O?

1

Si

LT)
r=l

CD

m

HEREDITY

IN RELATION TO EVOLUTION
AND ANIMAL BREEDING

HEREDITY

BY

WILLIAM E. CASTLE

PROFESSOR OF ZOOLOGY, HARVARD UNIVERSITY

NEW YORK AND LONDON

D. APPLE TON AND COMPANY

1911

COPYRIGHT, 1911, BY
D. APPLETON AND COMPANY

Published September, 1911

Printed in the United States of America

PREFACE

THIS little book is based on a course of eight
lectures delivered in November and December,
1910, before the Lowell Institute, Boston, as well
as . on a course of five lectures delivered before
the Graduate School of Agriculture held under
the auspices of the Association of Agricultural
Colleges and Experiment Stations at Ames,
Iowa, in July, 1910. The hope is entertained
that it may be of service to students and that it
will also interest the general reader.

The writer wishes to express his gratitude to
the Carnegie Institution of Washington for per-
mission, in its preparation, to draw freely upon
published and unpublished material derived from
investigations aided by the Institution.

Acknowledgment is also due to the following
persons, or to their publishers, for permission to
use figures from their publications, as indicated
in the text : Prof. E. B. Wilson and The Mac-
millan Co., Prof. H. S. Jennings and The Ameri-
can Naturalist, Dr. W. B. Kirkham and The
American Book Co.

W. E. CASTLE
JUNE, 1911

CONTENTS

PAGE

INTRODUCTION. — GENETICS, A NEW SCIENCE ... 1
CHAPTER

I. — THE DUALITY OF INHERITANCE 6

II. — GERM- PLASM AND BODY, THEIR MUTUAL

INDEPENDENCE 27

III. — MENDEL'S LAW OF HEREDITY 33

IV. — THE DETERMINATION OF DOMINANCE;

HETEROZYGOUS CHARACTERS AND THEIR
FIXATION; ATAVISM OR REVISION. ... 52

V. — EVOLUTION OF NEW RACES BY Loss OR

GAIN OF CHARACTERS 72

VI. — EVOLUTION OF NEW RACES BY VARIATIONS

IN THE POTENCY OF CHARACTERS .... 87

VII. — CAN MENDELIAN UNIT-CHARACTERS BE

MODIFIED BY SELECTION? 106

VIII. — MENDELIAN INHERITANCE WITHOUT DOMI-
NANCE, "BLENDING" INHERITANCE . . 128

IX. — THE EFFECTS OF INBREEDING 143

X. — HEREDITY AND SEX 153

INDEX 183

LIST OF ILLUSTRATIONS

FIG. PAGE

1. — Egg and sperm of the sea-urchin, Toxopneustes 9

2. — Fertilization of the egg of Nereis 12

3. — Egg of a mouse previous to maturation Facing 14

4. — Maturation and fertilization of the egg of

a mouse . • Facing 14

5. — Diagrams showing the essential facts of

chromosome reduction in the development

of the sperm-cells 17

6. — An ordinary fern 21

7. — The prothallus of a fern 23

8. — Diagram showing the chromosome number in

the spermatogenesis of ordinary animals

and of the wasp 24

9. — Diagram showing the relation of the body to

the germ-cells in heredity 29

10. — A young, black guinea-pig Facing 30

11. — An albino female guinea-pig Facing 30

12. — An albino male guinea-pig Facing 30

13. — Pictures of three living guinea-pigs and of the

preserved skins of three others . . . Facing 32

14. — A black, female guinea-pig, and her young

Facing 34

15. — An albino male guinea-pig Facing 34

16. — Two of the grown-up young of a black and of

an albino guinea-pig Facing 34

ix

x LIST OF ILLUSTRATIONS

FIG. PAGE

17. — A group of four young, produced by the

animals shown in Fig. 16 ... .Facing 34

18. — Diagram to explain the result shown in Fig. 17 35

19. — A shortened condition of the skeleton, par-

ticularly of the fingers Facing 36

20. — Radiograph of a hand similar to those shown

in Fig. 19 Facing 38

21. — Diagram showing the descent, through five
generations, of the condition shown in

Figs. 19 and 20 40

22. — A smooth, dark guinea-pig ..... Facing 40

23. — A rough, white guinea-pig Facing 40

24. — A dark, rough guinea-pig Facing 40

25. — A smooth, white guinea-pig Facing 42

26. — A short-haired, pigmented guinea-pig Facing 42

27. — A long-haired albino guinea-pig . . . Facing 42

28. — Offspring produced by animals of the sorts

shown in Figs. 26 and 27 ... .Facing 42

29. — A long-haired, pigmented guinea-pig, "Dutch-

marked " with white Facing 42

30. — Diagram to explain the results of a cross

between the sorts of guinea-pigs shown in
Figs. 26 and 27 43

31. — Diagram showing the kinds and relative

frequencies of the young to be expected in
F2 from the crossing of animals shown in
Figs. 26 and 27 46

32. — Along-haired, rough albino guinea- pig Facing 46

33. — Five new combinations of unit-characters

obtained in generation F2, by crossing the
animal shown in Fig. 32 with animals like
that shown in Fig. 22 . , , . . , Facing 48

LIST OF ILLUSTRATIONS xt

FIG. PAGE

34. — Diagram to show the gametic combinations

and segregations involved in a cross be-
tween guinea-pigs differing in three unit-
characters 49

35. — Diagram to show the gametic combination

and recombinations which occur in the
production and fixation of an atavistic coat-
character in guinea-pigs 66

36. — An imperfectly rough guinea-pig . . .Facing 101

37. — A silvered guinea-pig Facing 101

38. — A. Front feet of an ordinary guinea-pig.

B. Its hind feet D. Hind feet of a race
four- toed on all the feet. C. Ordinary
condition of the hind feet of young obtained
by crossing B with D Facing 101

39. — Diagram showing variation in the color-

pattern of hooded rats Facing 101

40. — Diagram showing the variations in size of

eight different races of paramecium ... 112

41. — Chart showing effects of selection in eight

successive generations upon the color-
pattern of hooded rats 122

42. — Skulls of three rabbits Facing 128

43. — A long-haired, albino rabbit, having erect

ears Facing 132

44. — A short-haired, sooty yellow rabbit, having

lop ears Facing 132

45. — A short-haired, black rabbit, son of the

rabbits shown in Figs. 43 and 44 . .Facing 132

46. — An F2 descendant of the rabbits shown in

Figs. 44 and 45 .,,.,,., .Facing 132

xii LIST OF ILLUSTRATIONS

FIG. PAGE

47. — Diagrams to show the number and size of the

classes of individuals to be expected from
a cross involving Mendelian segregation
without dominance 135

48. — Photographs to show variation in ear length

of two varieties of maize, of their Fx off-
spring, and of their F2 offspring . .Facing 138

49. — Diagram of sex-determination in partheno-

genesis 162

50. — Diagram of sex determination when the female

is homozygous, the male heterozygous . . 167

51. — Diagram of sex-determination when the female

is heterozygous, the male homozygous . . 170

52. — Diagram of sex-limited inheritance when the

female is a heterozygote 173

53. — Diagram of sex-limited inheritance when the

female is a homozygote, as in the red-eyed
Drosophila 175

HEREDITY

INTRODUCTION

GENETICS, A NEW SCIENCE

THE theory of organic evolution has prob-
ably influenced more fields of human
activity and influenced them more pro-
foundly than has any other philosophic deduc-
tion of ancient or modern times. By this theory
philosophy, religion, and science have been rev-
olutionized, while in the practical arts of educa-
tion and agriculture, twin foundation stones of
the state, man has been forced to adopt new
methods of procedure or to justify the old ones
in the light of a new principle.

The evolutionary idea has forced man to con-
sider the probable future of his own race on
earth and to take measures to control that fu-
ture, a matter he had previously left largely to

fate. With a realization of the fact that or-

1

HEEEDITY

ganisms change from age to age and that he
himself is one of these changing organisms man
has attained not only a new ground for humility
of spirit but also a new ground for optimism and
for belief in his own supreme importance, since
the forces which control his destiny have been
placed largely in his own hands.

The existence of civilized man rests ultimately
on his ability to produce from the earth in suf-
ficient abundance cultivated plants and domes-
ticated animals. City populations are apt to
forget this fundamental fact and to regard with
indifference bordering at times on scorn agri-
cultural districts and their workers. But let the
steady stream of supplies coming from the land
to any large city be interrupted for only a few
days by war, floods, a railroad strike, or any
similar occurrence, and this sentiment vanishes
instantly. Man to live must have food, and
food comes chiefly from the land.

A knowledge of how to produce useful animals
and plants is therefore of prime importance.
Civilization had its beginning in the attainment
of such knowledge and is limited by it at the
present day. If, therefore, this knowledge can

2

be increased, civilization may be advanced in a
very direct and practical way. Before Darwin
the practices of animal and plant breeders were
largely empirical, based on unreasoned past ex-
perience, just as was in antiquity the practice
of metallurgy. Good plows and good swords
were made long before a scientific knowledge
of the metals was attained, but without that sci-
entific knowledge the wonderful industrial de-
velopment of this present age of steel would
have been quite impossible. In a similar way,
if not in like measure, we may reasonably hope
for an advance in the productiveness of animal
and plant breeding when the scientific principles
which underlie these basic arts are better under-
stood. Two practical problems present them-
selves to the breeder: (1) how to make best use
of existing breeds, and (2) how to create new
and improved breeds better adapted to the con-
ditions of present-day agriculture. We shall
concern ourselves with the second of these only.
The production of new and improved breeds
of animals and plants is historically a matter
about which we know scarcely more than about

the production of new species in nature. Selec-

3

HEREDITY

tion has been undoubtedly the efficient cause of
change in both cases, but how and why applied
and to what sort of material is as uncertain
in one case as in the other. The few great
men who have succeeded in producing by their
individual efforts a new and more useful type
of animal or plant have worked largely by
empirical methods. They have produced a
desired result but by methods which neither
they nor any one else fully understood or
could adequately explain. So there exists as
yet no true science of breeding but only a
highly developed art which was practiced as
successfully by the ancient Egyptians, the
Saracens, and the Romans as by us. The
present, however, is an age of science; we
are not satisfied with rule-of-thumb methods,
we want to know the why as well as the how
of our practical operations. Only such knowl-
edge of the reasons for methods empirically
successful can enable us to drop out of our
practice all superfluous steps and roundabout
methods and to proceed straight to the mark
in the most direct way. The industrial his-
tory of the last century is full of instances in

4

GENETICS, A NEW SCIENCE

•which a knowledge of causes in relation to
processes, i. e. a scientific knowledge, has
shortened and improved practice in quite un-
expected ways. So we may not doubt the ulti-
mate value in practice of a science of breeding,
if such a science can be created.

