (3), by using fine
pipettes, have been able to isolate bacteria successfully
and even to drag individual chromosomes out of nuclei.
It would be far simpler to isolate the zygotes of Spiro-
gyra. Incidentally, it may be mentioned that the
zygotes of Spirogyra are sometimes produced by the
union of three sexual cells instead of two, and breeding
results in such a case should be very interesting.

For the following two reasons, therefore, a study of
inheritance in gametophytes is demanded and should
be undertaken: (i) to test our hypothetical mechanism
with its segregation of many kinds of factors; (2) to
study the lower plants where breeding phenomena are
not obscured by so many complexities.

In a secondary way, also, the gametophytc generation
becomes significant in genetics. Under the topics
of self-sterility, xenia, and hybrid vigor, inheritance
in gametophytes will be touched upon again. The
preceding discussion has been merely a suggestion of
the perspective of an entirely new territory for explora-
tion, a territory which should be made a part of that
which is already occupied.

In conclusion, the following quotation from EAST (2)
is pertinent: " Modern discoveries tend more and more|
to show that the sole function of the gametophyte of j

132 Plant Genetics

the angiosperms is to produce sporophytes. The
characters which they carry appear to be wholly sporo-
phytic, the factors which they carry functioning only
after fertilization." This is a logical inference from most
of the data on genetics. It does not mean, however,
that we should leave all. gametophytes out of considera-
tion, for active investigation of their behavior in inher-
itance should reveal facts of considerable theoretical
value.

1. BARBER, M. A., Univ. of Kansas Sci. Bull. no. 4. 1907 (p. 3).

2. EAST, E. M., and PARK, J. B., Studies on self-sterility. I.
The behavior of self-sterile plants. Genetics 2: 505-609. 1917.

3. KITE, G. S., The physical properties of the protoplasm of cer-
tain animal and plant cells. Am. Jour. Physiol. 22:146-173.

CHAPTER XIV
SELF-STERILITY

It has been discovered that many of the races of
cultivated plants, notably apples, do not develop viable
seed when self-pollinated. The practice adopted by
horticulturists, therefore, to meet this situation is to
make mixed plantings of several races. In this way
foreign pollen is always present, and this results in
viable seed where " own " pollen fails. This phenomenon
is known as self-sterility and seems to be well established
in many cases.

In connection with this phenomenon, the following
problem confronts us. Is self-sterility of any advantage
to plants in nature, or is it merely a necessary evil that
has arisen as a result of culture; or is it merely mechan-
ical or chemical necessity that has arisen in response to
certain conditions and is of no significance one way or
the other in the economy of plants ? Opinion is divided
on this point. Most biologists, however, claim that
self-sterility is of some ^advantage to the plant. This
opinion is related of course to belief in the Darwinian
doctrine of the survival of the fittest, survival being a
proof of usefulness. For this reason it has become
customary to regard a plant character as useful until
the contrary is proved. This point of view has been of
service in the past and is still useful, subject to certain
limitations.

133

134 Plant Genetics

If self-sterility is useful to the plant the n&tyuce of its
usefulness should be considered. The situation may
be outlined as follows. The first plants used only
asexual reproduction, and this has proved to be a very
simple and effective method. Why then was sex ever
developed? It will be remembered that WEISMANN
was the first to answer this question clearly, claiming
that sex was developed to secure greater variation.
Greater variation is desirable of course to secure adapta-
tions to diverse conditions of living. It is well known
that asexual reproduction is the surest way to duplicate
the parent form and would result in uniform plants
generation after generation. If a species remains uni-
form within certain narrow limits it will be impossible
to meet changing conditions. This will result in two
evils: (i) if conditions are changed where the uniform
species is growing the species will be exterminated;
(2) even if the local conditions remain the same the
species will be unable to spread into new localities where
the conditions are different. If, on the other hand, a
species shows wider variations, certain of these might
fit the new conditions. Such variations will survive
when conditions change and would always solve the
problem of migrating into new habitats. For these
reasons variation is necessary to the general success of
the species.

This increase in the range of variations, therefore, is
the significance of the introduction of sex, for the fusion
of two germ plasms in the sex act greatly increases
chances for variation. WEISMANN thought that sex
secured the small individual variations which are so
common in any species, but now most such variations

Self -Sterility 135

are regarded merely as fluctuations or responses. Such
variations are not inherited, and therefore can be of no
phylogenetic significance; that is, of no value to the
future of the species. Our recent knowledge of the
factor hypothesis, however, indicates that many of these
variations are more than fluctuations. They may
represent germinal differences, such as varying "doses,"
and such differences are inherited and may be of some
value therefore in the future of the species.

It will be realized, however, that it is possible to
have a homozygous race. If a race were strictly
homozygous in respect to every heritable factor the
sex act would result in no more variation than would
asexual reproduction, provided of course our ideas of the
mechanism of inheritance in sexual plants are correct.
The fact is that a perfectly homozygous plant is probably
rare. Supposing, however, that it were common, we
would have to resort to some other mechanism than the
sex act to secure the desirable variation. If sex does
not secure variation what can ?

The answer is cross-pollination. Although races
were all homozygous, cross-pollination would make
possible the crossing of different races, which would
inevitably result in great variation in the progeny.
Even if homozygous races are few it is certainly true that
in any event cross-pollination results in greater variation
than self-pollination. This is the biological basis of
the many devices in plants to secure cross-pollination;
in fact, this is carried to such an extreme in many
cases that self-pollination is absolutely prevented. Sta-
mens and carpels may mature at different times; and
the great number of dioecious plants must all be

136 Plant Genetics

cross-pollinated. DARWIN uncovered this situation and
concluded that there is a general tendency among
plants to secure cross-pollination. The significance of
this universal tendency seems to be to secure variation.

Some ascribe another value to cross-pollination.
Hybrid vigor is a well established phenomenon, which
will be discussed later. Continued inbreeding fre-
quently results in degeneracy, such inbreeding races
being spoken of as " running out." It is a well known
fact that to cross an inbred race with some other race
results in a kind of rejuvenescence, a regaining of the
original vigor that was present before inbreeding began.
Even in the case of two normally vigorous races a cross
will often produce a hybrid more vigorous than either
parent. Of course, one might regard hybrid vigor
itself as a device to secure crossing.

In any event, it seems to be safe to conclude that
cross-pollination is of advantage to plants, and of all
mechanisms to secure cross-pollination, self-sterility
may be regarded as the culmination. If "own" pollen
cannot produce viable seed in a plant cross-pollination
becomes a necessity. Inbreeding cannot continue in a
race that has become or is becoming self-sterile.

This raises the question why is "own" pollen not
effective where foreign pollen is? There are well
known cases in which pollen is not effective on stigmas
of the same race but quite effective on those of another
race. Although the biological advantage of such be-
havior is evident, this does not explain the mechanism
involved.

The theories offered in the explanation of self-sterility
are very numerous. Some years ago an attempt was

Self-Sterility 137

made to explain self-sterility as a response. It was
demonstrated that when a race of plants was grown
under conditions of unusually high humidity its pollen
grains would burst before they had a chance to function.
This explanation was obviously true in a few simple cases,
but the phenomenon of self-sterility usually involves
something much more perplexing than merely the non-
functioning of pollen. The real problem is to explain
why pollen will not function on "own" stigmas and
will function on the stigmas of other closely related races,
while it requires foreign pollen to function on "own"
stigmas. Such a complex situation could hardly be
explained as a response; more probably it is involved
with inheritance. Morphologists, cytologists, ecologists,
physiologists, and practical horticulturists have all
contributed explanations; in fact, it is not assured as yet
to which of these fields the phenomenon belongs.
In the meantime, what concerns us is whether self-
sterility can be explained by the principles of genetics.

Such a theory has been proposed. Several investi-
gators have suggested that there is a heritable factor for
self-sterility, and some investigations of this possibility
have been undertaken, but as yet most of the results
have been negative. It must be realized, however, that
if self-sterility is ever to be explained on the basis of
hereditary factors the mechanism involved will be
rather unusual. Although most of the explanations
proposed by geneticists have proved inadequate there is
at least one that deserves consideration. It is a typical
Mendelian explanation and very ingenious. It fitted the
facts perfectly when first proposed and, still better, it
has continued to fit the facts that have been uncovered

138 Plant Genetics

by subsequent breeding work. Furthermore, it is the
first true Mendelian explanation that has been based
on the germinal constitution in the gametophyte as a
deciding factor.

In 1914 BELLING (i) published a paper entitled "A
study of semi-sterility." Apparently this is not true
self-sterility, but it is a closely related phenomenon and
may be considered in this connection. BELLING was
conducting breeding experiments with two races of
beans, both of which were completely fertile. When
he crossed them, however, the resulting hybrids were
semi-sterile. Uniformly just one-half of the pollen
grains appeared empty and collapsed, while one-half
of the ovules had no embryo sacs. The sterile pollen
and ovules appeared in random distribution with the
fertile.

Inbreeding the semi-sterile hybrids, BELLING ob-
tained an F 2 generation which showed the following
features: one-half the plants had perfect pollen, the
other one-half had a mixture of equal numbers of good
and bad pollen grains in all their flowers. The plants
which had perfect pollen also had perfect ovules, while
the plants with 50 per cent sterile pollen also had 50
per cent sterile ovules. In the F 3 generation all the
descendants from the fertile plants had perfectly good
pollen and ovules; but the progeny of the semi-sterile
plants again split up into two classes (fertile and semi-
sterile) as before.

BELLING states his general conclusion as follows:
"The explanation of the random abortion of one-half
of the pollen and one-half of the embryo sacs must
apparently be by the segregation of Mendelian factors

Self-Sterility 139

among pollen grains and embryo sacs individually, and
not by the action of these factors on the zygotes."

To make this suggestive situation clear the diagram
(fig. 32) may be considered. It enlarges a little upon
SELLING'S original ideas as he stated them, and empha-
sizes the sporophyte-gametophyte relationship. Pollen
grains and embryo sacs are gametophytes in the sense
that they inclose the male and female gametophytes,
so that when the diagram shows sterile gametophytes it
is the same as saying that both pollen grains and embryo
sacs are sterile. This of course is just what BELLING
found; whenever one-half the pollen grains in random
distribution were sterile one-half the embryo sacs in
random distribution were also sterile.

It should be remembered that BELLING started with
two completely fertile races. Suppose that the parent
race A had a factor (X) whose absence brought steril-
ity in the gametophytes (pollen grains or embryo
sacs). Species B has a different factor (Y), with a
similar effect, but inherited independently. When
BELLING crossed these races, all of the resulting F x
hybrids were semi-sterile. In other words, in every
F! plant one-half the gametophytes were sterile. It is
easy to see why XO and OY are fertile, also why OO
is sterile (lacking both factors); but why should XY
be sterile when it has both factors ? BELLING explains
it by saying that gametophytes are unlike sporophytes
in that normally they have single factors instead of
double factors. The germinal capacity of a game to-
phyte is just one-half that of a sporophyte. It is as
if a gametophyte were " supersaturated " by a double
factor. Such a situation is abnormal for a gametophyte

140

Plant Genetics

PARENT)
RACE SPOROPHYTE:

PARENT

RACE

B

SPOROPHYTE:

F,i SPOROPHYTE:

FERTILE)

FERTILE)

TILE)

'(FERTILE)

TILE)

'(FERTILE)

'(STERILE)

(STERILE)

FIG. 32. Diagram illustrating Balling's explanation of semi-
sterility.

Self -Sterility 141

and brings abnormal results. Therefore the gameto-
phyte having the abnormal double dose (XY) is just as
sterile as the gametophyte with no dose (OO).

In developing the F 2 ratios of course only the fertile
gametophytes function. XY and OO are eliminated,
so far as posterity is concerned, so that we have to
deal only with the chance matings among the fertile
gametophytes (XO and OY). According to mathe-
matics there are four possible matings between these
two gametophytes (fig. 33). Out of these four result-
ing F 2 sporophytes two would evidently produce only
fertile gametophytes and would remain fertile as long as
they are inbred. The other two are exactly like the
original F r hybrid and therefore semi-sterile, giving
one-half sterile gametophytes. The whole dynasty
may be represented as follows:

P Fertile X Fertile

>
FI Semi-sterile

5%

, . ^ .

F 2 Fertile Semi-sterile

50% 50%

^ ^

Fertile Fertile Semi-sterile

v t y

F 4 Fertile Fertile Fertile Semi-sterile

This is a very ingenious scheme and, like others,
should be tested by further experiments. To a certain
extent it has already met this test, for BELLING (2)
has subsequently reported a few more generations
in which the breeding results were entirely consistent

142

Plant Genetics

with those of the earlier generations. Also he has
discovered two new pure races of beans which give
similar results. It remains to be seen whether this
explanation will fit other cases of semi-sterility.

This theory of semi-sterility of course may have
some value in throwing light upon the problem of com-
plete self-sterility. It seems to be of greater theoretical

interest, however, in
being the first attempt
to include inheritance in
gametophytes in an ex-
planation of Mendelian
results. BELLING him-
self does not stress this
gametophyte relation-
ship; in fact, he does not
use the word "gameto-
phyte"; and yet it is
probably this gameto-
phyte relationship that
represents the most
significant feature of his
work and will prove to be the most 'suggestive feature
in future investigations.

Attractive as this explanation may be, vte must
realize, however, that as yet it merely explains semi-
sterility and does not clear up the fundamental com-
plexity of the more general topic of self-sterility, that is,
why pollen fails on "own" stigmas and functions on
foreign stigmas.