A beginning has been made during the last
ten years, starting with the rediscovery of
Mendel's law of heredity in 1900. This book
will be concerned largely with the operations
of that law.

CHAPTER I

THE DUALITY OF INHERITANCE

A the outset we may with profit inquire
what is meant by heredity. When a
child resembles a parent or grand-
parent in some striking particular, we say it
inherits such-and-such a characteristic from
the parent or grandparent in question. By
heredity, then, we mean organic resemblance
based on descent.

Resemblances due to heredity may exist even
between individuals not related as ancestor and
descendant, as for example between uncle and
nephew. Here the resemblance rests on the
fact that uncle and nephew are both descended
from a common ancestor, and they resemble
each other simply because they have both in-
herited the same characteristic from that an-
cestor. This form of inheritance is sometimes
spoken of as collateral in distinction from direct

6

THE DUALITY OF INHEEITANCE

inheritance. In all cases alike community of
descent is the basis of resemblances which can
be ascribed to heredity, whether direct or col-
lateral. Mother and child, no less than uncle
and nephew, resemble each other because they
have received a common inheritance from a
common ancestor.

Three biological facts of fundamental im-
portance to a right understanding of heredity
were known imperfectly or not at all in the
time of Darwin and Mendel. These are (1)
the fertilization of the egg, (2) the maturation
of the egg, which must precede its fertilization,
and (3) the non-inheritance of " acquired "
characters. These we may consider in order.

Every new organism is derived from a pre-
existing organism, so far as our present ex-
perience goes. It may not have been so al-
ways. Indeed, on the evolution theory, we
must suppose that living matter originally
arose from lifeless, inorganic matter. But if
it did, this may have occurred, and probably
did occur, under physical conditions quite
different from those now existing. At the
present time the most exhaustive researches

7

HEREDITY

fail to reveal the occurrence of spontaneous
generation, that is, the origin of living beings
other than from pre-existing living beings.

In asexual methods of reproduction a new
individual arises out of a detached portion of
the parent individual. Such methods of origin
are varied and interesting, but do not concern
us at present. In all the higher animals and
plants a new individual arises, by what we call
a sexual process, from the union of two minute
bodies called the reproductive cells. They are
an egg-cell furnished by the mother and a
sperm-cell furnished by the father.

There is a great difference in size between
egg and sperm. The egg is many thousand
times greater in bulk, as seen in Fig. 1, for
example, yet the influence of each in heredity
appears to be equal to that of the other. This
fact shows unmistakably that the bulk of the
reproductive cell is not significant in heredity.
A large part of the relatively huge egg can
have no part in heredity. It serves merely as
food for the new organism, furnishing it with
building material until such a time as it can
begin to secure food for itself. The essential

8

THE DUALITY OF INHEKITANCE

material, so far as heredity is concerned, is
evidently found in egg and sperm alike. It
is plainly small in amount and possibly con-
sists merely in ferment-like bodies which ini-

FIG. 1. — Egg and sperm (s) of the sea-urchin, Toxopneustes,
both shown at the same enlargement. (After Wilson.)

tiate certain metabolic processes in a suitable
medium represented by the bulk of the egg.
The amount of a ferment used in starting a
chemical change bears no relation, as is well
known, to the amount of the chemical change
which it can bring about in a suitable medium.

9

HEKEDITY

The equal share of egg and sperm in deter-
mining the character of offspring is well shown
in the following experiment. An albino guinea-
pig is one which lacks in large measure the
ability to form black pigment. Apparently it
does not possess some ingredient or agency
necessary for the production of pigment. Now,
if an albino male guinea-pig, such as is shown
in Fig. 15, be mated with a black female guinea-
pig of pure race, such as is shown in Fig. 14,
young are produced all of which are black, like
the mother, none being albinos, like the father.
Fig. 16 shows black offspring produced in this
way. Exactly the same result is obtained from
the reverse cross, that is, from mating an al-
bino mother with a black sire. It makes no
difference, then, whether the black parent be
mother or father, its blackness regularly domi-
nates over the whiteness of the albino parent,
so that only black offspring result. This fact,
which has been repeatedly confirmed, shows that
the black character is transmitted as readily
through the agency of the minute sperm-cell
as through the enormously greater egg-cell.

Let us now consider what happens when egg
10

THE DUALITY OF INHERITANCE

and sperm unite, in what we call the fertiliza-
tion of the egg. The egg is a rounded body
incapable of motion, but the sperm is a minute
thread-like body which moves like a tadpole
by vibrations of its tail. In the case of most
animals which live in the water, egg and sperm-
cells are discharged into the water and there
unite and develop into a new individual, but
in the case of most land animals this union takes
place within the body of the mother. We may
consider an illustration of either sort.

The fertilization of the egg of a marine
worm, Nereis, is shown in Fig. 2. The thread-
like sperm penetrates into the egg. Its en-
larged head-end forms there a small nmtfear
body, which increases in size until it equals
that of the egg-nucleus, with which it then fuses.
The egg next begins to divide up to form the
different parts of a new worm-embryo. To
each of these parts the nuclear material of egg
and sperm is distributed equally. Since this
development takes place wholly outside the
body of either parent it is necessary that the
egg contain enough food to last until the young
worm can feed itself. This food material is

11

HEREDITY

FIG. 2. — Fertilization of the egg of Nereis.
A. The sperm has entered the egg and is forming a minute
nucleus at o^. The egg-nucleus is breaking up preparatory
to the first maturation division. B. The egg-nucleus is
undergoing the first maturation division. Notice the con-
spicuous rod-like chromosomes separating into two groups.
The sperm-nucleus ( £ ) is now larger and lies deeper in the
egg. C. A small polar-cell has been formed above by the
first maturation division of the egg. A second division is
in progress at the same point. The sperm-nucleus is now
deep in the egg and is preceded by a double radiation (am-
phiaster). D. Two polar-cells are fully formed. The ma-
tured egg-nucleus is now fusing with the sperm-nucleus.
An amphiaster indicates that division of the egg will soon
take place. (After Wilson.)

IS

THE DUALITY OF INHEKITANCE

represented in part by the conspicuous oil-
drops seen in the egg (the heavy circles in
Fig. 2).

The egg of a mouse needs no such store of
nourishment, since in common with the young
of other mammals the mouse-embryo nourishes
itself by osmosis from the body fluids of the
mother. The mouse-egg is accordingly smaller.
Stages in its fertilization are shown in Fig. 4.
In A the sperm has already entered the egg.
Eemnants of its thread-like tail may still be
seen there. Nearby is seen a nuclear body
derived from the sperm-head. Opposite is
seen the nuclear body furnished by the egg
itself. The two nuclear bodies fuse and their
united substance is then distributed to all
parts of the embryo-mouse, just as happens in
the development of the worm, Nereis.

There are reasons for thinking that the
nuclear material is especially important in re-
lation to heredity and that the equal share of
the two parents in contributing it to the em-
bryo is not without significance, for inheritance,
as we have seen, is from both parents in equal
measure. In cases where the inheritance from

13

HEREDITY

each parent is different it can be shown
that the offspring possess two inherited possi-
bilities, though they may show but one. Thus
in the case of a black guinea-pig, one of whose
parents was white, the other black, it can be
shown that the animal transmits both qualities
(black and white) which it received from its
respective parents, and transmits them in equal
measure. For, if the cross-bred black animal
be mated with a white one, half the offspring
are black and half of them white. The cross-
bred black animal inherited black from one
parent, white from the other. It showed only
the former, but on forming its reproductive
cells it transmitted black to half of these, white
to the other half. Hence the cross-bred black
individual was a duality, containing two possi-
bilities, black and white, but its reproductive
cells were again single, containing either black
or white, but not both.

Now it has been shown in recent years that
the nuclear material in the reproductive cells
behaves exactly as do black and white in the
cross just described. This nuclear material
becomes doubled in amount at fertilization,

U

FIG. 3. — Egg of a
mouse previous to
maturation. (After
Kirkham.)

\

FIG. 4. — Maturation and fertilization of the egg of a mouse.
A. The first maturation division in progress. B. The first
polar-cell fully formed ; the second maturation division
in progress. C. The second maturation division com-
pleted ; the second polar-cell is the smaller one ; near it,
in the egg, is the egg-nucleus, and at the left is the sperm-
nucleus. D. A view similar to the last, but showing only
one polar-cell, the second; note its twelve distinct
chromosomes ; near the sperm-nucleus in the egg, at
the left, is seen the thread-like remains of the sperm-tail.
(After Kirkham.)

THE DUALITY OF INHERITANCE

equal contributions being made by egg and
sperm. This double condition persists through-
out the life of the new individual in all its
parts and tissues. But if the individual forms
eggs or sperm, these, before they can function
in the production of a new individual, must
undergo reduction to the single condition.

This reduction process is called maturation;
it is well illustrated in the case of the mouse-
egg, whose fertilization has already been de-
scribed. The large nucleus of the egg-cell, as
it leaves the ovary, is either broken up or about
to break up preparatory to a cell-division. The
most conspicuous of the nuclear constituents
are some dense, heavily staining bodies called
chromosomes, about twenty-four in number.
In Fig. 3 each of these is split in two, prepara-
tory to the first maturation division. The egg
now divides twice, both times very unequally
(Fig. 4), forming thus two smaller cells called
polar cells, or polar bodies. They take no part
in the formation of the embryo. The chromo-
somes left in the egg after these two divisions
are only about half as numerous as before, or
about twelve in number. These form the chro-

15

HEREDITY

matin contribution of the egg to the production
of a new individual. It is possible that other
cell constituents undergo a similar reduction
by half during maturation, but of this we have
no present knowledge.

The known fact of chromosome reduction,
of course, favors the current interpretation
that the chromosomes are bearers of heredity,
though it by no means proves the correctness
of that interpretation. In the egg of Nereis,
as well as in that of the mouse, two matura-
tion divisions precede the fertilization of the
egg. See Fig. 2. In B the first maturation
division is in progress; in C the second is in
progress; and in D both polar cells are fully
formed, while egg and sperm nuclei are unit-
ing. Similar processes occur in eggs gener-
ally, prior to their fertilization.

Like changes occur also in the development
of the sperm-cells. In Fig. 5 the original or
unreduced condition of the chromosomes in a
cell of the male sexual gland is shown (at A) as
one of four chromosomes to a cell. After a series
of changes involving as in the maturation of the
egg two cell-divisions, we find (at H) that the

16

THE DUALITY OF INHERITANCE

products of the original cell contain in each case
two chromosomes, half the original number.

FIG. 5. — Diagrams showing the essential facts of chromosome
reduction in the development of the sperm-cells. (After
Wilson.)

These chromosomes make up the bulk of the
head of the sperm which forms from each of

17

these cells, its tail being derived from other
portions of the cell.

It follows that not only eggs but also sperms,
prior to their union in fertilization have passed
into a reduced or single state as regards their
chromatin constituents, whereas the fertilized
egg, and the organism which develops from it,
is in a double condition. It will be convenient
to refer to the single condition as the N condi-
tion, the double as the 2 N condition.

From a wholly different source we have
evidence strongly confirmatory of the conclu-
sion that the fertilized egg contains a double
dose of the essential nuclear material. By arti-
ficial means it has been found possible to cause
the development of an unfertilized egg. The
means employed may be of several different
sorts, such as stimulation with acids, alkalies,
or solutions of altered density. In such ways
the development has been brought about of the
eggs of sea-urchins, star-fishes, worms, and mol-
lusks, which normally require fertilization to
make them develop.

The sea-urchin egg has been made to develop
more successfully than any other. This has

18

occurred even after the egg had undergone
maturation, being reduced to the N condition.
From the development of such reduced but
unfertilized eggs fully normal sea-urchins have
been obtained which even contain developed
sexual glands. On the other hand it has been
found possible to break the egg into fragments
by shaking it, or cutting it into bits with fine
knives or scissors. It has also been found
possible to bring about the development of
an egg fragment so obtained, — a fragment
which contained no egg nucleus. This result
has been attained by allowing a sperm to enter
it and form there a nuclear body. No adult
organism has yet been reared from such a
fertilized egg-fragment, but so far as the de-
velopment has been followed it progressed
normally.

There can accordingly be no doubt that the
nuclear material of a sperm-cell has all the
capabilities of that of an egg-cell and can in-
deed replace it in development. Accordingly,
when, as in normal fertilization, both an egg
nucleus and a sperm nucleus are present in
the cell, a double dose of the necessary nuclear

19

HEREDITY

material is supplied. The second or extra dose
is, however, not superfluous. It probably adds
to the vigor of the organism produced, and in
some cases at least, materially affects its form.
For many animals and plants exist in two
different conditions, in one of which the nu-
clear components are simple, N, while in the
other they are double, 2 N. Thus in bees,
rotifers, and small Crustacea the egg may
under certain conditions develop without being
fertilized. If the egg develops before matura-
tion is complete, that is in the 2 N condition,
the animal produced is a female, like the
mother which produced the egg. But if the
egg undergoes reduction to the N condition
before beginning its development, then it pro-
duces a male individual, an organism, so far
as reproduction is concerned, of lower meta-
bolic activity.

In many plants, too, individuals of N and
of 2 N constitution occur, which differ markedly
in appearance. Thus the ordinary fern-plant
is a 2 N individual, but it never produces 2 N
offspring. Fig. 6 shows an ordinary fern-
plant, which produces spores on the under

20

THE DUALITY OF INHERITANCE

FIG. 6. — An ordinary fern, which reproduces by asexual spores.
The fern is shown reduced in size at 382; a portion of a
frond seen from below and slightly enlarged, at 383; a
cross-section of the same more highly magnified, at 384.
Notice in 384 the sporangia, and in 385 one of these dis-
charging spores. (After Wossidlo, from Coulter Barnes
and Cowle's Textbook of Botany.)

3

21

HEREDITY

surface of its fronds. Each of those spores
is a reproductive cell which, like the mature
eggs and sperm of animals, is in a reduced
nuclear condition (N). These spores germi-
nate, however, without uniting in pairs and
form a plant different from the parent, just
as the mature egg of a bee, if unfertilized,
develops into an individual different from the
parent, in that case a male. The plant which
develops from the spore of a fern is small and
inconspicuous and is known as a prothallus.
See Fig. 7. It produces sexual cells (eggs and
sperm ) which, uniting in pairs, form fern-plants,
2 N individuals. Thus there is a constant alter-
nation of generations, fern-plants (2 N), which
produce prothalli (N), and then these produce
again fern-plants (2 N).

The fact is worthy of note that in an animal
or plant which is in the single or N condition,
there occurs no chromatin reduction at the
formation of reproductive cells. Its cells are
already in the single condition, and they
probably cannot be further reduced without
destroying the organism. The 2 N fern-plant
forms reproductive cells, its spores, which are

22

THE DUALITY OF INHEKITANCE

in the reduced condition, N, and these germi-
nate into the prothallus, which accordingly is

FIG. 7. — The prothallus of a fern, which reproduces by sexual
cells, eggs and sperm. The eggs are borne in the sac-like
"archegonia," just below the notch in the figure. They,
like the sperm -forming "antheridia," lie on the under sur-
face of the flattened prothallus which is here viewed from
below. Notice the root-hairs or rhizoids by which the
plant feeds. Highly magnified. (After Coulter, Barnes,
and Cowles.)

N throughout. But when the prothallus forms
reproductive cells, no reduction occurs. Its
egg-cells and its sperm-cells in common with

23

HEREDITY

all other cells of the prothallus are already
in the reduced condition without any matura-
tion divisions. The result of their union in
pairs, at fertilization, is the formation of 2 N
combinations that germinate into fern-plants.
Similarly in the case of a male animal which

FIG. 8. — Diagram showing the chromosome number in the
spermatogenesis of ordinary animals (upper line) and of the
wasp (lower line).

has developed from a reduced but unfertilized
egg, no reduction occurs at the formation of
its sperm-cells. In an ordinary male animal,
one which is in the double or 2 N state, the
development of the sperms is attended by re-
duction to the N condition. In this process
there occur two cell-divisions producing from
each initial cell four sperms. See Fig. 5, and

24

THE DUALITY OF INHEEITANCE

Fig. 8, upper line. But in the male wasp,
whose cells are in the N condition at the be-
ginning, one of these divisions is so far sup-
pressed that the resulting cell products are of
very unequal size, and the smaller one contains
no nuclear material. The other then gives rise
to two sperm-cells, each possessing the origi-
nal N nuclear condition, while the small non-
nucleated cell degenerates. See Fig. 8, lower
line.

In conclusion, I wish to introduce two tech-
nical terms, which it will be convenient for us
to use in subsequent discussions. These are
gamete and zygote. A reproductive cell (either
egg or sperm) which is in the reduced condi-
tion (N) ready for union in fertilization is
called a gamete. The result of fertilization is
a zygote, a joining together of two cells each
in the N condition. The result is a new or-
ganism, at first a single cell, in the 2 N
condition.

HEREDITY

BIBLIOGRAPHY

CASTLE, W. E.

1903. "The Heredity of Sex." Bull. Mus. Comp. Zool-
ogy, 40, pp. 189-218.
DELAGE, Y.

1898. "Embryons sans noyau maternel." Compte rendu,
Academic des sciences, Paris, 127, pp. 528-531.

1909. "Le sexe chez les Oursins issus de parthe"-
nogenese experimentale." Compte rendus, Academic des
sciences, Paris, 148, pp. 453-455.

KlRKHAM, W. B.

1907. Maturation of the Egg of the White Mouse."
Trans. Conn. Acad. of Arts and Sciences, 13, pp. 65-87.

LOEB, J.

1899. "On the Nature of the Process of Fertilization and
the Artificial Production of Normal Larvae (Plutei) from
the Unfertilized Eggs of the Sea-urchin." Amer. Journ.
of Physiol, 3, pp. 135-138.

LOTSY, J. P.

1905. "Die X-Generation und die 2 X-Generation."

Biologisches Centralblatt, 25, pp. 97-117.
MEVES, F., und DUESBERG, J.

1908. "Die Spermatozytenteilungen bei der Hornisse
(Vespa crabo L.)." Arch. f. mik. Anat. u. Entwick., 71,
pp. 571-587.

WILSON, E. B.

1896. "The Cell in Development and Inheritance," 370
pp., illustrated. The Macmillan Co., New York.

CHAPTER II

GEBM-PLASM AND BODY, THEIR MUTUAL
INDEPENDENCE

IN the last chapter we discussed two bio-
logical principles which, if clearly grasped,
greatly simplify an understanding of the
process of heredity. These are as follows:

(1) A sexually produced individual arises
from the union of two reproductive cells (or
gametes), each of which contains, so far as
heredity is concerned, a full material equip-
ment for the production of a new individual.
Accordingly, the newly produced individual is
two-fold or duplex as concerns the material
basis of heredity.

(2) If the new individual becomes adult and
forms gametes, the production of these will be
attended by a reduction to the simplex or
single condition as regards the material basis
of heredity.