Very recently an explanation of the more general
problem of self-sterility has been proposed by EAST

@@

FIG. 33. Diagram showing how
the F 2 would be produced according
to SELLING'S idea of semi-sterility.

Self -Sterility 143

and PARK (3). This explanation seems to fit the facts
correctly but has certain theoretical disadvantages.
In any event it involves such complex Mendelian con-
ceptions that we shall not try to explain it here. One
statement, however, which was considered in the pre-
vious chapter again commands our attention. EAST
and PARK state "that modern discoveries tend more
and more to show that the sole function of the gameto-
phytes of the angiosperms is to produce sporophytes.
The characters which they possess appear to be wholly
sporophytic, the factors which they carry functioning
only after fertilization/' This statement would seem to
throw out of court all consideration of the gametophyte
generation and would regard as impossible BELLING'S
theory, which is based on the direct influence of the
germinal equipment of gametophytes upon gameto-
phytes themselves. We will not enter into any general
discussion of this point, but merely suggest that the
self-sterility in the two cases might have been involved
with distinctly different phenomena, since SELLING'S
material showed degeneration and sometimes complete
abortion in pollen and embryo sacs, while the material
(Nicotiana spp.) of EAST and PARK was self-sterile
merely because of the failure of pollen tubes. The
hereditary mechanism of the two cases must be quite
different.

i. BELLING, JOHN, A study of semi-sterility. Jour. Heredity

5^65-75. figs. 7. 1914-
2> ? A hypothesis of semi-sterility confirmed. Jour.

Heredity 7:552. 1916.
3. EAST, E. M., and PARK, J. B., Studies on self-sterility. I. The

behavior of self-sterile plants. Genetics 2:405-609. 1917.

CHAPTER XV
THE ENDOSPERM IN INHERITANCE

We have dealt chiefly with inheritance in the sporo-
phyte, in connection with which most work in plant
genetics has been done. Brief mention has been made of
inheritance in thej;ametophyte, in connection with which
there has been very little work. It is natural now to
consider inheritance in the endosperm. This classifica-
tion raises the question as to the nature of the endo-
sperm. In general it has been regarded as belonging
to the gametophyte generation, but since the discovery
of " double fertilization " in 1898 some have claimed
that it belongs to the sporophyte generation. On the
basis of chromosome numbers it is neither, so that there
is also the claim that the endosperm is neither sporophyte
nor gametophyte; at least we are justified in considering
inheritance in endosperm as a separate topic. As might
be inferred, endosperm shows some features characteristic
of a gametophyte, others characteristic of a sporophyte,
and still others peculiar to itself. Judgment as to its
nature, therefore, will depend upon which of these
features are emphasized.

It is generally believed that angiosperms have been
derived from gymnosperms, and it is natural therefore
to explain angiosperm structures by the corresponding
structures of gymnosperms. The gymnosperm and
angiosperm ovules are contrasted in fig. 34, which will
assist in the following discussion. In gymnosperms the

144

The Endosperm in Inheritance

situation is clear. After the germination of the mega-
spore, everything within the old megaspore wall is

Gymnosperni

Archegonium

Egg

Old Megaspore

Wall, Inclosing

emale Gametophyte'

Angiosperm

Egg

Fusion Nucleus

Old Megaspore

Wall, Inclosing

Female Gametophyte

OVULE

Gymnosperm

Embrya
Megaspore Wall

Endosperm

SEED

Angiosperm

Embryo
Megaspore Wall

Endosperm

FIG. 34. Diagrams contrasting young ovules and mature seeds of
gymnosperms and angiosperms.

gametophyte tissue; fertilization^ affects the egg only,
resulting in a sporophyte embryo. In the seed, there-
fore, the embryo is imbedded in nutritive tissue which is

146 Plant Genetics

evidently the vegetative body of the female gameto-
phyte, and this tissue is called the endosperm.

In angiosperms, however, a new situation intro-
duces doubt. It can be said as before that after the
germination of the megaspore everything within the
megaspore wall is a female gametophyte tissue, but
it cannot be said that fertilization affects the egg only,
for one of the sperms fuses as regularly with the fusion
nucleus as does the other sperm with the egg. It will
be remembered that the fusion nucleus is formed by
two nuclei, which have migrated from each end of the
sac, so that when the sperm enters into the fusion there
is a triple fusion. After fertilization the fertilized egg
of course forms the embryo sporophyte, but usually
every nucleus of the old gametophyte disappears except
the fertilized fusion nucleus, which then forms the
endosperm in which the young sporophyte is imbedded.
For this reason the fertilized fusion nucleus is usually
called the ^jidosgerm ^nucleus.

A comparison of angiosperm and gymnosperm seeds
reveals the f ollowing^contrast (see fig. 34) . In the appear-
ance of their essential structures they are exactly alike,
and for this reason some botanists claim that the endo-
sperm of angiosperms is the same as that of gymno-
sperms, that is, gametophyte tissue. The opposing claim
is that, although the gymnosperm endosperm is gameto-
phyte tissue, the situation in angiosperms is essentially
different. In angiosperms the endosperm does not arise
from morphologically unmodified gametophyte tissue, as
in gymnosperms, but entirely from the endosperm nu-
cleus, and this nucleus is clearly the product of the fusion
of male and female nuclei. With such an origin the endo-

The Endosperm in Inheritance 147

sperm nucleus is comparable with the zygote, and the
endosperm tissue is sister to the embryo sporophyte.
In other words, in angiosperms, embryo and endosperm
are twins. This means that the endosperm of angio-
sperms belongs to the sporophyte generation, although
of course it is a distinct individual which produces no
progeny. The embryo sporophyte is a parasite upon its
twin and destroys it.

It will be recognized that there is some reason for
each of these claims. Is there any way of testing the
claims, that is, of distinguishing between sporophyte and
gametophyte tissue ? The cytological distinction is that
the sporophyte is 2X tissue and the gametophyte x
tissue. Applying this test it is found that endosperm
tissue is neither x nor ix, but 3% as might be expected
from the triple fusion. The conclusion involves several
possibilities, as follows: $x is evidently nearer 2X than x,
and therefore endosperm tissue is more like sporophyte
tissue than gametophyte tissue; but on the other hand
two of the #'s have come from the female gametophyte,
and therefore two-thirds of the endosperm nucleus is
female gametophyte tissue. On the basis of predomi-
nance, therefore, endosperm tissue is more like female
gametophyte tissue than anything else. Finally, there
is a third alternative, and that is that the 3^ condition
deserves to be set apart in a category by itself, which
would mean that the endosperm is neither gametophyti
nor sporophyte.

These are the claims and the evidence as to the na-
ture of angiosperm endosperm. Opinion is not settled,
but the facts are clear. This prepares for a consid-
eration of the bearing of this situation upon genetics.

148 Plant Genetics

The geneticist is not much concerned about the exact
morphological or cytological nature of endosperm, but he
is much concerned about its behavior in inheritance, and
perhaps the phenomena of endosperm inheritance may
help to decide whether endosperm is gametophyte or
sporophyte or neither.

Certain races of corn have yellow endosperm, while
in other races it is white (colorless). If a cross is made
with pollen from the yellow endosperm race - on the
silks of the white endosperm race, what would the infer-
ence be as to the result ? We could assume that yellow is
dominant over white, since yellow is probably due to the
presence of a factor which is lacking in the white. In
making such a cross, therefore, we would expect a hybrid
embryo to be formed which would show the dominant
character of yellow endosperm when this embryo
becomes a plant bearing ears the next season. On the
contrary, we find that the dominant yellow character
appears, but it appears that same year. The cross
of course puts the yellow endosperm factor in the young
hybrid embryo, but we cannot imagine that this embryo
passed the character out into the endosperm that sur-
rounds it. The real mechanism is as follows.

Some time after this phenomenon was discovered in
1872, it was named xtmia (in 1881), the definition of the
term being the direct effect of foreign pollen upon the
endosperm. At the time of its discovery the mechanism
involved in xenia was not understood. Later, double
fertilization was discovered, and this furnished a complete
mechanism. A pollen grain from the yellow endosperm
race contains two male gametes, and each gamete con-
tains the factor for yellow endosperm. One of the

The Endosperm in Inheritance 149

gametes fertilizes the egg and produces a hybrid embryo,
which, in the next generation, behaves as a heterozygote
for yellow endosperm. ^The other male gamete fertilizes
the fusion nucleus and produces the endosperm nucleus,
which therefore contains the factor for yellow endosperm,
the result being that the endosperm is yellow, although
the ovule belongs to the white race. Xenia means,
therefore, that the endosperm is a hybrid as well as
the embryo, and "triple fusion" involves the trans-
mission of hereditary characters. Fertilization of the
fusion nucleus is just as essential as fertilization of the
egg, and so far as inheritance is concerned the endosperm
and embryo are sister sporophytes.

Xenia has caused much discussion in genetics. It
throws light upon the nature of endosperm and suggests
that it belongs to the sporophyte generation because it is
a product of an act of fertilization. Because of its
behavior in inheritance geneticists would naturally
regard the endosperm as a sporophyte, an abnormal
sister to the embryo.

Cases of xenia are not limited to yellow endosperm.
Xenia appears also in the crosses between sweet and
starchy corn. Crosses of red corn and purple corn also
show xenia, but in this case additional details appear.
A section of a grain of corn is shown in fig. 35. There is
first the pericarp or "seed coat," which is the ovary
wall, belonging to the old sporophyte, and therefore
does not concern us. Within this is a thin aleurone
layer, which is the outer layer of endosperm, while the
bulk of the seed consists of the starchy endosperm.
Since aleurone is endosperm, colors peculiar to it would
show xenia in inheritance. This was shown in the case

Plant Genetics

Pericarp

Endosperm

of EAST'S red and purple corn, the colors being located
in the aleurone.

There is another phase of ibis situation to which
attention should be called. By pollinating the silks
of a white-grained individual with pollen from a red-
grained individual xenia is secured, the resulting grains
being red like those of the pollen parent. In the recip-
rocal cross, however, that is, pollinating silks on a red-
grained individual with
pollen from a white-
grained individual, a differ-
ent result is obtained.
The resulting grains are
not white like those of the
pollen parent, but red like
those of the ovule parent.
There is no xenia, there-
fore, for the pollen has no
immediate effect upon the
developing endosperm.
This seeming difficulty,
however, is easily explained. When the pollen parent
is white and the ovule parent is red, the endosperm gets
its characters from both parents, and since red is domi-
nant over white the resulting endosperm will be red
because the female nuclei that entered into the triple
fusion carried the factor for red endosperm; and there-
fore the pollen from the white parent seemed to have
no effect. The mechanism works in all cases, but owing
to dominance, xenia appears only in certain cases.
There is no need to discuss all of the Mendelian situations
in which xenia may occur. An understanding of the

Embryo

Cross-Section of
Corn-Seed

FIG. 35. Diagram of corn seed

The Endosperm in Inheritance 151

fundamental situation as described should enable us
to analyze such cases and reach a conclusion as to the
expected results.

A law which EAST (i) has formulated in reference to
xenia is pertinent: "When two races differ in a single
visible endosperm character, in which dominance is
complete, xenia occurs only when the dominant parent
is male (pollen parent). When the two races differ in
a single endosperm character, in which dominance is
incomplete, or when they differ in two characters, both
of which are necessary for the development of the visible
differences, in both of these cases xenia occurs when
either parent is male/' This may be called "normal"
xenia. What may be called "abnormal" xenia must
now be considered.

WEBBER (3) in 1900, in experimenting upon xenia in
corn, uncovered some interesting abnormalities. Pollen
from a red-grained race, applied to silks of a white-
grained race, should result in solid red grains if xenia is
normal. In most cases WEBBER obtained this result;
but sometimes there appeared grains which showed
blotches of red and white in a kind of irregular mosaic.
For these cases he constructed an ingenious and reason-
able explanation.

Normally the second male nucleus fuses with the
fusion nucleus, and the result is a solid red grain. In
some cases, however, the second male nucleus does not
join with the two other nuclei, but all three divide
separately without having fused, and in such a case the
mature grain is white with red streaks. In still other
cases the two polar nuclei fuse, but the second male
nucleus does not enter into the complex, and in this case

152 Plant Genetics

the fusion nucleus and male nucleus divide separately
to form endosperm. WEBBER concludes that in this
last case there will be very mucj^ white and very little
red in the mature grain; for the fusion of the polars
to form tKe fusion nucleus has resulted in the usual
growth stimulus that is evident in all fusions. In the
previous case in which all three divided separately, if
more white than red develops, it is simply because there
are two female polars (white-producing) to one male
gamete (red-producing). In the latter case, however,
the fusion of the two female polars stimulates them to
more rapid division, so that they greatly preponderate
over the single unstimulated male gamete, the result
being that much white and little red appears in the
mature grain.

The appearance of red and white blotches rather than
a more regular distribution of colors is explained by the
usual method of endosperm formation. Endosperm
formation begins with free nuclear division, the resulting
nuclei being free in the cytoplasm of the sac. The
cell walls, which limit them, are not formed for some
time; sometimes not until all the nuclei have been
formed. Before a large number of free nuclei have
appeared they move from the central region of the sac
and usually become placed near the wall, where free
nuclear division continues. When walls begin to appear,
separating the nuclei, wall formation begins at the
periphery of the sac and extends toward the center,
in what is called centripetal growth. This program,
which is common in seed plants and is known to occur
in wheat, is doubtless the program in corn. If then the
second male gamete fails to unite with the fusion nucleus

The Endosperm in Inheritance 153

and each divides separately, when their progeny nuclei
move out to the periphery of the sac the nuclei of male
and female origin doujbtless become more or less inter-
spersed. In their further division there would be groups
of cells of male origin interspersed among groups of
female origin. The result would be reel and white areas
on the mature grain, intermingled as irregular blotches.