27

HEREDITY

To these two principles we may now add
a third, viz.: — (3) The individual consists
of two distinct parts : first, its body destined
to die and disintegrate after a certain length
of time; and, secondly, the germ-cells con-
tained within that body, capable of indefinite
existence in a suitable medium.

The fertilized egg or zygote begins its in-
dependent existence by dividing into a number
of cells. These become specialized to form
the various parts and tissues of the body,
muscle, bone, nerve, etc., and by becoming thus
specialized they lose the power to produce any-
thing but their own particular kind of special-
ized tissue; they cannot reproduce the whole.
This function is retained only by certain un-
differentiated cells found in the reproductive
glands and known as germ-cells. They are
direct lineal descendants of the fertilized egg
itself. If they are destroyed the individual
loses the power of reproduction altogether.

External influences which act upon the body
may of course modify it profoundly, but such
modifications are not transmitted through the
gametes, because the gametes are not derived

28

GERM-PLASH AXD BODY

from body-cells, but from germ-cells. This
relationship first pointed out by Weismann
may be expressed in a diagram, as in Fig. 9.
Only such environmental influences as directly
alter the character of the germ-cells will in
any way influence the character of subsequent
generations of individuals derived from those

Line of succession.

© Line of inheritance.
C

FIG. 9. — Diagram showing the relation of the body (<S) to the
germ-cells (G) in heredity. (After Wilson.)

germ-cells. Body (or somatic) influences are
not inherited. This knowledge we owe largely
to Weismann, who showed experimentally that
mutilations are not inherited. The tails of
mice were cut off for twenty generations in
succession, but without effect upon the char-
acter of the race. Weismann also pointed out
the total lack of evidence for the then current
belief that characters acquired by the body
are inherited. The correctness of his view that
body and germ-cells are physiologically distinct

29

HEREDITY

is indicated by the results obtained when germ-
cells are transplanted from one individual to
another.

Heape showed some twenty years ago that
if the fertilized egg of a rabbit of one variety
(for example an angora, i. e. a long-haired,
white animal) be removed from the oviduct of
its mother previous to its attachment to the
uterine wall, and be then transferred to the
oviduct of a rabbit of a different variety (for
example a Belgian hare, which is short-haired
and gray), the egg will develop normally in
the strange body and will produce an individual
with all the characteristics of the real (an-
gora) mother unmodified by those of the foster
mother (the Belgian hare). Young thus ob-
tained by Heape were both long-haired and
albinos, like the angora mother. To this ex-
periment the objection might be offered that
the transplanted egg was already full-grown
and fertilized when the transfer was made, and
that therefore no modification need be expected,
but if the egg were transferred at an earlier
stage the result might have been different. In
answer to such a possible objection the follow-

30

FIG. 10. — A young, black guinea-pig, about three
weeks old. Ovaries taken from an animal like
this were transplanted into the albino shown
below.

FIG. 11. — An albino female guinea-pig. Its
ovaries were removed, and in their place were
introduced ovaries from a young, black guinea-
pig, like that one shown in Fig. 10.

FIG. 12. — An albino male guinea-pig, with which
was mated the albino shown in Fig. 11.

GERM-PLASM AND BODY

ing experiment performed by Dr. John C. Phil-
lips and myself may be cited.

A female albino guinea-pig (Fig. 11) just at-
taining sexual maturity was by an operation
deprived of its ovaries, and instead of the re-
moved ovaries there were introduced into her
body the ovaries of a young black female
guinea-pig (Fig. 10), not yet sexually mature,
aged about three weeks. The grafted animal
was now mated with a male albino guinea-pig
(Fig. 12). From numerous experiments with
albino guinea-pigs it may be stated emphati-
cally that normal albinos mated together, with-
out exception, produce only albino young, and
the presumption is strong, therefore, that had
this female not been operated upon she would
have done the same. She produced, however,
by the albino male three litters of young,
which together consisted of six individuals, all
black. (See Fig. 13.) The first litter of young
was produced about six months after the oper-
ation, the last one about a year. The trans-
planted ovarian tissue must have remained in
its new environment therefore from four to
ten months before the eggs attained full growth

31

HEREDITY

and were discharged, ample time, it would seem,
for the influence of a foreign body upon the
inheritance to show itself were such influence
possible.

In the light of the three principles now
stated, viz. (1) the duplex condition of the
zygote, (2) the simplex condition of the
gametes, and (3) the distinctness of body and
germ-cells, we may proceed to discuss the
greatest single discovery ever made in the
field of heredity, — Mendel's law.

BIBLIOGRAPHY

CASTLE, W. E., and PHILLIPS, JOHN C.

1911. "On Germinal Transplantation in Vertebrates."
Carnegie Institution of Washington, Publication No.
144, 26 pp., 2 pi.
HEAPE, W.

1890. "Preliminary Note on the Transplantation and
Growth of Mammalian Ova within a Uterine Foster-
mother." Proc. Roy. Soc., 48, pp. 457-458.
1897. "Further Note," etc. Id. 62, pp. 178-183.
WEISMANN, A.

1893. "The Germ-Plasm." Translation by Parker and
Romfeldt. Chas. Scribner's Sons, New York.

FIG. 13. — Pictures of three living guinea-pigs (A, B, C),
and of the preserved skins of three others (D, E, F);
all of which were produced by the pair of albinos shown
in Figs. 11 and 12.

CHAPTER III

MENDEL'S LAW OF HEREDITY

G^EGOR JOHANN MENDEL was a
teacher of the physical and natural
sciences in a monastic school at Briinn,
Austria, in the second half of the last cen-
tury. He was, therefore, a contemporary of
Darwin, but unknown to him as to nearly
all the great naturalists of the period. Al-
though not famous in his lifetime, it is clear
to us that he possessed an analytical mind
of the first order, which enabled him to plan
and carry through successfully the most origi-
nal and instructive series of studies in hered-
ity ever executed. The material which he used
was simple. It consisted of garden-peas, which
he raised in the garden of the monastery.
The conclusions which he reached were like-
wise simple. He summed them up, the results
of eight years of arduous work, in a brief
paper published in the proceedings of the local

33

HEREDITY

scientific society. There they remained un-
heeded for thirty-four years, until their author
had long been dead. Meantime biological sci-
ence had made steady progress. It reached
the position Mendel had attained in advance
of his time, and Mendel's law was rediscov-
ered simultaneously in 1900 by De Vries in
Holland, by Correns in Germany, and by
Tschermark in Austria. It gratifies our sense
of poetic justice that to-day the rediscovered
law bears the name, not of any one or of all of
its brilliant rediscoverers, but of the all-but-
forgotten Mendel.

The essential features of this law can best
be explained in connection with some illustra-
tions, which I choose for convenience from my
own experiments. If a black guinea-pig of
pure race (Fig. 14) be mated with a white one
(Fig. 15), the offspring will, as explained on
page 10, all be black; none will be white.
To use Mendel 's terminology, the black char-
acter dominates in the cross, while white
recedes from view. The black character is,
therefore, called the dominant character;
white, the recessive character.

34

MENDEL'S LAW OF HEREDITY

But, if now two of the cross-bred black in-
dividuals (Fig. 16) be mated with each other,
the recessive white character reappears on the
average in one in four of the offspring (Fig.

W ) ( W W ) Zygotes

FIG. 18. — Diagram to explain the result shown in Fig. 17.

17). Its reappearance in that particular pro-
portion of the offspring may be explained as
follows (see Fig. 18) : The gametes which
united in the original cross were, one black,
the other white in character. Both characters

35

HEREDITY

were then asociated together in the offspring;
but black from its nature dominated, because
white in this case is due merely to the lack of
some constituent supplied by the black gamete.
But when the cross-bred black individuals on
becoming adult form gametes, the black and
the white characters separate from each other
and pass into different cells, since, as we have
seen, gametes are simplex. Accordingly, the
eggs formed by a female cross-bred black are
half of them black, half of them white in char-
acter, and the same is true of the sperms
formed by a male cross-bred black. The com-
binations of egg and sperm which would natu-
rally be produced in fertilization are accord-
ingly 1BB:2BW:1WW, or three combina-
tions containing black to one containing only
white, which is the ratio of black to white off-
spring observed in the experiment.

Now the white individual may be expected
to transmit only the white character, never the
black, because it does not contain that char-
acter. Experiment shows this to be true.
White guinea-pigs mated with each other pro-
duce only white offspring. But the black in-

36

o

ao;
efj

O 03

.3

•3 (H

MENDEL'S LAW OF HEREDITY

dividuals of this generation are of two sorts,
- B B and B W in character. The B B indi-
vidual is pure, so far as its breeding capacity
is concerned. It can form only black (B)
gametes. But the B W individuals may be
expected to breed exactly like the cross-bred
blacks of the previous generation, forming
gametes, half of which will carry B, half W.
Experiment justifies both these expectations.
The test may readily be made by mating the
black animals one by one with white ones.
The pure (or B B) black individual will pro-
duce only black offspring, whereas those not
pure, but B W in character, will produce off-
spring half of which on the average will be
black, the other half white. These two kinds
of dominant individuals obtained in the second
generation from a cross we may for conven-
ience call homozygous and heterozygous, fol-
lowing the convenient terminology of Bateson.
A homozygous individual is one in which like
characters are joined together, as B with B;
a heterozygous individual is one in which unlike
characters are joined together, as B with W.
It goes without saying that recessive individ-
4 37

HEREDITY

uals are always homozygous, as W W for ex-
ample. For they do not contain the dominant
character, otherwise they would show it.

It will be observed that in the cross of black
with white guinea-pigs black and white behave
as units distinct and indestructible, which may
meet in fertilization but separate again at the
formation of gametes. Mendel's law as illus-
trated in this cross includes three principles:
(1) The existence of unit-characters, (2) domi-
nance, in cases where the parents differ in a
unit-character, and (3) segregation of the units
contributed by the respective parents, this seg-
regation being found among the gametes formed
by the offspring.

The principles of dominance and segregation
apply to the inheritance of many characteristics
in animals and plants. Thus in guinea-pigs a
rough or ro-setted coat (Figs. 23 and 24) is domi-
nant over the ordinary smooth coat. If a pure
rough individual is crossed with a smooth one,
all the offspring are rough ; but in the next gen-
eration smooth coat reappears in one fourth of
the offspring, as a rule. Again, in guinea-pigs
and rabbits a long or angora condition of the

38

FIG. 20. — Radiograph of a hand similar to those shown in
Fig. 19. Notice the short, two-jointed fingers. (After
Farabee.)

MENDEL'S LAW OP HEREDITY

fur is recessive in crosses with normal short
hair. All the immediate offspring of such a
cross are short-haired, but in the next genera-
tion long hair reappears in approximately one
fourth of the offspring.

In cattle, the polled or hornless condition is
dominant over the normal horned condition;
in man, two- jointed fingers and toes (Figs. 19
and 20) are dominant over normal three-jointed
ones. This is clear from an interesting pedi-
gree given by Farabee of the inheritance of
the abnormality in a Pennsylvania family (see
Fig. 21). In no case was an abnormal mem-
ber of the family known to have married any
but an unrelated normal individual. It will
be seen that approximately half the offspring
throughout the four generations of offspring
shown in the table were of the abnormal sort,
— short-bodied and with short fingers and toes.

In each of the cases thus far considered a
single unit-character is concerned. Crosses in
such cases involve no necessary change in the
race, but only the continuance within it of two
sharply alternative conditions. But the result
is quite different when parents are crossed

39

c%..

0-.

<— eH

I Q ^

HO+ I

0+

o-

CH-
0+

O

CH
*0

CH-

o-

.a

13
O4- C

*O G

o* .2

"O

bo a

a s

03

•^ Q

40

FIG.23.

FIG. 22. — A smooth, dark guinea-pig.
FIG. 23. — A rough, white guinea-pig.
FIG. 24. — A dark, rough guinea-pig. The new
combination of characters obtained when
• animals are mated like those shown hi

MENDEL'S LAW OF HEREDITY

which differ simultaneously in two or more
independent unit-characters. Crossing them»be-
comes an active agency for the production of
new varieties.

In discussing the crosses now to be described
it will be convenient to refer to the various
generations in more precise terms, as Bateson
has done. The generation of the animals orig-
inally crossed will be called the parental gen-
eration (P) ; the subsequent generations will
be called filial generations, viz. the first filial
generation (Fx), second filial (F2), and so on.

When guinea-pigs are crossed of pure races
which differ simultaneously in two unit-charac-
ters, the Fj offspring are all alike, but the F2
offspring are of four sorts. Thus, when a
smooth dark animal (Fig. 22) is crossed with
a rough white one (Fig. 23) the Fj offspring
are all rough and dark (Fig. 24), manifesting
the two dominant unit-characters, — dark coat
derived from one parent, rough coat derived
from the other. But the F2 offspring are of
four sorts, viz. (1) smooth and dark, like one
grandparent, (2) rough and white, like the
other grandparent, (3) rough and dark, like

41

HEREDITY

the F! generation, and (4) smooth and white,
a new variety (Fig. 25). It will be seen that
the pigmentation of the coat has no relation
to its smoothness. The dark animals are either
rough or smooth, and so are the white ones.
Pigmentation of the coat is evidently a unit-
character independent of hair-direction, and as
.new combinations of these two units the cross
has produced two new varieties, — the rough
dark and the smooth white.

Again, hair-length is a unit-character inde-
pendent of hair-color. For if a short-haired
dark animal (either self or spotted, Fig. 26)
be crossed with a long-haired albino (Fig. 27),
the Fj offspring are all short-haired and dark
(Fig. 28) ; but the F2 offspring are of four
sorts, viz. (1) dark and short-haired, like one
grandparent, (2) white and long-haired, like
the other, (3) dark and long-haired, a new com-
bination (Fig. 29), and (4) white and short-
haired, a second new combination (compare
Fig. 25).

Now the four sorts of individuals obtained
from such a cross as this will not be equally
numerous. As we noticed in connection with

42

FIG. 25. — A smooth, white guinea-pig. A second new combination of
characters, but obtained first among the grandchildren of such an-
imals as are shown in Figs. 22 and 23.

FIG. 26. — A short-haired, pigmented guinea-pig. (" Dutch-marked "
with white.)

FIG. 27. — A long-haired, albino guinea-pig.

FIG. 28. — Offspring produced by animals of the sorts shown in Figs. 26
and 27. One shows the " Dutch-marked " pattern as a belt of
pale yellow ; the other does not. Both are short-haired and
pigmented (not albinos).

FIG. 29. — A long-haired, pigmented guinea-pig, " Dutch-marked "
with white. Its parents were like the animals shown in Fig. 28 ;
its grandparents like those shown in Figs. 26 and 27.

MENDEL'S LAW OF HEREDITY

the black-white cross, dominant individuals are
to the corresponding recessives as three to one.
Therefore, we shall expect the short-haired in-

Gamete

Gamete*

FlG. 30. — Diagram to explain the result of a cross between the
sorts of guinea-pigs shown in Figs. 26 and 27. L stands
for long hair, S for short hair, D for dark hair, and W for
white hair. Dominant characters are indicated by heavy
type.

dividuals in F2 to be three times as numerous
as the long-haired ones, and dark ones to be
three times as numerous as white ones. Fur-
ther, individuals which are both short-haired

43

HEREDITY

and dark should be 3 X 3 or 9 times as nu-
merous as those which are not. The ex-
pected proportions of the' four classes of F2
offspring are accordingly 9:3:3:1, a propor-
tion which is closely approximated in actual
experience. The Mendelian theory of inde-
pendent unit-characters accounts for this re-
sult fully. No other hypothesis has as yet been
suggested which can account for it.

Suppose that each unit has a different mate-
rial basis in the gamete. Let us represent the
material basis of hair-length by a circle, that
of hair-color by a square, then combinations
and recombinations arise as shown in Fig. 30.
The composition of the gametes furnished by
the parents is shown in the first line of the
figure; that of an F± individual (or zygote),
in the second line; that of the gametes
formed by the Fj individual in the third line.
L meets S and W meets D in fertilization to
form an Fx individual double and also hetero-
zygous as regards hair length and hair color,
but these units segregate again as the gametes
of the Fj individuals are formed, and it is a mat-
ter of chance whether or not they are associated

44

MENDEL'S LAW OF HEEEDITY

as originally, L with W and S with D, or in a
new relationship, L with D and S with W.
Hence we expect the Fx individuals to form
four kinds of gametes all equally numerous, —
L W, S D, L D, and S W. By chance unions of
these in pairs nine kinds of combinations be-
come possible, and their chance frequencies will
be as shown in Fig. 31. Four of these com-
binations, including nine individuals, will show
the two dominant characters, short and dark;
two classes, including three individuals, will
show one dominant and one recessive charac-
ter, viz. dark and long; two more classes, in-
cluding three individuals, will show the other
dominant and the other recessive character,
viz. short and white; and lastly, one class, in-
cluding a single individual, will show the two
recessive characters, long and white. The four
apparent classes, or, as Johannsen calls them,
phenotypes, will accordingly be as 9:3:3:1.
This is called the normal Mendelian ratio for
a dihybrid cross, — that is, a cross involving
two unit-character differences.

One individual in each of these four classes
will, if mated with an individual like itself,

45

HEREDITY

breed true, for it is homozygous, containing
only like units. The double recessive class,
long white, of course contains only homozygous
individuals, but in each class which shows a
dominant unit, heterozygous individuals out-
number homozygous ones, as 2 : 1 or 8 : 1. Now
the breeder who by means of crosses has pro-

Short Dark.

Long Dark.

Short White.

Long White.

i S D. S D

1 L D. 1

- D

15^

7. S w

1 L W.

L W

2 S D. L D

3 L D. 1

'* w

8 S v

7. L W

2 S D. S w

£S D. L w

9

~9

¥

T

FIG. 31. — Diagram showing the kinds and relative frequencies
of the young to be expected in F2 from the crossing of
animals shown in Figs. 26 and 27.

duced a new type of animal wishes, of course,
to " fix " it, — that is, to obtain it in a condi-
tion which will breed true. He must, therefore,
obtain homozygous individuals. If he is deal-
ing with a combination which contains only
recessive characters, this will be easy enough,
for such combinations are invariably homozy-
gous. His task will become increasingly diffi-
cult the more dominant characters there are
included in the combination which he desires
to fix.

46

FIG. 32. — A long-haired, rough albino guinea-pig ;
male, 2002.

MENDEL'S LAW OF HEREDITY

The most direct method for him to follow is
to test by suitable matings the unit-character
constitution of each individual which shows the
desired combination of characters, and to reject
all which are not homozygous. In this way a
pure race may be built up from individuals
proved to be pure. Such a method, however,
though sure, is slow in cases where the desired
combination includes two or more dominant
unit-characters, for it involves the application
of a breeding test to many dominant individ-
uals, most of which must then be rejected. It
is therefore often better in practice to breed
from all individuals which show the desired
combination, and eliminate from their offspring
merely such individuals as do not show that
combination. The race will thus be only grad-
ually purified, but a large stock of it can be
built up much more quickly.

We, may next discuss a cross in which three
unit-character differences exist between the par-
ents, instead of two. If guinea-pigs are crossed
which differ simultaneously in three unit-char-
acters, color, length, and direction of the hair,
a still larger number of phenotypes is obtained

47

HEREDITY

in F2, namely, eight. A cross between a short-
haired, dark, smooth guinea-pig (compare Fig.
22) and one which was long-haired, white, and
rough (Fig. 32) produced offspring in Fx which
were short-haired, dark, and rough (compare
Fig. 24), these being the three dominant char-
acters, two derived from one parent, one from
the other. The F2 offspring were of eight dis-
tinct types, two like the respective grandpar-
ents, one like the Fj individuals (parents), and
the other five new, shown in Fig. 33. They are
short white rough, short white smooth, long
white smooth, long dark rough, and long dark
smooth. The largest of the eight apparent
classes (phenotypes) was the one which mani-
fested the three dominant characters, short,
dark, and rough, which had been the exclusive
Fj type; the smallest class was the one which
manifested the three recessive characters, long,
white, and smooth. Theoretically these two
classes should be to each other as 27 : 1. Of
the twenty-seven triple-dominants, twenty-six
should be heterozygous.