Later, EAST (i) met the same phenomenon in his
experiments. He found, however, that sometimes
half the grain was red and half white, with a definite
boundary line between the two areas. Such a situa-
tion evidently did not agree altogether with WEBBER'S
conception of the mechanism; in fact, WEBBER himself
found similar cases and explains them as follows. When
the migration of the nuclei from the center of the embryo
sac to the periphery occurs, if the nuclei from the male
gamete have remained grouped together, as might well
occur, and the nuclei from the fusion nucleus have
remained grouped, it is probable that in their migra-
tion to the periphery those of one group would come
to occupy one portion of the periphery and those of the
other group the other portion. In other words, the
two groups would migrate en bloc to different regions of
the sac wall. This would lead to the production of
grains in which approximately half the endosperm would
resemble one parent and the other half the other parent.

We have called these cases abnormal xenia. The
explanation is ingenious, based upon facts known to be
true in other seed plants. Whether these phenomena
occur in corn or not remains to be proved, but the
angiosperm embryo sac program is so uniform that
we can hardly doubt it. Inasmuch as corn is probably

154 Plant Genetics

the most important material in plant genetics today it is
very desirable that the histology of the sac should be
investigated in detail. It will be realized that xenia
supplies a very convenient index for breeders. Corn
breeders especially must realize fully the fact of xenia,
and they are able to work more intelligently if they under-
stand the mechanism, since the principal work in corn
breeding has to do with the character of the endosperm.

Having introduced the phenomenon of double fertili-
zation in connection with xenia, its use in explaining other
things than the mere fact of xenia may be considered.
It will be recalled that EAST (2) has done a large amount
of experimental work with sweet and starchy corn.
Careful examination enabled him to distinguish two
distinct races of starchy corn. Both races of course had
starchy grains, but the starch seemed to be laid clown
in different ways in the two races. In one race the
starch occurred in a loose powdery or floury condition,
while in the other race it was compacted into a hard,
flinty, or so-called corneous condition. The two races,
therefore, may be spoken of as floury and corneous races
of starchy corn.

EAST made various crosses between these two races
to discover the method of inheritance of the two endo-
sperm characters. Naturally such characters would
be expected to show xenia. In the following descrip-
tion, therefore, when the F x generation is referred to,
both the hybrid embryo and the hybrid endosperm
surrounding it will be included.

When EAST used the floury race as the pollen parent
and the corneous race as the ovule parent, the F x genera-
tion was all corneous. When he made the reciprocal

The Endosperm in Inheritance 155

cross (corneous pollen and floury ovule), the F t genera-
tion was all floury. This result certainly suggests
maternal inheritance, for in both cases it is the character
of the ovule parent that is transmitted, while the pollen
parent seems to have no effect. If it is assumed that
this is a case of maternal inheritance, similar to the
four-o'clocks of CORRENS, with their diseased cytoplasm,
two problems are encountered: (i) to prove that this
behavior is not due merely to parthenogenesis; (2) to
discover the mechanism to explain maternal inheritance
in this case.

In the first place EAST established the fact that there
was no possibility of parthenogenesis. Continuing his
investigation he inbred the F x generation in each case
and examined the F 2 progeny. If we are dealing with a
case of maternal inheritance, what should the F 2 genera-
tion show? It should be exactly the same as the F r
generation, for in true maternal inheritance a race will
go on forever breeding true to the maternal character,
whether it is self -pollinated or cross-pollinated. If this
had been a case of true maternal inheritance EAST
should have obtained the following situation:

Floury X Corneous Corneous X Floury

N|/ ^

FI Corneous Floury

i i

F 2 Corneous Floury

I *

etc. etc.

Actually, however, he obtained the following result:

Floury X Corneous Corneous X Floury

^ I

F! Corneous Floury

F 2 ^ Floury - Corneous \ Corneous | Floury

156 Plant Genetics

The conclusion is therefore that this is not a true case
of maternal inheritance. EAST offers a very reasonable
explanation of these results, based upon double fertiliza-
tion plus the cumulative factor idea. These characters
appear superficially to be maternal for the following
reasons. The endosperm nuclei are 3$, 2X from the
female and ix from the male. In the characters under
discussion the presence of two factors always dominates
the presence of one factor; thus corneous female crossed
with floury male produces progeny which are all pheno-
typically (in appearance) corneous, while floury female
crossed with corneous male for the same reason produces
progeny which are all phenotypically floury. The
cumulative factor idea will be recognized. The mother
always determines the character of the hybrid endo-
sperm because there are always two female nuclei to
predominate over the single male nucleus. In the
embryo, however, this predominance does not occur,
for there only a single female nucleus has fused with
a single male nucleus. When this hybrid embryo
matures, therefore, it is evident that it will produce
gametes of two sorts, 50 per cent corneous and 50 per
cent floury. It is evident that the female is really the
only decisive factor, so far as the appearance of endo-
sperm is concerned, so that the ratios appearing among
the female gametes in the F x generation will be the
ratios that will appear also in the F 2 endosperms. In
other words, 50 per cent of the F 2 endosperms will be
corneous and 50 per cent floury, no matter what may
be the source of the pollen. It is obvious that the
explanation of this peculiar form of apparently maternal
inheritance depends entirely upon a clear conception of

The Endosperm in Inheritance 157

the phenomenon of triple fusion. On the other hand,
this type of inheritance indicates that the triple fusion,
so far from being an unimportant cytological process,
is really significant.

It may be questioned why the other endosperm
characters are not inherited in this way, as for example
aleurone colors, just as well as floury and corneous
characters. The question cannot be answered satis-
factorily, any more than it can be explained why some-
times there is complete dominance and sometimes
only partial dominance. All that can be said is that
corneous and floury characters appear to be inherited in
a different way from endosperm colors, where the
presence of one color factor is sufficient to cause the
complete determination of color.

1. EAST, E. M., and HAYES, H. K., Inheritance in maize. Conn.
Agric. Exper. Sta. Bull. no. 167. pp. 142. pis. 25. 1911.

2. HAYES, H. K., and EAST, E. M., Further experiments on
inheritance in maize. Conn. Agric. Exper. Sta. Bull. no. 188.
pp. 31. pis. 7. 1915.

3. WEBBER, HERBERT J., Xenia, or the immediate effect of
pollen in maize. U.S. Dept. Agric. Bull. no. 22. pp. 44.
pis. 4. 1900.

CHAPTER XVI
HYBRID VIGOR

The phenomenon of hybrid vigor has been mentioned
in a general way, but its relation to genetics remains to be
considered. Geneticists have proposed various explana-
tions of hybrid vigor, and some of these will be presented.

The first record of observations on hybrid vigor is
that of JK.OLREUTHR in 1776, who states that crossing
results in an increase in general vegetative luxuriance
and an increase in the facility of vegetative propagation
and viability. Later GARTNER, discussed the same
phenomenon but gave no important new ideas. Finally,
hybrid vigor attracted the attention of DARWIN, who
states that crossing hastens the time of flowering and
maturing and increases the size of the individual. He
adds the very important fact that it is not mere crossing
that gives the stimulus, but crossing forms that differ
in the constitution of their sex elements; in other words,
crossing between individual flowers on the same plant
gives no advantage. It is DARWIN who is responsible
for bringing hybrid vigor to the attention of botanists,
although the modern popular impression would be that
BURBANK deserves this credit because of his experience
in producing some remarkably fast growing, large, and
vigorous hybrids. Among the modern investigators of
hybrid vigor, the work of SHULL and EAST will be con-
sidered, both of whom have worked upon the subject
during the last ten years.

158

Hybrid Vigor 159

SKULL'S (6) conclusions up to the year 1910 may be
summarized as follows. His work was entirely with
corn, and the conclusions contained some very significant
points.

1. "The progeny of every self-fertilized corn plant is
of inferior size, vigor, and productiveness, as compared
with the progeny of a normally cross-bred plant derived
from the same source." In general this conclusion
would be admitted by everyone, but it raised one
question. It was known that when two races have been
inbred for many generations they frequently "run out/'
gradually losing their vigor. In such a case a cross
between the two races tends to restore the original vigor.
The remaining question, however, is whether the same
thing may be effected by a cross between two inbred
races which have not run out but remain in normal vigor.
SHULL answers that hybrid vigor is exhibited when both
parents are above the average condition as well as when
they are below it.

2. Another question which naturally arises is as fol-
lows. When these crosses are made it is of course the
F! generation that shows the hybrid vigor. If the F x
generation is inbred, what is the status of the F 2 gen-
eration with reference to vigor? SHULL answers this
question in the following general way. "The decrease
in size and vigor which accompanies self-fertilization is
greatest in the first generation, and becomes less and less
in each succeeding generation, until a condition is reached
in which there is (presumably) no more loss of vigor."
The facts involved in this statement may be represented
in fig. 36. In this figure it can be seen clearly how
the great loss of vigor comes immediately after self-

i6o

Plant Genetics

fertilization again begins. After that self-fertilization
brings additional loss in vigor, but this loss is less with
each succeeding generation. It is as though a very
definite limit were being approached and each genera-
tion goes down one-half of the remaining distance
toward that limit. Just why and in what way this limit
is approached will be considered later in connection with
the work of EAST.

Parent
Races

F,

F 7 F a etc.

FIG. 36. Illustrating status of hybrid vigor in F x and later genera-
tions. Vigor represented by height of rectangles.

3. "A cross between_sibs (sister and brother) within
a self-fertilized family shows little or no improvement
over self-fertilization in the same family/' This, it will
be noticed, is simply carrying a little farther the point
that DARWFN originally discovered. DARWIN found
that crosses between flowers on the same plant did not
result in hybrid vigor. SHULL now finds that crosses
between different individual plants in the same race is
of no effect. Wejrealize that an inbred race should be
homozygous; therefore all the individuals involved
would have the same germinal constitution. A cross
between any two such individuals certainly could not

Hybrid Vigor 161

produce a hybrid in any sense, so that it would not be
surprising that such a cross fails to bring hybrid^vigor

4. "A cross between plants belonging to two self-
fertilized families results in a progeny of as great vigor
size, and productiveness as are possessed by families
which have never been self- fertilized/' The conclusior
from this is that inbreeding results in no permanent
loss of vigor. A race may "run out" if inbred con-
tinuously, but when crossed with another race it im-
mediately seems to regain all the original vigor. It
is as though all germ plasm contains the potentiality
of developing vigorous individuals. This potentiality,
however, cannot express itself until the proper combina-
tion of conditions arises, and this proper combination
seems to be connected in some way with hybridizing.

5. " Reciprocal crosses between two distinct self-
fertilized families are equal" in producing hybrid vigor.
When reciprocal crosses are equal it immediately sug-
gests Mendelian segregation. Is it possible that hybrid
vigor may be explained in terms of Mendelism ? These
are the five "laws" of hybrid vigor developed by SHULL
in 1910. It should be noted that they are not hypotheses
but observed facts. The hypotheses were developed
later when more of the facts were in.

A practical suggestion made by SHULL in connection
with hybrid vigor is of interest. Granted that hybrid
vigor is an established fact, the question of its practical
use in connection with crop plants should be taken into
account. If a farmer after years of work has finally
developed a desirable new strain of corn by selection, or
isolation, or both, he is not likely to favor hybridizing
with some other strain in any wholesale way. He must

l62

Plant Genetics

preserve his pure strain at all costs. SHULL has sug-
gested the following solution of this practical problem, as
indicated in fig. 37. Two desirable strains (A and B)
are developed. One small plot (I) is planted entirely
with A, and at some distance another small plot (II) is
planted with A and B in alternating rows. Plot I is
used only to perpetuate strain A in pure condition. In
plot II all the A plants are detasseled. The silks of
these A plants, therefore, are pollinated by B pollen only,

A B A B A B

PLOT I PLOT

FIG. 37. SHULL'S scheme of planting for making practical use of
hybrid vigor in corn.

and the resulting grains in the A rows are all bound to be
hybrids. Using these grains as seed for the crop, hybrid
vigor will be obtained. At the same time both A and B
are perpetuated in the pure condition, since the B rows
in plot II are always self -pollinated, as there is no
other pollen in that neighborhood. This is a very
simple solution of the problem, without necessitating
laborious hand-pollination.

The investigations and conclusions of EAST (2) may
next be considered. SHULL did his work entirely with
corn, but EAST investigated the problem in a more
wholesale way. After assembling an extensive collec-

Hybrid Vigor 163

tion of data, he gives the summarizing statement that
59 out of 85 angiosperm crosses showed a noticeable
increase in vigor. EAST of course did not continue to
investigate all of these 85 types, but concentrated upon
two representatives chiefly. Corn was selected as rep-
resenting species normally cross-fertilized in nature,
while tobacco was used as representing those species
generally self-fertilized in nature. EAST'S results with
corn need not be discussed, for they confirmed in every
point SHULL'S results. One very valuable addition,
however, was made; namely, that some crosses resulted
in relatively less hybrid vigor than others. For this re-
sult EAST developed a very significant explanation, which
has revolutionized ideas on hybrid vigor, and that is that
hybrid vigor is proportional to the number of factors in
which parents differ. This situation may be visualized
from the diagram, in connection with which certain
situations may be developed.

Parents Fi

AABBCCDDXAABBCCdd -AABBCCDd = little hybrid vigor
AABBCCDDXAABBccdd = AABBCcDd = more hybrid vigor
AABBCCDDXAAbbccdd = AABbCcDd = still more hybrid

vigor
AABB CCDD X aabbccdd = AaBbCcDd = most hybrid vigor

It is the Fj generation of course that shows the vigor,
but what index can be obtained from the germinal
formula of the F x generation as to the amount of hybrid
vigor that it will show ? It is evident that the index is
that hybrid vigor is proportional to the number of
factors in the F x generation in the heterozygous con-
dition. Thus in the first case shown in the diagram
there is only a single heterozygous set (Dd), and the

164 Plant Genetics

result is little hybrid vigor. Following down the dia-
gram it will be noted that 2, 3, and 4 of these hetero-
zygous sets show an increasing amount of hybrid vigor.
These are the facts which lie at the basis of EAST'S theory
which he calls heterozygosis. This term should not be
confused with "heterosis," which is commonly used to
express merely the fact of hybrid vigor.

We shall now consider how this conception of hetero-
zygosis explained the phenomena that SHULL had previ-
ously discovered in reference to hybrid vigor.

1. The fact of hybrid vigor. Heterozygosis claims
that hybrids are vigorous because of the heterozygous
sets.

2. The decrease in vigor after self-fertilization begins
again. The greatest loss in vigor comes between the F r
and the F 2 generations. Thereafter the loss becomes
gradually less each generation, approaching a definite
limit when no more loss in vigor occurs. Heterozygosis
explains this as follows:

AABBCCDDXaabbccdd

I
AaBbCcDd

In this case the F r generation is 100 per cent hetero-
zygous, all four sets being heterozygous, and therefore it
is very vigorous. In later generations, as is well known,
more or less homozygous sets will be split off. Intro-
ducing homozygous sets into some individuals will reduce
the aggregate heterozygous condition of the whole crop
to something less than 100 per cent; there will there-
fore be a corresponding loss in vigor. Mathematics
will show why this loss is greatest between the F x and the
2 generations. Thereafter the loss gradually approaches

Hybrid Vigor 165

the limit when the perfectly homozygous condition is
reached for the whole population, and then there can
be no more loss of vigor.

3. A cross between sister and brother effects nothing.
-This is evident, for it introduces no heterozygosity.

4. "A cross between plants belonging to two self-
fertilized families results in a progeny of as great vigor,
size, and productiveness as are possessed by families
which have never been self -fertilized/' Heterozygosis
explains this by showing that a cross between two pure
lines may bring into the hybrid a maximum number
of heterozygous sets, quite as many as are normally
present in normally cross-fertilized families.

5. Reciprocal crosses are equivalent. This obviously
follows from the theory of heterozygosis.

In connection with this work EAST was under-
taking to discover the nature of hybrid vigor. In what
respects are such hybrids vigorous? In reply to this
question EAST offers the following analysis of hybrid

it is an increase and acceleration of cell

.

division; in other words an increase in the power of
assimilation. One can early observe a slight increase
in the size of the cotyledons. The more rapid growth
and earlier maturity of the seedlings is quite noticeable.
Then one sees a distinct increase in the size of the roots.
In the stem there is no increase in the number of nodes,
but the internodal development is striking. Usually the
stem growth is greater than the leaf growth, but the
increase of the latter can be definitely traced. The size
of the flower is usually not affected, nor is there any
change in the size of small fruits, such as tobacco. In

1 66 Plant Genetics

fleshy fruits, such as tomato and eggplant, there is a
marked increase. On the individual plant there are
distinctly more flowers and fruits, and in some cases
separate inflorescences are longer, such as the ears of
corn.

EAST next studied tobacco as representing those
species which are generally self -fertilized in nature.
It is the common impression that tobacco is a striking
exception in the matter of hybrid vigor. In tobacco
crosses the hybrid progeny, instead of being more
vigorous, are frequently less vigorous than either parent.
EAST admits that there are certain cases of this kind,
but points out a number of other cases which are quite
normal. As yet there is no very satisfactory explana-
tion of the tobacco situation, and we must be satisfied
at present with the tentative conclusion that in tobacco
there appear certain unknown limitations to hybrid
vigor. In general it may be said that hybrid vigor
appears prominently in species normally cross-pollinated
in nature and less prominently in species generally self-
fertilized in nature.

In this connection it should be stated that hybrid
vigor is also manifested in some regions of the animal
kingdom. It seems, however, that it is not so general
a phenomenon among animals as among plants. In
fact, many zoologists refuse to recognize in hybrid
vigor any general law, pointing out the many cases
among animals in which hybridizing apparently results
in loss of vigor. It should be noted, therefore, that the
present discussion of hybrid vigor applies primarily to
plants and should not be extended in any general way
to the animal kingdom.

Hybrid Vigor 167

Having considered the salient facts of heterozygosis,
a statement by EAST on the "Value of heterozygosis in
evolution" is pertinent.

It can hardly be doubted that heterozygosis does aid in the
development of the mechanisms whereby flowers are cross-
fertilized. Variations must have appeared that favored cross-
fertilization. Those plants producing a cross-fertilized progeny
would have had more vigor than their self-fertilized relatives.
The crossing mechanism could then have become homozygous and
fixed, while the advantage due to cross-fertilization continued.
But was this new mechanism an advantage ? It must have been
often an advantage to the species as a whole. In competition
with other species the general vigor of those which were cross-
fertilized would aid in their survival. But the mechanism may
not have been useful in evolving real vigor in the species, because
of the survival of weak strains in combination. In self -fertilized
species, new characters that weakened the individual would have
been immediately eliminated. Only strains that stood by them-
selves, that survived on their own merits, would have been
retained. On the other hand, weak genotypes in cross-fertilized
species were retained through the vigor that they exhibited when
crossed with other genotypes. The result is, therefore, that self-
fertilized strains that have survived competition are inherently
stronger than cross-fertilized strains. On this account weak geno-
types may often be isolated from a cross-fertilized species that as
a whole is strong and hardy.

Some recent investigations have furnished striking
confirmation of the theory of heterozygosis. The work
was done originally by COLLINS and KEMPTON (i), and
later confirmed and extended by JONES (3). In brief it
is as follows. If corn sporophytes exhibit heterozygosis,
will the endosperm also show the same phenomenon ?
Endosperms, as has been stated, are genetically equiva-
lent to sporophytes in several ways. If crossing increases

168 Plant Genetic,

vigor and size of sporophytes, therefore, it might be
expected to increase the size of endosperms also.
Furthermore, endosperms have considerable advantage
over sporophytes as material for such investigation. We
say that hybrid sporophytes are more vigorous than
pure-bred sporophytes, but just how much more vigor-
ous cannot be stated with exactness. In order to
demonstrate this clearly it would be necessary to have
the hybrid and the pure-bred stock growing side by side
in exactly the same conditions, but such conditions
cannot be controlled with exactness. The environ-
mental factors affecting the size and vigor of a corn plant
are numerous, complex, and to a large extent uncon-
trollable. Thus two different plants, growing side by
side, might be in a distinctly different environment
without the fact being recognized. It cannot be
said, therefore, with much certainty that one plant
shows hybrid vigor and one does not when there are
so many other unknown factors that might affect size
and vigor. On the other hand, if it is claimed that the
endosperm of one grain shows hybrid vigor while the
endosperm of the grain next to it upon the same ear
does not, the statement would be more exact, for the
two endosperms have developed under conditions which
are unquestionably much more constant than the condi-
tions surrounding the different sporophytes in a corn-
field.

JONES selected a plant with white endosperm and
pollinated it with a mixture of its own pollen and pollen
from a yellow endosperm race. In the mature ear,
therefore, he had a mixture of yellow endosperm grains
and white endosperm grains. The former grains of

Hybrid Vigor 169

course were hybrid, since the yellow factor was intro-
duced by the foreign pollen, while the white endosperm
grains must have resulted from "own" pollen and were
homozygous. In this way JONES obtained side by side
in the same ear endosperms obviously hybrid and en-
dosperms obviously homozygous. When he weighed
these two types he found that the hybrids exceeded the
homozygotes in weight by from 5 to 35 per cent, or an
average of 20 per cent.

He then made the reciprocal cross, using a similar
mixture of yellow and white pollen on silks of the yellow
race. Of course all the resulting endosperms were
yellow, but the hybrids which had the yellow factor only
from the female side were distinctly lighter yellow than
the homozygotes which had the yellow factor from both
male and female sides. Weighing these two types JONES
obtained the same results as before, the hybrids exceeding
the others in weight by an average of 20 per cent. This
is the clearest demonstration of heterozygosis that has
yet been given, for the conditions of the experiment
were ideally constant.

The theory of heterozygosis accounting for the
phenomenon of hybrid vigor is the one most generally
accepted. Its claim is that hybrid vigor appears in
proportion to the number of factors in which the parents
of the cross differed.

This claim may be considered briefly. Is hetero-
zygosis really an explanation of the phenomenon of
hybrid vigor? It seems obvious that it is not. It
was known that hybrids are vigorous because they are
hybrids. Heterozygosis states that hybrids are vigorous
to the degree that their parents differed in hereditary

170 Plant Genetics

factors; in other words, this is merely a statement that
hybrids are vigorous because they are hybrids, with
the addition that the more hybrid a hybrid is the more
vigorous it is. It follows therefore that heterozygosis is
not an explanation of hybrid vigor, but merely a restate-
ment of the phenomenon in Mendelian terms, with the
additional idea that there may be various degrees of
hybrid vigor. It is not the intention to discredit
heterozygosis as a valuable conception, but merely to
point out that it is not a real explanation, merely a more
intelligent statement of the facts. Restatements should
not be confused with explanations.

A suggested explanation may be considered. In
nature a " struggle for existence " occurs among species
and individuals. There occurs also a struggle for ex-
istence among unit characters. If a unit character
is undesirable it is .eliminated, for the individual or
species that carries it is eliminated. The unit charac-
ters, therefore, that have survived and appear in the
plants of today are for the most part desirable ones,
although some undesirable ones also may have survived,
having been carried through in association with the
desirable characters. The majority of unit characters
today, however, may certainly be regarded as desirable
ones, and a majority is sufficient for our present con-
sideration.

The question may be raised as to what constitutes a
desirable character. It nlay be any one of a number of
things, but is there not some feature common to all
desirable characters ? The common feature of all de-
sirable characters would seem to be vigor. Each desir-
able character must add somewhat to the vigor of

Hybrid Vigor 171

the plant that contains it, and if vigor is increased such
things as size and productiveness will also be increased.
Those plants, therefore, will be most vigorous which have
in combination the greatest number of desirable char-
acters, and it is obvious that the plants which have the
greatest combination of such characters are hybrids.

A diagram similar to that which was used to explain
heterozygosis may be considered:

p. P. F

AABBCCDDXAABBCCdd =AABBCCDd -little hybrid vigor
AABBCCDDXAAbbccdd =AABbCcDd~ = still more hybrid

vigor

In that explanation it was stated that the first case
showed little hybrid vigor because it had only one
heterozygous set (Del), while the other case showed
more hybrid vigor because it had three such hetero-
zygous sets. Hybrid vigor, therefore, appeared in
proportion to the number of heterozygous sets in the
hybrid. This diagram served the purpose in explaining
heterozygosis, but it will now be discarded because
it does not represent the most important result when
two races are crossed. The important result is repre-
sented in the following diagram:

Px-AABBCCddeeff *' , i -,

=more hybnd Vlgor

P,-AABBccddeeff \ ATJ , , , ., .

Pa-aabbCCDDeeff JAaBbCcDdeeff = less hybrid vigor

The thought is that in each of these two cases the
hybrid is more vigorous than either parent, not because
it contains more heterozygous sets but because it
contains more positive, dominant factors. For example,

172 Plant Genetics

in the first case each parent contains three factors, the
small letters representing merely the absence of factors.
The F! generation, therefore, contains six factors, and
for this reason is more vigorous than either parent. It
is stated in the diagram that in the first case there is
"more hybrid vigor " and in the second case "less
hybrid vigor," simply because hybrid vigor is a relative
term. It represents merely how much more vigorous
the hybrid is than either parent. In the first case the
parents have three factors and the hybrid six, and the
increase is three, which measures the amount of hybrid
vigor. In the second case each parent has two and the
hybrid four; the increase therefore is only two, and for
this reason there is less hybrid vigor in the second
case than in the first.

Assuming that the majority of factors are desirable,
and the desirable factors make for general vigor, it would
follow that the most vigorous plant will be the one
containing the greatest number of positive factors.
It has been shown that plants that contain the greatest
number of factors are hybrids, and for this reason
hybrids are more vigorous.

The following question may be raised. JTf it is
granted that all desirable factors tend somewhat to
increase the general vigor, do they all do this to the
same degree? The obvious answer is in the negative,
and this has no bearing upon the validity of the explana-
tion. On the other hand, if heterozygosis be accepted
for an explanation the question would present a diffi-
culty. Heterozygosis would claim that Aa induces
vigor, not because of any particular factor that it repre-
sents, but because it is a heterozygous set. It seems to

Hybrid Vigor 173

be more reasonable and accurate to suppose that certain
factors induce more vigor than others. EAST recognizes
this in the following statement: "This stimulus to
development is cumulative up to a limiting point, and
varies directly with the number of heterozygous factors
in the organism, although some factors may have
more powerful action than others."