A comparison of this case with the one just
previously described shows what an increas-

48

FIG. 33. — Five new combinations of unit-characters obtained in gen-
eration F2, by crossing the animal shown in Fig. 32 with animals
like that shown in Fig. 22.

MENDEL'S LAW OF HEEEDITY

ingly difficult thing it is to fix types obtained by
crossing, if the number of dominant characters

Omnocen

FIG. 34. — Diagram to show the gametic combinations and
segregations involved in a cross between guinea-pigs differ-
ing in three unit-characters. L stands for long hair, S for
short hair, W for white hair, and D for dark hair; R for
rough, and Sm for smooth coat. Compare Figs. 22 and 32.

in the selected type increases. On the theory of
unit-characters the gametic combinations and
segregations involved in this cross are as shown

49

HEEEDITY

in Fig. 34. The nature of the gametes formed
by the parents crossed is shown in the first row ;
the composition of the Fj individuals, immedi-
ately below. In the two lower rows are shown
four different sorts of gametic splittings which
may occur in the Fx individuals, producing thus
eight different kinds of gametes. If, in re-
ality, the F! individuals form eight kinds of
gametes, all equally numerous, and chance
unions in pairs occur among them, there should
be produced eight corresponding sorts of indi-
Viduals numerically as 27:9:9:9:3:3:3:1.
In a total of 64 individuals there should be on
the average one pure individual in each of the
eight different classes. The class numerically 27
in 64 manifests three dominant characters ; those
which are numerically 9 in 64 manifest two domi-
nant characters ; those which are numerically 3
in 64 manifest one dominant character. Among
each of these there will be on the average one
pure individual, but the class which contains 1
individual in 64 is a pure recessive, £br it contains
no dominant character. This combination, then,
requires no fixation. It will breed true from the
start.

50

MENDEL'S LAW OF HEEEDITY

BIBLIOGRAPHY

CASTLE, W. E.

1903. " Mendel's Law of Heredity." Proc. Am. Acad.
Arts and Sci., 38, pp. 535-548; also, Science, N. S., 18,
pp. 396-^06.

1905. " Heredity of Coat-characters in Guinea-pigs and
Rabbits." Carnegie Institution of Washington, Publica-
tion No. 23, 78 pp., 6 pi.
FARABEE, W. C.

1905. " Inheritance of Digital Malformations in Man."
Papers of the Peabody Museum, Harvard University, vol.
3, No. 3, pp. 69-77, 5 pi.

CHAPTER IV

THE DETERMINATION OF DOMINANCE; HETEKOZY-
GOUS CHARACTERS AND THEIR FIXATION; ATA-
VISM OR REVISION

WE have noticed that when a black
guinea-pig of pure race is mated
with a white one, only black off-
spring are produced; and that when rough-
coated guinea-pigs are mated with smooth-
coated ones, only rough-coated young are pro-
duced; and that when short-haired guinea-
pigs are mated with long-haired ones, only
short-haired young are produced. The char-
acter which in each case is seen in the young
we call dominant, that which is unseen we
call recessive. Thus black is dominant over
white, rough coat over smooth coat, and short
coat over long coat.

A question which has given much concern to
students of heredity is this, — upon what does

52

DOMINANCE

dominance depend? Why should black domi-
nate white rather than the reverse?

In poultry, indeed, the relations are often
reversed, white dominating black. Why is
this? Several attempted explanations have
been made, but none of them is thoroughly
satisfactory. The one which has found wid-
est acceptance is this : In the dominant in-
dividual something is present which is want-
ing in the recessive. Thus, in the black guinea-
pig there is present some ferment-like body
or some ingredient of black which is wanting
in the albino. Accordingly, the black guinea-
pig forms pigment, a thing which the albino
can do only feebly or not at all. The distinc-
tive something of the black parent therefore
dominates a corresponding nothing of the
white parent. White fowls, on the other hand,
are not albinos. They have pigmented eyes.
Accordingly they do not lack the power to
form pigment, owing to the absence of some
necessary ferment or pigment ingredient.

White guinea-pigs occur which are in a way
comparable with white fowls. They look ex-
actly like albinos, except that their eyes are
5 53

black, whereas the eyes of the albino are pink.
If such a black-eyed white guinea-pig is crossed
with an albino of the sort shown in Fig. 15,
the young produced will be black all over.
Now this result shows that the black-eyed
white animal possesses what is lacking in the
albino as compared with the all-black animal.
It would seem, therefore, that it lacks some-
thing different from what the albino lacks, and
that a cross of the two supplies both lacks, the
albino supplying what is wanting in the black-
eyed white, and vice versa. Accordingly, wholly
black offspring result from the crossing of the
two white races.

But the case of white poultry is different
from this, since white poultry lack nothing
that is necessary to produce the complete
black plumage. For when white fowls crossed
with black ones produce white offspring, if
these offspring are then bred with each other,
they produce both white offspring and black
ones in the ratio 3 to 1. White fowls, there-
fore, are able to produce the black condition.
This ability is in the white individual held in
abeyance, it is not exercised. Why, we do

54

not know. Some suppose it to be held in
check by an additional unit-character, an in-
hibiting factor, but we have no direct evidence
that such a factor exists. All that we are
warranted in saying at 'the present time is
that black and white in poultry represent dif-
ferent conditions of pigmentation, alternative
to each other in heredity. In crosses of the
two, white is ordinarily dominant over black,
but in crosses between certain strains of white
and black poultry this relationship is reversed,
as Bateson has shown.

In still other cases, a cross of white with
black fowls produces offspring which resemble
neither parent closely, but which are in reality
intermediate. They are known as blue or
Andalusian fowls. They manifest a dilute
condition of black, such as one might obtain
by mixing lampblack with flour; they are in
reality a fine mosaic of black with white.
Such a condition has thus far been obtained
only from a cross of black fowls with a pecu-
liar strain of impure sooty whites. This strain
undoubtedly contains the mosaic pattern but
without sufficient black pigment to make it

55

HEREDITY

plainly visible. A cross with a black race
makes it visible. No one, however, has suc-
ceeded in " fixing ': a blue race, that is,
in obtaining a strain which would "breed
true.

When two blue individuals are bred together
they, produce black, blue, and white offspring
in the ratio 1:2:1. The blacks are homo-
zygous, B B ; the whites also are homozygous,
WW, but the blues are invariably heterozy-
gous, B W. Blue accordingly in this case is
called a heterozygous character, one which is
due to the presence in one zygote of two unlike
unit-characters, which invariably segregate from
each other at the genesis of gametes, but which
jointly produce a different appearance from
what either produces by itself. If a strain of
Andalusian fowls should ever be secured which
would breed true, it would have to come about
by the association of black with white in a
non-segregating relationship, so that both
would be transmitted in the same gamete.
That is, one would have to secure in the same
gamete with white enough black pigment to
bring out the latent mosaic pattern, and fur-

56

DOMINANCE

ther, one would have to secure a homozygous
race of fowls which formed such gametes.

Success would be most likely to attend the
experiment if one selected always the sootiest
whites obtained from blue parents, for blue
results, as we have seen, from the association
of more black with the white and in the pat-
tern borne already by the white race.

A much-debated case of inheritance which
involves this principle of unfixable heterozy-
gous characters occurs among fancy mice, in
the variety known as yellow. A wonderful
series of color varieties exists among mice
kept as pets, equalling or perhaps surpassing
that known in the case of any other mammal.
All these varieties appear to be derivatives
of the common house-mouse, with which they
cross readily. All are capable of explanation
as unit-character variations from the condition
of the house-mouse. Among all these varie-
ties yellow is most peculiar in its behavior.
In crosses it is dominant over all others, yet
is itself absolutely unfixable.

If certain strains of yellow mice are crossed
with black ones, the offspring produced are of

57

HEREDITY

two sorts equally numerous, yellow and black.
From this result alone it is impossible to say
which is the dominant character, but breeding
tests of the offspring show that yellow is the
dominant character. For the black offspring
bred together produce only black offspring,
but the yellows bred together produce both
yellow offspring and black ones. The curious
feature of the case is that when yellows are
bred with each other no pure yellows, that is,
homozygous ones, are obtained. Hundreds of
yellow individuals have been tested, but the
invariable result has been that they are found
to be heterozygous; that is, they transmit yel-
low in half their gametes, but some other color
in the remaining gametes — it may be black,
or it may be brown, or it may be gray. The
black, brown, or gray animals obtained by
mating yellow with yellow mice never produce
yellow offspring if mated with each other.
This shows that they are genuine recessives
and do not contain the yellow character, which
is dominant.

Now ordinary heterozygous dominants, when
mated with each other, produce three domi-

58

DOMINANCE

nant individuals to one recessive. Accordingly
we should expect yellow mice, if, as stated,
they are invariably heterozygous, to produce
three yellow offspring to one of a different
color, but curiously enough they do not. They
produce two yellows (instead of the expected
three) to every one of a different color. About
the ratio there can be no reasonable doubt.
It has been determined with great accuracy
by my pupil, Mr. C. C. Little, who finds that
in a total of over twelve hundred young pro-
duced by yellow parents almost exactly two-
thirds are yellow. Instead of the regular
Mendelian ratio 3:1, we have then in this case
the peculiar ratio 2:1, and this requires ex-
planation. The explanation of this ratio is to
be found in the same circumstance as is the
total absence of pure yellows. Pure yellow
zygotes are indeed formed, but they perish for
some unaccountable reason. For a yellow in-
dividual forms gametes of two sorts with equal
frequency, viz. yellow and non-yellow (let us
say black). For, if yellow individuals are
mated with black ones, half the offspring are
black, half yellow, as already stated.