It is obvious that the suggestion made above is that
of a real explanation of hybrid vigor and not merely a
restatement. It is rather an obvious explanation that
has probably occurred to a number of geneticists.
KEEBLE and PELLEW (5) suggested it in 1910, and
since that time it has been somewhat discussed in the
literature, being referred to as "the hypothesis of
dominance (accounting for hybrid vigor).' 7 At first
statement the theory seems sound, but actually it does
not fit the facts. The two chief objections to this
theory of dominance may be found in the publications
of SHULL, EMERSON, and EAST (2).

, f>. If hybrid vigor were due to dominance it would be
possible in generations subsequent to the F 2 to recom-
bine in one race all of the dominant factors. Thus
there could be isolated a race that was " 100 per cent
vigorous," and since it would be homozygous its vigor
would not be lost by inbreeding. Actually, though,
hybrid vigor cannot be fixed in this way; "all maize
varieties lose vigor when inbred."

2.) Experience assures us that the distribution of
individuals in the F 2 generation with reference to hybrid
vigor is represented graphically by a symmetrical curve,
similar to the normal probabilities curve; the class
containing the greatest number of individuals is that

174 Plant Genetics

which shows the medium amount of hybrid vigor,
while on either side of this class the fall in the curve is
regular, reaching its lowest points in the two small
extreme classes which show respectively greatest hybrid
vigor and least hybrid vigor. According to the domi-
nance hypothesis the largest class of the F 2 individuals
would be that showing greatest hybrid vigor, while
the smallest class would be that showing least hybrid
vigor. The curve representing such a situation would
be unsymmetrical and strikingly different from that
which actually occurs.

For these two reasons the dominance hypothesis
seems to have been discarded. Although it is theoreti-
cally attractive its failure to satisfy these two important
details of the hybrid vigor situation has condemned it.

Very recently JONES (4) has ingeniously modified the
dominance hypothesis so as to avoid these difficulties.
At first consideration his theory seems to be clearly the
most reasonable explanation of hybrid vigor that has
yet been presented, although in time it may suffer
from destructive criticism. The argument is essentially
the same as that for the old dominance hypothesis, with
the following important modification. Under the old
hypothesis it was stated that in the cross A Abb XaaBB =
AaBb the hybrid showed vigor because it combined
the two dominant determiners AB. This met with the
practical objections mentioned above. JONES visualizes
the situation as represented in fig. 38. In this case it is
a question of linkage. The vigor of one parent is due
to the two dominant determiners A and D, while that
of the other parent is due to the two dominant deter-
miners C and B. The hybrid is more vigorous than

Hybrid Vigor

175

either because it combines all four dominant determiners.
The attractiveness of this scheme is that it escapes the
objections that were made to the older dominance
hypothesis.

1. The fact that 100 per cent hybrid vigor cannot be
fixed is quite in accordance with JONES'S scheme, for
it is obviously impossible to isolate a homozygous race
with all four factors A, B, C, and D.

2. A simple mathematical demonstration will show
that the distribution of F 2 individuals is quite what it

FIG. 38. Diagram to aid in visualizing JONES'S explanation of
hybrid vigor by dominance of linked factors.

should be, represented by a symmetrical curve similar
to the curve of probabilities.

In fact this new theory, "the dominance of linked
factors/ 7 seems altogether sound. We should reason-
ably expect that each chromosome would contain one or
more dominant determiners (conducive to vigor) linked
with one or more recessives.

The discussion has left the impression that hybrid
vigor is to be explained by dominance (and linkage). In
this day of factors and determiners such a hypothesis is
quite appropriate. It may be, however, that in the

176 Flant Genetics

future such a phenomenon as hybrid vigor may be
explained in terms of the stabilities and reactivities of
the constituents of specific protoplasts.

In conclusion, attention should be called to the danger
of confusing phenomena of hybrid vigor with those of
cumulative factors. The two situations are somewhat
similar, but careful consideration will discover distinct
differences. It is perhaps unsatisfactory to assume
absolute lack of dominance in the case of cumulative
factors and the essential presence of dominance (partial
at least) in explaining hybrid vigor. Such assumptions,
however, are not necessarily contradictory. The two
things should be recognized as distinct phenomena,
although they are at work simultaneously and probably
interact to give complex results.

i. COLLINS, G. N., and KEMPTON, J. H., Effects of cross-pollination
on the size of seed in maize. U.S. Dept. Agric. Circular 124.

r ^3-

7 EAST, E. M., and HAYES, H. K., Heterozygosis in evolution and
in plant breeding. U.S. Dept. Agric., Bur. PL Ind. Bull. 243.
pp. 68. pis. 8. 1912.

3. JONES, D. F. ; Bearing of heterosis upon double fertilization.
Bot. Gaz. 65:324-333. figs. j. 1918.

4. , Dominance of linked factors as a means of accounting

for heterosis. Genetics 2:466-479. 1917.

5. KEEBLE, F., and PELLEW, C., The mode of inheritance of
stature and of time of flowering in peas (Pisum sativum). Ge-
netics 1:47-56. 1910.

6. SHULL, G. H., Hybridization methods in corn breeding. Amer.
Breeders Mag. 1:98-107. 1910.

CHAPTER XVII

SEX DETERMINATION

The subject of sex determination must be included in
any consideration of genetics, but it is extensive enough
as measured by the amount of investigation to form the
subject of a book. In consequence in this connection it
can only.be given brief consideration. The current the-
ories of sex determination fall naturally into two general
categories :

1. Many biologists believe that sex is predetermined
by the chromosome equipment that enters into the
zygote.

2. Other biologists believe that sex may be deter-
mined otherwise than by chromosomes; that it may be
determined at fertilization, or before, or even after-
ward; in any event, that it is not the chromosomes
that determine sex but some general physiological
condition.

These are the two schools in regard to sex determina-
tion; in fact, each school includes a number of theories,
but the fundamental idea in each group is the same.
These two general views will be referred to as the
Chromosome theories and the physiological theories.
Representative examples of each will be given first, and
since sex determination has been investigated more
thoroughly and for a much longer time in animals than in
plants the illustrations will be taken from animals.

177

178

Plant Genetics

CHROMOSOME THEORIES. A classic example of the
simplest kind is found in the nematode worms. Fig. 39
will indicate how sex is determined in this case. The
mechanism is obvious. Both male and female have ten

MALE

FEMALE

MATURE INDIVIDUALS

ooo o

> O

ooo o

Zygotes

oo oo

> >

oooo

(This type of zygote
produces a male)

(This type of zygote
produces a female)

FIG. 39. Illustrating behavior of sex chromosomes

chromosomes to determine most of their somatic char-
acters; but in addition there are extra chromosomes that
determine sex, known variously as sex chromosomes,
^chromosomes , and heterochromosomes. In this case the
male contains only one sex chromosome while the female
contains twg. At the reduction division, connected
with gamete formation, the chromosome equipment

Sex Determination 179

is reduced one-half. It is obvious that in the female
each egg receives one sex chromosome, and therefore
all eggs are alike in this feature. In the male, however,
with one sex chromosome, at the reduction division
the solitary sex chromosome goes to one pole, leaving
the other pole without such a chromosome. __A a
result there are two kinds of sperms, one half containing
a sex chromosome, the other half containing none. At
fertilization, if an egg mates with a sperm having a sex
chromosome the zygote ^contains two, and this results
obviously in a female, for females are characterized by
two sex chromosomes. -With a sperm of the other type,
the zygote receives only one sex chromosome and must
produce a male individual. As a result males and
females are produced in equal numbers, and sex is
determined by the type of sperm that enters into the
sex fusion.

In such cases females are homozygous for sex and
males are heterozygous, for in the male the sex chromo-
some is paired with its absence. This is the commonest
situation, although it would be theoretically possible to
have females heterozygous for sex, having the single
chromosome, while the male had a pair. In this case
sex would be determined by the type of egg that entered
into the sex fusion. In addition, there are cases where
one sex has one sex chromosome and the other sex has
none. Finally, there are casjes in which one sex has a
pair of large chromosomes, while the other sex has one
large and one small sex chromosome. In all of these
cases, however, the fundamental mechanism is the same.
Either male or female (commonly the former) is hetero-
zygous for sex, producing two kinds of gametes, as

180 Plant Genetics

shown by the chromosome situation. Sex, therefore,
is_predetermined by the kind of gamete that enters into
the sex fusion.

This statement represents the chromosome theories
of sex determination, according to which it seems evident
that sex is absolutely predetermined. Outside conditions
can effect no_ change, and therefore sex cannot be con-
trolled. It results, therefore, with mathematical pre-
cision in 50 per cent males and 50 per cent females, just
as a Mendelian ratio comes out 3:1. , t
{ PHYSIOLOGICAL THEORIES. In 1906 RICHARD HERT-
WEG (5) performed some sex determination experiments
with frogs. The eggs are laid free in the water before
fertilization, so that they furnish unusually good material
for such experiments. Normally the eggs are fertilized
very soon after they are laid, with the result that the
progeny consists of approximately 50 per cent males and
50 per cent females. HERTWEG took some of these
eggs and allowed them to overripen before fertilization
took place ; that is, he put aside some eggs as soon as they
were laid and allowed them to remain unfertilized for an
unusually long period. While these eggs were standing
in the water he found that they absorbed an unusual
amount of water, and the obvious conclusion was that
very ripe eggs show higlj water content. He then
allowed these overripe eggs to be fertilized, and the
resulting progeny wei^ iQo,.per ^ c^t_males. His con-
clusion was that sex was not determined by the chromo-
some equipment, but by the physiological conditions in
the egg, high water content resulting in males.

This theory was confirmed in a striking way in
1912 by Miss KING (7), who performed the converse of

Sex Determination 181

HERTWEG'S experiments, using toads' eggs. Taking some
newly laid eggs she withdrew water from them, then
allowed them to be fertilized, and the resulting progeny
were 90 per cent females. The obvious conclusion is that
eggs with low water content produce females.

Finally, there is the remarkable work of PJJDDLE (n)
with pigeons. HERTWEG and Miss KING had found
that sex is determined by the physiological factor of
water content. RIDDLE has investigated a little more
critically, analyzing the physiological conditions of both
male and female, and gives the following contrast:

Male Female

High percentage of water Low percentage of water

Low percentage of fat High percentage of fat

Low percentage of phosphorus High percentage of phosphorus

High metabolism Low metabolism

It appears from this that high water, low fat, and
low phosphorus are male attributes or conditions,
while the female attributes are the reverse. The main
feature of difference, however, to which the other con-
trasting conditions are subordinate is that the male
shows high metabolism and the female low metabolism.
The idea is that any physiological conditions that
affect water content or fat content or phosphorous con-
tent in -the egg will affect the sex in the resulting
progeny.

Following these ideas, RIDDLE was able to^control
sex by various means; furthermore, he makes the
somewhat startling statement that sex is a quantitative
phenomenon; that is, the difference between male and
female is a difference in degree only. A diagram
(fig. 40) will illustrate the situation. It represents a

i8 2 Plant jpenetics

graduated scale based on relative water, fat, and phos-
phorous content. The egg may be at any point on the
scale, and the sex of the individual produced by the
egg will depend upon its position on the scale. An
egg in any position to the left of the middle results in a
male and to the right in a female. It should be noted,
however, that if the egg is near one of the extremes the
progeny will be either a very masculine male or a very
feminine female; while if the egg lies near the middle
point, on one side or the- other, the progeny will be a

High H 2 I.ow H,0

|^ w p a{ Very Male Some Female Some Male Very female jj i p flt

Males with Females with

Very Male

Some Female Some Male

Very Female

Males

Characteristics Characteristics

Females

V

\/

^

Low P ^_ ^i* ^ ^_Jt- " lgh P

High MALES FEMALES Low

Metabolism Metabolism!

FIG. 40. Illustrating RIDDLK'S idea of sex. Sexes differ only
quantitatively, and it is possible to find various degrees of maleness
and femaleness at different points along the scale.

male with some female characteristics or a female with
some male characteristics; in other words, a feminine
male or a masculine female. In fact, RIDDLE is actually
able to bring this about, obtaining at will males with all
degrees of maleness, etc. He even suggested at one
time that when we learn more about this method of sex
control we might be able to eliminate from human
society such undesirables as the effeminate man or the
masculine woman.

There are three possible conclusions in reference to
these contradictory theories: i) an acceptance of the
one and rejection of the other; (2) the claim that both
amount to the same thing, that is, that they express the

Sex Determination 183

same fundamental facts in different terms or by the use of
different indices; (3) the claim that both are true but
cover different territories; that one of them explains
certain types of cases and the other explains other types
of cases. At present the third alternative seems to be
the most reasonable and least difficult of explanation.

In view of the evidence on both sides one can hardly
say that either the chromosome theory or the physio-
logical theory is absolutely wrong. Again, to interpret
either theory in terms of the other is theoretically very
difficult. The cytologists certainly present a striking
demonstration of the chromosome theory, showing
the sex chromosomes and the constant occurrence of the
proper numbers in a given sex. They also show the
machine at work with its inevitable sex product of 50-50;
and of course they are logical in concluding that sex
cannot be controlled.

On the Bother hand, the physiologists show just as
clearly that sex can be controlled artificially; by chan-
ging the chemistry or physics of the situation they change
the sex. This of course is cytologically unorthodox;
and as a consequence some cytologists have attempted
to clear up the situation by the following claim. Granted
that certain chemical changes may change the sex,
the result is clue to the fact that the mitotic figures are
disturbed and the sex chromosome situation changed.
It is always the sex chromosomes that determine sex,
but they may rearrange themselves in consequence of
abnormal chemical stimuli. Such an explanation seems
like the last resort. One might imagine that the nucleus
might cast out a sex chromosome and thus change an
expected female into .a male ; but to make the change in

184 Plant Genetics

the other direction the nucleus would have to manu-
facture an additional sex chromosome, which is difficult
to imagine. It seems safe to conclude, therefore, that
physiological conditions determine sex in some cases and
sex chromosomes determine sex in other cases.

With this background we Will consider briefly such
work as has been done with plants. In the first place,
the sex chromosome has not been established for
plants. SYKES, DARLING, STRASBURGER, and others
have searched for it diligently but with ^negative results.
It is true that certain forms show odd chromosomes,
such as Ginkgo, Galtonia, Melandrium, Fagopyrum, and
Houstonia, but they have never been satisfactorily
connected with sex. The most hopeful work has been
done by^ALLEN (i) with Sphaerocarpus, which will be
mentioned later. It must not be concluded, however,
that the chromosome theory of sex determination does
not hold for plants. Many cytologists fully expect
sex chromosomes to be demonstrated in plants, and at
present they point out certain indirect indications of
the chromosome situation, such as the even separation
of sexes at the reduction division. The theory therefore
is firmly believed by many plant cytologists, even
though the sex chromosome situation has been estab-
lished in no plant. Attention may be called to the
fact, however, that plants differ from animals in ways
that present obstacles to the chromosome theories of
sex. Without discussing these obstacles fully atten-
tion may be called to two of them.

In the first place, hermaphrpditism is very common
in plants, and this is difficult to explain in terms of the
sex chromosome theory. In the second place, all the

Sex Determination i8c

higher plants exhibit alternation of generations, a phe-
nomenon practically unrepresented in animals. In at-
tempting to apply the sex chromosome theory througt
the life-history of such a plant, with its alternation ol
generations and heterospory, great theoretical difficulties
are encountered. In general, therefore, we reach th(
conclusion that the sex chromosome theory has not been
established in plants and that certain obstacles seem
to stand in the way of its establishment.

On the other hand, no very general physiological
theory has been established; in short, sex determination
in plants is not yet well understood. A few suggestive
facts are known, however, and when more facts are
available some general theories doubtless will be devel-
oped. A general survey of the plant kingdom will show
the suggestive facts that are available.

ALGAE. No conclusive work has been done with this
group, but some of the experimental work of KLEBS (8)
suggests that sex may be controlled in algae.

FUNGI. Mention will be made only of the work
of BLAKESLEE and BURGEFF. In his experimental
work with Mucor, BLAKESLEE (2) found in general
three different sexual types of mycelia, two of which he
called plus and minus strains. Although they looked
alike in every particular, he concluded that they were
sexually different for the following^reason. Neither
strain by itself is capable of producing zygotes, but
when plus and minus strains are brought together sexual
branches from the one meet sexual branches from the
other and produce abundant zygotes. The natural
conclusion is that BLAKESLEE 's plus and minus strairis
represent the male and female conditions, although the

186 Plant Genetics

sex cannot be distinguished. The third type of my-
celium he called the neutral strain. Although this
cannot produce zygotes by itself it does produce them
when mated with either a plus or minus strain. Evi-
dently, therefore, the sex branches from the neutral
strain contain both male and female potentialities.

The answer to the question as to where and how sex
is determined in these forms involves a number of
possibilities. When a zygote germinates, one or more
sporangia are produced jyory early, and this raises the
question as to the sexes developed. In Mucor itself
the segregation of sex is evidently completed before the
formation of the spores in the sporangium, for all the
spores in a given sporangium are of a given strain,
producing either male or female mycelia. The sporan-
gium as a whole, therefore, is either male or female. In
Phycomyces, however, a different behavior is regular.
Tife zygote produces a sporangium, but the sporangium
is not completely of one sex. It produces three types of
spores: spores producing the plus strain, spores pro-
ducing the minus strain, and spores producing the
neutral strain. The plus strain then perpetuates
only plus strains through its spores, which means that
the sex is fixed in this case. The minus strain behaves
in the same way. The neutral strain, however, pro-
duces spores of all three types, an interesting situation,
for it suggests the Mendelian segregation. The Mucor
group deserves to be studied further for suggestions as
to sex determination.

The work of BURGEFF (3) is brief but striking. He
took BLAKESLEE'S plus and minus strains and found
that they always bred true to their plus and minus

Sex Determination 187

conditions. Then by using some remarkable technique
he grafted parts of one strain on to the other, and the
result was a neutral strain. This seems surprising for the
moment, but when it is remembered that Mucor is coeno-
cytic the result is what might be expected, for the two
sets of nuclei can mingle freely.

^LIVERWORTS. A great deal of experimental work
has been clone with this group, but only one conclusive
result has been obtained. In dioecious liverworts, as
Marchantia, the sexes appear in about equal numbers.
It is natural for cytologists to conclude from this that
sex is determined at the reduction division in the forma-
tion of spores. Thus when a mother-cell produces a
tetrad, two of the spores would produce male gameto-
phytes and the other two female gametophytes. This
of course was an assumption, for no sex chromosome was
disco vereel, nor was it known that the spores of a tetrad
always divide in this 50-50 way as to sex.

The assumption was strengthened by some work by
NOLL (10). He found that gemmae from male game-
tophytes always produce male gametophytes, while
gemmae from female gametophytes always produce
females. This established the fact that in the mature
gametophyte sex has been fixed, but it did not indicate
where it had been fixed.

It remained for STRASBURGER (13) to answer this
question. He selected Sphaerocarpus, which is pecul-
iarly favorable material for such work. It is^dioecious,
like many liverworts, but a remarkable feature is that
the spores hang together in the tetrael. Orelinarily
when spores are mature the tetrads are no longer dis-
tinguishable. Sowing such free spores one may get the

1 88 Plant Genetics

50-50 ratio of males and females, but it cannot be
inferred that the sexes are evenly divided in every
tetrad; it may have been only an equal division in
the capsule as a whole. If, for example, the upper
half of a capsule produced all females and the lower
half all males, sex would not have been determined at
the reduction division, where, according to the sex
chromosome theory, it is determined. S phaerocar pus ,
therefore, gave STRAS BURGER an opportunity to deter-
mine this situation, for he was able to isolate mature
individual tetrads, the four spores hanging together.
He sowed these tetrads in separate pots, obtaining four
gametophytes in each case. The experiment was con-
ducted upon rather a large scale, and the results were
strikingly constant. Two male and two female game-
tophytes were always produced from a tetrad. This
certainly established the fact that in this case sexes
were evenly separated at the reduction division, and
indicated strongly that the sex chromosome theory
applies in this group. What these facts established for
Sphaerocarpus has been taken for granted for the other
dioecious liverworts.

Recently J^LLEW (i) has claimed to have demon-
strated the x chromosome in Sphaerocarpus. He says
that, one large chromosome, exceeding in length and
thickness the other chromosomes, characterizes the cells
of the female gametophyte, while the cells of the male
gametophyte are characterized by one very small
chromosome. His investigation seemed to show that in
spore formation two of the spores of the tetrad receive
the large chromosome, and the other two the small
chromosome. These chromosomes have not been defi-

Sex Determination 189

nitely connected with sex determination, but the situa-
tion is most suggestive.

The sex chromosome theory cannot apply, of course,
to the monoecious liverworts, where the sexes must be
separated at a much later period than the reduction
division. In these cases obviously every spore contains
the potentialities for both sexes, and the two are sepa-
rated in connection with some of the vegetative divisions
of the bisexual gametophyte. In some species of
Riccia we know that a developing gametophyte pro-
duces only antheridia for a time and at a later period
produces archegonia only. The structural basis of this
change is not known nor the exact point at which it
occurs.

MOSSES. A single very significant piece of work
by MARCHAL (9) will be mentioned. Fjtnaria is a
dioecious moss, and hence it may be assumed that the
sexes are separated at the reduction division in the
formation of spores. Each spore carries the potentiali-
ties for one sex only; but of course the sporophyte as
a whole before the reduction division must carry the
potentialities for both sexes. MARCHAL, by a peculiar
technique of his own, clipped a fragment from a young
sporophyte and induced it to reproduce aposporously;
that is, the sporophyte fragment produced a gameto-
phyte directly. The fragment must have contained the
potentialities for both sexes, since it consisted of tissue
in which reduction division had not yet occurred.
Logically the gametophyte should be bisexual, pro-
ducing both antheridia and archegonia, and this was the
result obtained. It is quite in accord with the sex
chromosome theory and is a striking confirmation of it.

190 Plant Genetics

PTERIDOPHYTES. In this group we are confronted
by a new complexity. In the preceding groups it was
necessary to consider separately monoecious and dioe-
cious forms, assuming that in the latter the sexes are
separated at the reduction division. In the higher
Pteridophytes, however, heterospory appears, and
establishes once for all the dioecious condition in the
gametophyte generation. In heterosporous forms there-
fore the gametophytes are always dioecious, but it can-
not be said that the sexes are separated at the reduction
division. Obviously no such claim can be made, for it
would have to assume that each tetrad consists of two
megaspores and two microspores, a thing which prob-
ably never occurs, unless possibly in Equisetum, which
should be investigated. It is safe to conclude therefore
that sex in heterosporous forms is differentiated before
the reduction division. A megasporangium produces
only megaspores, although in doing so the reduction
division occurs. In Selaginella, which is perhaps the
best example, one can distinguish a megasporangium
from a microsporangium as soon as the mother-cells
are formed. Clearly the sexes are separated before
the reduction division. In attempting to explain this
situation cytologists claim that the sexes are separated
at one of the vegetative divisions of the sporophyte
previous to the formation of mother- cells. This is
merely an assumption and an extremely unsatisfactory
one, for only in connection with the reduction division
is there a proper mechanism for the separation of sexes
according to the sex chromosome theory. Furthermore,
at present there is just as good reason to believe that sex
in Selaginella is determined by physiological conditions.

Sex Determination 191

Before leaving Pteridophytes the homosporous
forms, notably the true ferns, should be considered
briefly. In the ordinary bisexual fern prothallium,
antheridia and archegonia appear, as was described for
Riccia. A young prothallium can produce antheridia
only, while a mature prothallium produces archegonia
only. These facts are interpreted in two ways. From
the physiological standpoint it is claimed that more
nutrition is required for the production of archegonia
and eggs than for the production of antheridia and
sperms. Young prothallia, therefore, have not the
nutritive capacity to produce archegonia and can pro-
duce antheridia only. Mature prothallia, of course, have
a greater nutritive capacity and can produce archegonia.
This nutrition theory has now considerable vogue and is
applied not only to ferns but also to other plant groups.

Another interpretation is given by cytologists.
They claim that the sexes are separated by some sex
chromosome mechanism at some one of the vegetative
divisions midway in the life of the prothallium. This
claim is difficult to accept on account of what it involves.
In accordance with this view a cell division occurs, so
that thereafter only archegonia will be produced; while
before this division only antheridia were produced.
It is hard to understand how a cell division can affect
the sex of its ancestors, the cells that came . before.
This difficulty is avoided by saying that a sex chromo-
some whose presence determines maleness drops out
of existence at this intermediate cell division, so that
only archegonia could be produced thereafter. The
physiological claim certainly seems to be the more
reasonable one in this instance.

192 Plant Genetics

There is an apparently easy experiment which should
test these claims and decide between them, but as yet
it has not been performed. Fern prothallia are known
at times to produce proliferations which reproduce the
parent prothallium vegetatively. If a proliferation
could be produced on the archegonium-bearing region of
a mature prothallium its sexual behavior should throw
much light upon this problem. According to the physio-
logical claim, this proliferation should first produce
antheridia, for it will pass through a young stage of low
nutritive capacity. According to the cytological claim
the cells from which the proliferation arose have female
potentialities only, and therefore the proliferation should
never produce antheridia but only archegonia.

SEED. PLANTS. Most of the work on sex determina-
tion has been carried on with seed plants. Experiments
have been performed to show the effect of manuring,
light, temperature, and other environmental factors
on the sex ratio. The results have been both positive
and negative. Some conclude that sex is rigidly pre-
determined in the seed, while others claim that it is
determined by the later environmental conditions.
The general results are not conclusive, so that they may
be passed by with this brief statement.

There are certain theories of sex determination,
however, with which every botanist should be familiar,
for they are an important part of botanical literature.
They may be stated briefly.

ORRENS (4) thought that sex was determined at the
reduction division of the pollen mother-cell. Two pol-
len grains are produced having female tendencies and
two having male tendencies. The sex of the progeny,

Sex Determination 193

therefore, depends upon which kind of pollen grain
functions. Such an idea, of course, would be quite in
keeping with a sex chromosome theory, where the male
is the heterozygote for sex.