59

HEREDITY

If now yellow individuals are mated with
each other we expect three sorts of young to
be produced numerically, as 1:2:1, viz. 1 Y Y,
2 Y B, and IBB. But since observation shows
that only two combinations are formed which
contain yellow tos one not containing yellow,
and since further all yellows which survive
are found to be heterozygous (YB), it must
be that the expected YY individual either is
not produced or straightway perishes. As to
which of these two contingencies happens we
also have experimental evidence. Mr. Little
finds that yellow mice when mated to black
ones produce larger litters of young than when
they are mated to yellow ones. The average-
sized litter contains something like 5.5 young
when the mate is a black animal, but only 4.7
when it is a yellow animal. It is evident, then,
that about one young one out of a litter per-
ishes when both parents are yellow, and this
undoubtedly is the missing yellow-yellow zy-
gote. The yellows which are left are hetero-
zygous yellow-black zygotes, and they are to
those that perish as 2:1. They are also to
the non-yellow zygotes as 2:1, the ratio ob-

60

DOMINANCE

served also among the surviving young of
yellow by yellow parents.

This interpretation of the 2 : 1 ratio observed
in this case is strongly supported by a similar
case among plants, in which the evidence is
even more complete. A so-called " golden '
variety of snapdragon, one in which the foli-
age was yellow variegated with green, was
found by the German botanist, Baur, to be
unfixable, producing when self -pollinated fully
green plants as well as golden ones, in the
ratio 2 golden : 1 green. The green plants were
found to breed true, that is, to be recessives,
while the golden ones were invariably found
to be heterozygous. Baur found, however, by
germinating seeds of golden plants very care-
fully, that there were produced in addition to
green plants and golden ones a few feeble
seedlings entirely yellow, not variegated with
green, as the golden plants are. These, for
lack of assimilating organs (green chlorophyl),
straightway perished. Clearly they were the
missing pure yellow zygotes.

Some Mendelian characters, while not them-
selves heterozygous and so unfixable, are never-

61

HEEEDITY

theless produced only when two independently
inherited factors are present together. A
character of this sort does not itself conform
with the simple Mendelian laws of inheritance,
but its factors do. Herein lies the explanation
of atavism or reversion, and the process by
which reversionary characters may be fixed.

Atavism or reversion to an ancestral con-
dition is a phenomenon to which Darwin re-
peatedly called attention. He realized that it
is a phenomenon which general theories of
heredity must account for. He supposed that
the environment was chiefly responsible for the
reappearance in a species of a lost ancestral
condition, but that in certain cases the mere
act of crossing may reawaken slumbering an-
cestral traits. Thus he noticed that when
rabbits of various sorts are turned loose in a
warren together, they tend to revert to the
gray-coated condition of wild rabbits. And
when pigeons are crossed in captivity they
frequently revert to the plumage condition of
the wild rock pigeon, Columba livia. In plants,
too, Darwin recognized that crossing is a fre-
quent cause of reversion. The explanation

62

DOMINANCE

which he gave was the best that the knowl-
edge of his time afforded, but it leaves much
to be desired. This lack, however, has been
completely supplied by the Mendelian princi-
ples. An illustration or two may now be
cited.

When pure-bred black guinea-pigs are mated
with red ones, only black offspring are as a
rule obtained. The hairs of the offspring do
indeed contain some red pigment, but the black
pigment is so much darker that it largely
obscures the red. In other words, black be-
haves as an ordinary Mendelian dominant. In
the next generation black and red segregate
in ordinary Mendelian fashion, and the young
produced are in the usual proportions, three
black to one red, or 1 : 1 in back-crosses of the
heterozygous black with red. All black races
behave alike in crosses with the same red in-
dividual, but among red animals individual
differences exist. Some, instead of behaving
like Mendelian recessives, produce in crosses
with a black race a third apparently new con-
dition, but in reality a very old one, the agouti
type of coat found in all wild guinea-pigs, as

63

HEREDITY

well as in wild rats, mice, squirrels, and other
rodents. In this type of coat reddish yellow
pigment alone is found in a conspicuous band
near the tip of each hair, while the rest of the
hair bears black pigment. The result is a
brownish or grayish ticked or grizzled coat,
inconspicuous, and hence protective in many
natural situations.

Some red individuals produce the reversion
in half of their young by black mates, some
in all, and others, as we have seen, in none,
this last condition being the commonest of the
three. It is evident that the reversion is due
to the introduction of a third factor, additional
to simple red and simple black. It is evident
further that this new third factor, which we
will call A (agouti), has been introduced
through the red parent, and that as regards
this factor, A, some individuals are homozy-
gous (AA) in character, others are heterozy-
gous (transmit it in half their gametes only),
while others lack it altogether. Further ob-
servations show that it is independent in its
inheritance of both black and red; it is in
fact an independent Mendelian character, which

64

DOMINANCE

can become visible only in the presence of both
black and red, because it is a mosaic of those
two pigments. If the Fx agouti individuals
are bred together they produce in the next
generation (F2) three sorts of young, viz.
agouti, black, and red, which are numerically
as 9:3:4. This evidently is a modification of
the dihybrid Mendelian ratio 9:3:3:1, result-
ing from the fact that the last two classes are
superficially alike. They are red animals with
and without the agouti factor respectively; but
this agouti factor is invisible in the absence
of black, so that both sorts of reds look alike.
Together they number four in sixteen of the
F2 offspring.

Fig. 35 is intended to show how the inde-
pendent factors behave in heredity. The black
parent contributes the factor B, the red par-
ent, R and A, so that the zygote, or new indi-
vidual, contains the three factors necessary to
the production of agouti. When the new in-
dividual forms gametes (sex-cells), these will
be of four different kinds, for A is independ-
ent of B and of R and may pass out with
either one in the reduction division which sepa-

65

HEREDITY

= Agouti (reversion)

BL RA) BA (RjGametes

[BRA)

(RABA)

(BJBA)

(BAJBA)

x^x^*/

i

^-— - •*-— *^

1

|

AoouLi

D

Agouti

Agouti

Agouti (fixed)

Black

Fed

Black

Red

4

2

2

1

FIG. 35. — Diagram to show the gametic combination and re-
combinations which occur in the production and fixation of
an atavistic coat-character in guinea-pigs.

Row 1 shows the character of the gametes formed by the
parents crossed; row 2 shows the character of the FI agouti
individuals resulting from the cross; row 3 shows the two
different sorts of gametic splittings which may occur in the
production of gametes by the FI agoutis, and how four differ-
ent kinds of gametes result; row 4 shows how among such
gametes four different kinds of unions may occur that will
produce agouti young. The BA-BA combination, it will
be understood, could result only from the union of a BA
gamete with another gamete of like constitution. Below
each of the four combinations is indicated the kinds of young
which an animal of that sort would produce if mated with
an animal like itself. The numerals show the expected
relative frequencies of the four sorts of combinations.

66

DOMINANCE

rates B from E. That division accordingly
may occur either so as to form gametes B and
E A respectively, or what is equally probable,
so as to produce gametes B A and E. Obser-
vation confirms this interpretation, for it is
found that the reversionary agoutis do not
breed true, but produce young of the three
sorts, agouti, black, and red, as expected. We
expect black individuals from unions of B with
B, or of B with E; we expect red individuals
from unions of E with E or with E A, and
from unions of E A with E A ; we expect agoutis
to be produced by any gametic union which
brings together the three factors B, E, and A.
There are six chances in sixteen for the oc-
currence of such a union, when the rever-
sionary agoutis are bred together. In fact,
however, agoutis are produced much oftener.
Approximately nine out of sixteen of the
young have been found to be agoutis. The
unexpected excess of agoutis in our experi-
ments was fully explained when these second-
generation agoutis were tested individually.
It was then found that they are of four sorts
as regards breeding capacity. The first sort

67

HEREDITY

produces the three kinds of young, agouti, black,
and red; the second sort produces only agouti
young and red young; the third sort produces
only agouti young and black young. The fourth
sort produces only agouti young, i. e. repre-
sents the fully fixed agouti type, the completely
recovered wild type.

In the chart (Fig. 35) are indicated cer-
tain gametic unions which would lead to the
production of these four classes of agoutis.
The probable frequencies of their occurrence
on the basis of chance are 4:2:2:1.

Experiment made it clear that R as an inde-
pendent gametic factor is not necessary to the
production of the agouti character, as was at
first thought to be the case, but that any
gametic union which includes both B and A
will produce an agouti individual whether
R is or is not present. Yet a microscopic
examination of the agouti hair shows that
red pigment is present in a distinct band
near the hair-tip. As a matter of fact all
black individuals, even when they breed true,
probably form some red pigment along with
the black, but its presence is overlooked when

68

DOMINANCE

the more opaque black is distributed through-
out the whole length of the hair. When, how-
ever, black is excluded from the hair-tip, the
red then becomes visible as the agouti mark-
ing; elsewhere the hair appears black. Red,
then, we may assume, is always present with
black in sufficient quantities to produce the
agouti marking if the factor A is present (ab-
sence of black from the hair-tip). This ex-
plains why blacks never give the reversion in
any sort of cross, but it is always brought
about through the agency of the red parent.
If a black individual contained the factor A,
it would no longer be a black individual, but
an agouti one.

The existence of a third factor, A, in cases
of reversion in coat-character among rodents
was long overlooked merely because it does
not represent a distinct pigment or set of pig-
ments, but consists in a particular kind of
pigment distribution on the individual hairs.
The agouti hair is due to a definite cycle of
activity of the hair follicle in forming its pig-
ments,— first black, then red, then black; the
wholly black hair is due to a continuous process
6 69

HEEEDITY

of pigment formation without alternation in the
character of the pigments produced.

In rabbits as well as in guinea-pigs rever-
sion to the original wild type, in this case
gray, may be obtained by crossing a black ani-
mal with a yellow one. In guinea-pigs the
yellow (or red) animal which will yield this
result cannot be distinguished in appearance
from one which will not; but in rabbits the
yellow animal which will give reversion has
a white belly and tail, while the one which
will not give reversion is not so distinguished.