STRASBURGER (13) took exception to this view. He
thought that the phylogenetic development of the seed
plants had completely separated the male and female
tendencies through heterospory. As a consequence, he
would regard CORRENS' idea that half the pollen grains
have female tendencies as impossible. STRASBURGER'S
idea is that the pollen mother-cell develops pollen
grains with stronger and weaker male tendencies, while
the megaspore mother-cell develops eggs with stronger
and weaker female tendencies. It is therefore the
algebraic sum of the two as they meet in fertilization
that determines the sex of the progeny. If a pollen
grain with strong male tendencies fuses with an egg
with weak female tendencies the progeny will be male,
and similarly for the other combination. This flavors
of RIDDLE'S work. Just what would happen if two
strong male and female tendencies should meet would
be interesting.

STRASBURGER'S strongest evidence for this theory is
his classic experiment with Mercurialis annua, with
which every botanist should be familiar. Mercurialis
may be called imperfectly dioecious. Certain plants are
prevailingly female but bear a few weak male flowers.
It would seem as though the female tendencies of such a
plant are stronger than the male. When such a plant is
inbred, using pollen from the weak male flowers on the
stigmas of the strong female flowers, the resulting
progeny is 100 per cent female, which is in accordance

194 Plant Genetics

with the theory. Other plants are prevailingly male but
bear a few weak female flowers, and inbreeding in this
case results in TOO per cent males. Finally, there are oc-
casional plants which are evenly monoecious; that is, half
of their flowers are strong males and the other half strong
females. From inbreeding such a plant, the resulting
progeny is 50 per cent male and 50 per cent female. It
is obvious that from such results STRASBURGER would be
convinced of his theory of male and female tendencies.

Such a theory involves too much of RIDDLE'S quanti-
tative idea to be acceptable to cytologists. They seem
confident that sex determiners will be demonstrated
eventually as distinct physical entities and that such
phrases as "male and female tendencies" will disappear.

It will be interesting to apply STRASBURGER'S theory
to cases of parthenogenesis and vegetative apogamy and
discover the difficulties encountered. In the first place,
in monoecious forms it is found that a parthenogenetic
egg produces a monoecious plant. Should it not produce
a strictly female plant, because eggs contain only
female tendencies ? STRASBURGER avoids this difficulty
by saying that these parthenogenetic eggs are diploid;
hence they are not truly eggs but merely sporophyte
cells. The process of producing the new plant, therefore,
is merely vegetative multiplication, which would natu-
rally perpetuate the tendencies of the parent body. The
parent was monoecious, and quite properly the progeny
of a diploid egg should be monoecious.

STRASBURGER used this same situation to confirm
his theory. In dioecious forms parthenogenetic eggs
should produce only female plants. As a matter of fact
this is true, as STRASBURGER shows by a number of

Sex Determination 195

examples. The dioecious Thalictrum, however, behaves
differently in cases of parthenogenesis. The progeny
are sometimes female, but just as often they are male
This situation deserves further investigation. There
is no question but that STRASBURGER'S theory is in-
genious, reasonable, and promising.

A striking bit of work on sex determination was done
by CIESIELSKY (14) of Lemberg, who worked with
Cannabis sativa, the Indian hemp. When he used pollen
fresh from the anthers the progeny were 85 per cent
males. When he took pollen from the anthers and
allowed it to stand for twelve hours or more before
using it the resulting progeny were 92 per cent females.
The conclusion is obvious that fresh pollen results in
males and stale pollen results in females. This can be
combineel with STRASBURGER'S idea, for one might
expect fresh pollen to show stronger male tenelencies
than stale pollen. CJESIELSKY'S work deserves to be
repeated upon a more extensive scale.

Probably the most unorthodox behavior of sex is
that which was rather recently described in the Papaya
(6). A few cases were describeel in which male Papaya
trees had been cut back. The whole head of the tree
had been cut off, eliminating all of the male flowers.
When new branches put up from the cut trunk they
bore in some cases hermaphrodite anel in other cases
prevailingly female flowers. That sex could be so com-
pletely changed by such an operation was incredible.
Further examination showed that this result was not
always obtained and that in fact there were only a few
such cases on record. Most botanists today doubt the
validity of this reporteel behavior.

196 Plant Genetics

We have a suggestion of the same sort of behavior in
the case of begonias. Gardeners claim that in Begonia
one can secure female flowers by pinching off the male
inflorescences, female inflorescences developing later in
the same place. One gardener, at least, has been prac-
ticing this remarkable process for some years.

In concluding this brief survey of sex determination
in plants it is obvious that some of the evidence points
to the sex chromosome theory, while other evidence
points to some physiological theory. It is perhaps best
at this time to conclude that one theory holds for
certain cases and the other theory for other cases.

This topic should not be left without mention of the
classic experiments of CORRENS and SHULL. CORRENS
(4) crossed the dioecious Bryonia dioica with the her-
maphroditic B. alba, while SHULL (12) crossed Lychnis
dioica with hermaphroditic mutants from the same.
The sexual behavior of the progenies in the two cases
was not identical, but both suggested a sex chromosome
theory, with the male heterozygous for sex. The theo-
retical explanations are so complex and dubious that they
cannot be conveniently discussed here, but the experi-
ments themselves represent a very significant type of
work.

1. ALLEN, CHARLES E., A chromosome difference correlated with
sex differences. Science 46:466-467. 1917.

2. BLAKESLEE, A. F., Sexual reproduction in the Mucorineae.
Proc. Amer. Acad. 40:205-319. 1904.

, Differentiation of sex in thallus gametophyte and

sporophyte. Bot. Gaz. 42:161-178. 1906.

3. BURGEFF, H., Uber Sexualitat, Variabilitat, und Verebung
bei Phycomyces nitens. Ber. Deutsch. Bot. Gesell. 30:679-
685. 1912.

Sex Determination 197

4. CORRENS, C., and GOLDSCHMIDT, R., Die Verebung u. Be-
stimmung des Geschlcchts. Berlin. 1913.

5. HERTWEG, R., Verhandl. Deutsch. Zool. Gesell. 1906; see
also Biol. Centralbl. 32:1. 1912.

6. HIGGINS, J. E., Growing melons on trees. Jour. Heredity 7:
208-220. figs. 7. 1916.

7. KING, H. D., Jour. Exp. Zool. 12:19. 1912.

8. KLEBS, G., Uber Probleme der Entwicklung. Biol. Centralbl.
24:8-9, 14-15? 257-267, 289-305, 444-614. 1904.

9. MARCHAL, EL. et EM., Aposporie et sexualite chez les Mousses.
I, II, III. Bull. Acad. Roy. Belgique. Cl. Sci. 1907. 765-789;
1909. 1249-1288; 1911. 750-778.

10. NOLL, FR., Vorlaufiger Abschluss der Versuche liber die
Bestimmung des Geschlechts bei diocischen Pflanzen. Sitz-
ungsber. Niederrh. Gesell. Bonn. Naturwiss. Abt. 1907.
S. 68.

11. RIDDLE, OSCAR, The control of the sex ratio. Jour. Wash
Acad. Sci. 7:319-356. 1917.

12. SHULL, G. H., Reversible sex mutants in Lychnis dioica. Bot.
Gaz. 52:329-368. 1911.

13. STRASBURGER, E., Versuche mit diocischen Pflanzen in
Riichsicht auf Geschlechtsverteilung. Biol. Centralbl. 20:
657. 1900.

, tJber geschlechtbestimmende Ursachem. Jahrb.

Wiss. Bot. 48:427-520. 1910.

14. WESTER, P. J., The determination of sex. Jour. Heredity
5:207-208. 1914.

CHAPTER XVIII
THE BEARERS OF HEREDITARY CHARACTERS

The conclusion reached in the preceding chapter was
that _sex may be determined by chromosomes, but that it
also may be determined by physiological conditions.
To the geneticists this suggests the following question:
If the sex character may be determined by physiological
conditions, why may not other plant characters be
determined in the same way ? In other words, this
opens up the general question of the chromosome
hypothesis in reference to inheritance. It has been
taken for granted in the preceding discussion that
chromosomes are the bearers of hereditary characters,
but is this position altogether justifiable? It will be
well to consider some of the arguments for and against
the claim that chromosomes are the bearers of what are
called hereditary characters.

ARGUMENTS IN FAVOR OF THE CHROMOSOME HYPOTHESIS

This evidence will be considered under two heads,
namely, that from cytology and that from breeding.

CYTOLOGY.- (T) The program involved in the
ordinary mitotic division is very suggestive evidence.
We discover a very exact mechanism worked out, the
result of which is to divide the chromosomes evenly, so
that each daughter-nucleus receives an exactly similar
set of chromosomes, a set exactly similar to the original
one in the mother-nucleus. These chromosomes are

198

Bearers of Hereditary Characters 199

thus carefully divided and evidently perpetuated without
alteration. Furthermore, nothing else can be discovered
in the cell which is so carefully divided and perpetuated
without alteration. It is natural to conclude, therefore,
that chromosomes are the bearers of hereditary char-
acters, for if these bearers were anything that is not so
carefully divided inheritance would be chaos instead
of the remarkable cosmos it is.

(2) The behavior of chromosomes in the reduction
division is a remarkable piece of evidence. The fun-
damental features of this division are so familiar that
they need not be repeated. It is sufficient to say that
reduction division supplies a remarkable mechanism
which explains perfectly MENDEL'S theory of segrega-
tion. This remarkable coincidence of the observed
cytological facts with the preconceived theoretical
explanation of inheritance is probably the strongest
argument in favor of the chromosome hypothesis.

(3) An argument may also be obtained from the
cytology of fertilization. In the higher plants the
contributions of male and female to the structure of
the zygote are as follows. The egg contributes a
nucleus plus a portion of the cytoplasm in the embryo
sac, which may well contain many types of cell con-
stituents. The male, however, contributes only the
nucleus stripped of its cytoplasm. It is safe to infer
that the male contributes quite as much to the off-
spring as the female, at least in the matter of he r 'Ktary
characters, so that the conclusion may be that *
characters are carried entirely in the

cytologists claim that the nucleus is entir
chromatin, and since chromatin is *

200 Plant Genetics

which chromosomes are composed the male contributes
only chromosomes to the progeny, and judging from
breeding results the female contributes nothing more.
Thus by elimination the conclusion is reached that
chromosomes are the bearers of hereditary characters.

(4) In addition to the foregoing arguments there
are some minor cytological evidences that confirm the
chromosome theory. Tetraploid sporophytes, that is,
sporophytes with twice the normal number of chromo-
somes, have appeared as mutants, as in Oenothera, or in
apospory, as in MARCHAL'S mosses; such sporophytes
are distinctly larger and more vigorous than their
normal ancestors. This is indirect confirmation of the
chromosome hypothesis. Many hybrids are sterile, and
in some cases cytology shows that this sterility is the
result of, or at least associated with, a chaotic condition
of the chromosomes in the germinal tissue of the hybrid,
while everything else in the hybrid seems normal.

BREEDING. (i) MENDEL'S theory of segregation,
which must be regarded as well established, requires that
the bearers of hereditary characters must behave in a
certain particular way at maturation of the gametes.
This necessary behavior is found to occur quite regularly
among the chromosomes at the reduction division, and
nothing but chromosomes can be shown to behave in
this way.

(2) Mention has been made of MORGAN'S unique

work r linkage and crossing over. It is only necessary

at his unusual results are associated with

>gularities in the behavior of chromosomes,

trong confirmation of the chromosome

Bearers of Hereditary Characters 201

(3) If chromosomes bear hereditary characters, and
there are only two pairs of chromosomes in a plant, the
result would obviously be nothing above the dihybrid
ratio. Although the plant might show many more
than two Mendelian characters these characters should
be linked into two groups and dihybrid ratios result.
If trihybrid ratios were obtained, one should have to
conclude that something else than two pairs of chromo-
somes was involved. It is a mechanical necessity that
the ratios should not exceed the number of chromosomes.
For example, do we ever have pentahybrid ratios in a
plant with four pairs of chromosomes? This question
cannot be answered fully for plants, because all the plants
used in experiments have an ample number of chromo-
somes to account for the highest ratios discovered. In
MORGAN'S fruit fly, however, there are actually only
four pairs of chromosomes, and in all the great mass of
data collected on inheritance in the fruit fly tetra-
hybrid ratios occur regularly, but pentahybrid ratios
never.

(4) Finally, the fact that the chromosome hypothesis
is coming into more and more use in plant-breeding is a
general evidence in its favor. The plant breeder does
not usually question whether the theory of inheritance
he uses is true or false, but merely whether it is useful
or barren in leading to new results or connecting up
known facts. On this basis the chromosome hypothesis
is proving very useful.

ARGUMENTS AGAINST THE CHROMOSOME HYPOTHESIS

(i) Without going into "details the general fact may
be stated that the chromosome hypothesis does not

2O2 Plant Genetics

appeal to many of the most prominent physiologists and
physiological chemists.

(2) Physiological chemists claim that chromosomes
in all plants are chemically identical. If chromosomes
differ so much in the factors they carry, one should
expect them to show some chemical differences in the
various races of plants, but they show no such differences.
On the other hand, there are other plant constituents
that show a remarkable chemical specificity. Plant
proteins, for example, are chemically different in about
every plant species. Plant starches, also, have been
shown to be peculiar in their chemical constitution to
the race of plants in which they occur. It is natural
to ask why are not these proteins and starches, with
their remarkable specificity, more like what the carriers
of hereditary characters ought to be than the chromo-
somes which are chemically the same throughout the
plant kingdom.

(3) Such work as that of RIDDLE is directly against
the chromosome hypothesis. If it is found that char-
acters which were supposed to be determined by chromo-
somes can be controlled artificially by regulating the
physiological conditions, it makes one conclude that
chromosomes are not of controlling importance in
inheritance.

(4) Finally, there is the growing impression that
chromosomes are consequences instead of causes.
Instead of actually determining hereditary characters,
chromosomes may well be merely a rather useful super-
ficial index of the hereditary situation. It may be that
this is all the claim made for the chromosome hypothe-
sis by many cytologists. In this connection we may

Bearers of Hereditary Characters 203

quote from WILSON (3): "For the present, at least, all
the requirements of investigation are met if we think of
the chromosomes, or that which they carry, only as
differential factors in heredity, not as primary or exclusive
' determiners/ 77 It is a very dominant conception in
biology today that all plant and animal phenomena are
to be explained ultimately on a physico-chemical basis.
Cytology, therefore, should not be satisfied with observ-
ing and recording the peculiar contortions of chromo-
somes in their different phases. Such a study is justified
only in so far as it furnishes some useful working hy-
potheses. The ultimate thing to be striven for is to be
able to control plants and animals artificially," and this
can be clone only by understanding the physics and
chemistry that underlie the chromosomes as well as all
other life-phenomena.

With this presentation of the arguments for and
against the chromosome hypothesis of inheritance, a
conclusion seems to be difficult, but a suggestion may be
made that seems reasonable and may be useful. That
chromosomes are the visible bearers of hereditary char-
acters is probably true, at least in most cases upon which
there are data. Chromosomes, however, are not the
ultimate things; behind lies the physico-chemical
background, and it is this which is the ultimate aim
of biology to understand.

Such a statement may seem unsatisfactory. It has
been said that chromosomes determine sex in some cases
and that physiological conditions determine sex in other
cases; then it was inferred that the same may be true of
other plant characters as well. Can we not unify these
two alternative conclusions in some single theoretical

2O4 Plant Genetics

scheme ? Two attempts to devise such a unifying con-
ception will be presented briefly for what they are worth.
Both are highly theoretical.

The first of these theories is rather simple, sketching
a picture which is easy to visualize as a whole but is
inexact in its detail. Sex determination may be taken
as an example of the more general situation. The deter-
mination of sex may be likened to a pair of balances. If
one side goes clown a male is the result; if the other side
goes down a female is the result. What are the weights
that operate the balance ? They are of two kinds; one
is represented by the sex chromosomes, the other by
the physiological conditions. The sex chromosomes are
very heavy weights, so that if they are put on one side of
the balance, it would require a great many physiological
weights on the other side to counterbalance them.
Nevertheless, this can be done. With the heavy chro-
mosome weights all on one side they can still be over-
balanced by piling all of the physiological weights on
the other. In the organisms in which the sex chromo-
some theory seems rigidly established, it means that the
chromosome weights are relatively heavy and cannot
be counterbalanced by all of the physiological weights
on the other side; or it may merely mean that all of the
physiological weights have not yet been discovered.
In other forms the chromosome weights are not so heavy,
and can be overbalanced by the known physiological
weights. 1

It may be asked why the chromosome weights are
imagined to be so much heavier in some forms than in
others. The explanation may be that where the chromo-

ir This analogy of weights has been suggested by DONCASTER (i).

Bearers of Hereditary Characters 205

some weights are heaviest there is rigidity in inheritance,
but where the physiological weights are more prominent
there is plasticity in inheritance. Speaking in terms of
ontogeny, all embryo tissues are plastic and mature
tissues are rigid. Applying this thought to phylogeny
it may be assumed that young unit characters are plastic
and may be modified by physiological conditions; while
old unit characters are fixed, rigid, and determined by
chromosomes, against which the influence of physiological
conditions is feeble and avails little or nothing.

The second theory contains more physiological
complexities, sketching a picture which is harder to
visualize as a whole but is more exact in its details.
As previously mentioned, the physiologist would like
to believe that the protoplasts of each variety of plant
show a certain specificity in their chemical make-up.
The physiologist finds that the cytoplasm is the seat of
the most important physiological activities, and con-
sequently expects to find that the cytoplasm is also in
some way the bearer of hereditary characters. He does
not readily agree with the ' cytologist that hereditary
characters are limited entirely to the chromosomes. He
cannot, however, controvert the great bulk of data,
collected by cytologists and Mendelians, indicating that
chromosomes are the bearers of hereditary characters.
He must reconcile his views with these undeniable facts.
Perhaps he may be able to do this in the following
manner.

In the first place how must he visualize the so-
called "determiners" ? Physiologically they seem to
behave more like enzymes than anything else that we
can think of. If, then, these enzymes are transmitted

2o6 Plant Genetics

from generation to generation in the orderly manner
that the Mendelian data indicate, can we conceive of
their being located in the cytoplasm ? Enzymes dif-
fuse rather readily through the cytoplasm; but if each
determiner were represented by an enzyme, what a
chaotic mixture this cytoplasm would be, with the
numerous determiner-enzymes intermingling, interact-
ing, and probably losing their specificity. This is not
the orderly picture required to account for the orderly
Mendelian results actually obtained.

If, however, we conceive of these determiner-
enzymes inclosed in a linear arrangement along the
chromosomes we can paint a much more orderly picture.
The chromosome itself is not an active substance chemi-
cally; it is merely a framework, an " apartment house"
for determiner-enzymes, having, however, the power of
reproduction. Each determiner-enzyme is contained
(and "nourished") within its own compartment. When
the chromosome reproduces, each determiner-enzyme
reproduces within the chromosome and is passed on, at
all times, however, being well bound in by the walls of its
compartment and thus kept from reacting in the cyto-
plasmic medium which surrounds the chromosomes.
Sooner or later, as the automatic and inevitable result
of a natural series of physiological changes that has
taken place in the cell dynasty of the individual plant,
certain physiological conditions appear in the cytoplasm
which bring about the " release" of certain of these
determiner-enzymes, which then react in the cytoplasm
to produce the specific plant characters of which they
are the determiners. If all these determiner-enzymes
had been released at once, or if they had never been con-

Bearers of Hereditary Characters 207

fined, chaos rather than cosmos would have resulted.
Actually we know that cosmos exists, and our highly
theoretical scheme at least can account for that. After
all, what is this scheme but the "intracellular pangene-
sis" of DE VRIES (see chapter xiv) ?

Applying the scheme still further, we may make two
assumptions: (i) certain unusual physiological con-
ditions are capable of " releasing" determiner-enzymes
earlier than they would naturally be released. Thus
they are free to react in an unusual medium, or to inter-
act with other determiner-enzymes which they would
not have met in the natural course of events. In
either event the reaction is an unusual one, and the
final result may be the appearance of an unusual plant
character. (2) Unusual physiological conditions may
per se bring about unusual reactions and result in
unusual plant characters. The determiner-enzymes
would not be involved in this.

These two possibilities would account for some of the
unusual (hereditary) responses that have appeared as
the result of unusual physiological conditions, as, for
example, the mutants obtained by MACDOUGAL (2) when
he injected the ovaries of Onagra with various chemicals.
In short, such a theory may help to reconcile some of the
apparently contradictory data of the physiological school
and the cytological school.

To apply this theory to the matter of sex determina-
tion would involve certain difficulties, since sex is a
unique character and is inherited in a unique way. It
is quite possible, however, to overcome these theoretical
difficulties and to make our scheme reconcile the two con-
flicting views even on the matter of sex determination;

208 Plant Genetics

but it would not be appropriate to enter into the
theoretical complications any further. In this con-
nection, however, we may recall the work of CIESIELSKY
(see page 195), in which it is claimed that staleness
of pollen affects sex. Staleness probably means an
unusual amount of oxidation and would be a typical
instance of " unusual physiological conditions." One
might predict that staling of pollen might be found to
affect other plant characters as well as sex. This sug-
gests a new field of investigation. The male gamete
of angiosperms, however, is pretty well insulated within
the pollen coat. If work of this kind is to be done,
the algae and fungi would furnish much more favorable
material (see page 130).

These are two theoretical pictures in which both
chromosomes and physiological conditions may play
their roles. The latter is more complex, but at the same
time more exact and more suggestive of further investi-
gation. Both are highly theoretical and should not be
confused for a moment with any of the actual known
facts of inheritance. At best either one of them may
help to unify the situation that has been discussed, and
unless some such scheme is constructed it will be found
difficult to keep one's bearings among conflicting theories
based upon apparently conflicting facts.

1. DONCASTER, L., The determination of sex. Cambridge Univ.

Press. 1914.

2. MACDOUGAL, D. T.,see Heredity and eugenics (CASTLE, COUL-
TER, DAVENPORT, EAST, TOWER). Chicago. 1912 (p. 205).

3. WILSON, E. B., Heredity and microscopical research. Science
N.S. 37:814-826. 1913.

INDEX

INDEX

Acquired characters, inheritance
of, 6, 1 6

Adzuki beans, inheritance in, 45

Albomaculata type of Mirabilis, 98

Aleuronc layer, inheritance of color
in, 149 >,

Algae: inheritance in, 121, 130;
sex determination in, 185;
sexual structures of, 118

Allelomorphs, 31

Allen, C. E., 184, 187

Antagonistic characters, 86, 91

Antennaria, parthenogenesis in,
116

Anthocyanins in inheritance, 55

Apogamy: and true-breeding hy-
brids, 48; vegetative, 115

Apospory, 116, 189

Appetency, 16

Artificial selection, effect of, 80

Babcock, E. B., v

Bacteria, isolation of, 131

Barber, M. A., 131

Bateson, W., 40, 42, 55

Begonia: hybrids, 47; sex in, 196

Belling, John, 138

Biophores, 13

Blakeslee, A. E., 45, 185

Blends, 43/127

Bolley, H. L., 23, 26

Bryonia, sex determination in, 196

Bryophytes, inheritance in, 130

Burbank, Luther, 47, 158

Burgeff, H., 186

Camerarius, discovery of sex in
plants, 4

Cannabis, work on, 195

Capsella and alpine characters,

20

Castle, W. E., 104, 105
Cell architecture of Weismann,

13

Chlorophyll inheritance, 98, 101

Chromogens in inheritance, 55

Chromosomes: numbers in plants,
126; as bearers of hereditary
characters, 198; as differential
factors in heredity, 203; isola-
tion of, 131; theories of sex de-
termination, 178

Chrysanthemum, chromosomes in,
126

Cicsielsky, T., 195, 208

Clausen, R. E., v

Collins, G. N., 167

Complementary factors, 52, 57,
78, 127

Compositae, parthenogenesis in,
116

Continuity of germ plasm, 7

Continuous variations, 113

Corneous endosperm, inheritance
of, 154

Correlation, 96; physical and
gamctic, 73

Correns, C., 40, 43, 98, 192, 196

Coupled characters, 86, 91

Crepis, chromosomes in, 126

Crossing over, 90

Cross-pollination, significance of,
135

Crypthybrids, 48

Cumulative factors, 58, 61, 76, 78,
109, 128

Czapek, P, and M. E., 55

211

212

Darling, C. A., 184

Darwin, Charles, i, 4, 14, 68, 84,

136, 158

Determinants, 13
Determiner, 32, 52
De Vries, Hugo, 2, 14, 80, 84,

207

Dihybrid ratio, 37
Disease transmission, 26
Dominance, 30, 156; failure of, 45;

and hybrid vigor, 170
Doncaster, L., 204
Dose, double and single, 34
Double fertilization, 144
Drosophila, work on, 93
Duplex, 42

East, E. M., 27, 40, 46, 48, 50, 5 2 >
6g, 80, 112, 131, 142, 151, 153,
154, 162, 173
Elimination factor, 24
Embryo, sources of, 117
Emerson, R. A., 50, 86, 173
Endosperm, inheritance in, 144
Engler, Arnold, 22
Environment, effects of, 20
Enzymes in inheritance, 55, 205
Equisetum, sex in, 190

Factor hypothesis, 52
Fagopyrum, heterochromosomes

in, 184

Failure of dominance, 45
Ferns: apogamy in, *n6; sex

determination in, 191

Fixation of hybrids, 83

Floury endosperm, inheritance of,

154

Fluctuating variations, 70
Four-o'clock, inheritance in, 44
Fruit fly, inheritance in, 93
Funaria, sex determination in, 189
Functionless organs and inherit-
ance, 25

Fungi, sex determination in, 185
Fusion nucleus, 146

Gartner, J., 158
Galton, Francis, 16, 51

Galtonia, heterochromosomes in,
184

Gametes, purity of, 31
Gametic correlation, 73
Gametophytcs, inheritance in, 123,

138 -
Gemmules, 6
Genotype, 36
Geranium, work on, 100
Germ cells, origin of, 18

Germinal continuity, 6, 8; selec-
tion, 6, 9, 68

Ginkgo, heterochromosomes in,

184

Hemp, sex in, 195
Hertweg, Richard, 180
Heterochromosomes, 178
Heterosis, 164

Heterozygosis and hybrid vigor,
164, 169

Hetcrozygote, 35

Hieracium, parthenogenesis in,

116

Higgins, J. E., ^95
Homozygote, 35 ".

Houstonia, heterochromosomes in,

184 .

Hurst, C. C., 40
Hybrid vigor, 158

Idants, 12

Ids, 12

Indian hemp, sex in, n

Inheritance: of acqi

acters, 6, 16; in endosperm, 144,

in gametophytes, 138
Inhibitory factors, 57, 58, no, 128
Intermediate hybrids, 43