We now know what is implied in the fixa-
tion of a heterozygous character obtained by
crossing. When A and B are crossed we ob-
tain a third condition, C. G is due either to
the simple coexistence of A with B, or to the
coexistence with them of a third factor intro-
duced with one or the other. In either case
fixation will consist in getting into the gamete
all the factors necessary to the production of
C. In the first supposed case the zygote is
A-B and the resultant is equivalent to C. Fix-
ation will consist in getting a zygote of the
formula AB • AB. In the second supposed

70

DOMINANCE

case the zygote produced is either A-CB or
AC -B ; fixation will consist in obtaining a zy-
gote ACB-ACB; every gamete formed will
then contain the three factors A, C, and B.

BIBLIOGRAPHY

BATESON, W.

1909. "Mendel's Principles of Heredity," 393 pp., illus-
trated. University Press, Cambridge; also G. P. Put-
nam's Sons, N. Y. [Contains translation of Mendel's
original papers.]

BAUR, E.

1907. " Untersuchungen tiber die Erblichkeitsverhaltnisse
einer nur in Bastardform lebensfahigen Sippe von An-
tirrhinum majus." Ber. d. Deuisch Bot. Gesellsch., 25,
p. 442.
CASTLE, W. E.

1907. "On a Case of Reversion Induced by Cross-Breed-
ing and its Fixation." Science, N. S., 25, pp. 151-153.

1907. "The Production and Fixation of New Breeds."
Proc. Amer. Breeders' Ass'n, 3, pp. 34-41.

CASTLE, W. E. and LITTLE, C. C.

1910. "On a Modified Mendelian Ratio Among Yellow
Mice. " Science, N. S., 32, pp. 868-870.

CUENOT, L.

1908. "Sur quelques anomalies apparentes des pro-
portions Mendelie'nnes." Arch. Zool. Exper. (4), Notes
et Revue, p. vii.

DAVENPORT, C. B.

1906. "Inheritance in Poultry." Carnegie Institution
of Washington, Publication No. 52, 104 pp., 17 pi.

1909. "Inheritance of Characteristics in Domestic Fowl."
Carnegie Institution of Washington, Publication No.
121, 100 pp., 12 pi.

71

CHAPTER V

EVOLUTION OP NEW KACES BY LOSS OB GAIN OF
CHARACTERS

OUR knowledge of Mendelian phenom-
ena is most complete in the case of
color-inheritance. We find that the
flower-colors of plants and the coat-colors of
mammals are alike complex, and that what
seem at first sight simple results may really
depend on several independent factors acting
conjointly. By analysis of such complex cases
we are able to gain some idea of what the
probable course of evolution has been in the
production of the color varieties found among
cultivated plants and domesticated animals.

Thus among rodents (mice, rabbits, guinea-
pigs) the coat is grayish, consisting of black,
brown, and yellow pigments mingled together
on the same individual hair in a pattern of
greater or less complexity.

72

EVOLUTION BY LOSS

The simplest variation from this ancestral
type of coloration is albinism, a wholly unpig-
mented condition in which the eyes are pink.
This is due to the loss of the capacity to form
pigment. Albinism is recessive in crosses. We
explain it by assuming that something neces-
sary for color production is wanting in the
albino, and call that something the color-factor
C, without necessarily making any assumption
as to its nature. Another common variation
is the loss of the pattern-factor of the indi-
vidual hair, the agouti or A factor. An ac-
count of the discovery of this factor was given
in the last chapter. In consequence of the loss
of this factor the pigments become mingled
together without order, and the result is a
uniform black, the denser pigment hiding the
others.

A third variation is the loss of the capacity
to form black pigment (factor B), only brown
and yellow pigments being left. Thus arise
brown and cinnamon varieties. Through these
three independent loss-variations there arise
eight different color-varieties as follows:

73

HEEEDITY

Gray (or agouti) = C B Br A; Cinnamon = C Br A

Black = C B Br; Brown = C Br

Albino (1) = B Br A; Albino (3) = Br A

Albino (2) = B Br; Albino (4) = Br

Proof of the correctness of this interpretation
may be obtained from crosses. Suppose the
four kinds of albinos described be crossed with
the same colored variety, brown; albino 1 will
produce gray offspring, albino 2 will produce
black ones, albino 3 will produce cinnamon
ones, and albino 4 will produce brown ones.
The cross with albino 1 brings together all the
four factors entering into the production of
gray, viz. C, B, Br, and A, hence the young are
gray. The cross with albino 2 brings together
the factors C, B, and Br only. The result is
black. The cross with albino 3 brings together
the factors C, Br, and A; result, a cinnamon
animal. The cross with albino 4 brings to-
gether no factors except C and Br; result, a
brown animal.

Thus far we have considered merely varia-
tions which arise by loss of one or more of
the three unit-characters, A, B, and C. We
may now consider variations which arise

74

EVOLUTION BY LOSS

by modification without loss of these same
factors.

Yellow varieties owe their origin to a re-
duction in the amount of black or brown pig-
ment in the fur, and to a corresponding in-
crease in the amount of yellow. In some
yellow animals, such as the sooty yellow rab-
bit, black and brown pigments are not wholly
lacking in the fur, but are only greatly re-
duced in amount. They always persist in
the eye. In other yellow animals, mice for
example, the black or brown pigments are
wholly absent from the fur, and they may also
be greatly reduced in amount in the eye, as
in the variety known as pink-eyed yellow, but
in no yellow animal, so far as I am aware, is
the production of black and of brown pigments
wholly suppressed.

In any mammal which possesses yellow varie-
ties we can produce by suitable crosses as many
different varieties of yellows as there are of
gray, black, cinnamon, and brown varieties
combined. For example, in mice, yellow indi-
viduals of which, as was shown in the last
chapter, are invariably heterozygous and pro-

75

duce some other variety than yellow, even when
mated with yellows, we can recognize the fol-
lowing varieties distinct in breeding capacity,
though all looking very similar.

1. Yellows which produce yellow young and gray
ones;

2. Yellows which produce yellow young and black
ones;

3. Yellows which produce yellow young and
cinnamons;

4. Yellows which produce yellow young and brown
ones.

Albino varieties occur which correspond with
each of these yellow varieties, viz. (1) albinos
which if crossed with brown will produce yel-
low young and gray ones; (2) albinos which
crossed with brown produce yellow young and
black ones ; (3) albinos which crossed with brown
produce yellow young and cinnamon ones ; and
(4) albinos which crossed with brown produce
yellow young and brown ones. Such albinos,
of course, differ from the corresponding yellow
varieties merely by the general color factor C,
which the albino lacks. If this is added by a
cross, they produce the same visible result as

76

EVOLUTION BY LOSS

the corresponding yellow variety in the same
cross.

In addition to the modification which pro-
duces yellow varieties, we can recognize sev-
eral other modified conditions of the unit-
characters A, B, C, and Br, which modifica-
tions produce whole series of color varieties.
For a modified condition of a single unit-
character is capable of producing as many
new varieties as there are possible combina-
tions of the modified character with other unit
characters.

One who attends a poultry-show cannot fail
to be impressed with the great number of color
varieties among poultry. Let him first observe
these among fowls of common size, and if he
then visits the bantam section he will find them
all duplicated in miniature among the bantams.
If a new color variety is brought out, it is only
a short time until it finds its place among the
bantams as well as among fowls of common
size. The dwarf size of the bantam is clearly
due to a modified condition of one or more unit-
characters capable of combinations with as many
different kinds of coloration as occur among

HEREDITY

poultry. The various combinations are of
course brought about by crossing, and two gen-
erations suffice theoretically for securing them.

In mice, if one possessed only the albino
variety last described, — the one which corre-
sponds with the brown-eyed yellow variety, —
he could easily produce within six months every
one of the various color varieties which have
been mentioned. All he would have to do
would be to catch some wild mice and cross
these with his albinos. The immediate off-
spring produced by the cross might seem un-
promising; they would either be gray, exactly
like wild mice, or else yellow. But if our
breeder possessed the faith to breed a second
generation from these animals, he would be
rewarded by seeing all the color varieties which
I have described put in an appearance, viz.
yellows with black eyes, and yellows with
brown eyes, blacks, browns, cinnamons, and
grays, and albinos corresponding in character
with each colored variety except for the lack
of the color-factor C.

It may be of interest to consider how some
additional color varieties of mice have arisen,

78

EVOLUTION BY LOSS

for of all mammals bred in captivity the mouse
is probably richest in color varieties. In one
series of these the capacity to form black or
brown pigments is greatly weakened, so that
the coat is less heavily pigmented and the eye
is almost wholly unpigmented, and looks pink,
due to the red color of the blood in the eye.
This series we may call the pink-eyed series.
All the common color varieties occur in a pink-
eyed as well as in a dark-eyed series. Thus
there are pink-eyed grays, pink-eyed blacks,
pink-eyed cinnamons, pink-eyed browns, and
pink-eyed yellows, as well as albinos which
transmit the pink-eyed condition in crosses.

Given a single pink-eyed individual in any
one of these varieties, all the others may be
produced from it by suitable crosses. Thus
a pink-eyed gray crossed with brown produces
in F! reversion to the condition of the wild
house-mouse, but in F2 (that is, among the
grandchildren) occur eight varieties, — four
dark-eyed and four pink-eyed. Gray, black,
cinnamon, and brown occur, both in dark-eyed
and in pink-eyed individuals, the latter being
also far lighter in color than the dark-eyed

79

HEEEDITY

varieties. The pink-eyed condition is there-
fore in mice a unit-character modification of
the pigmentation, independent of any of the
pigment factors previously mentioned, since
it can be transferred by crosses from asso-
ciation with one of these to association with
another. It may also be transmitted equally
well through colored and through albino in-
dividuals, though it produces a visible effect
only in colored individuals.

Another unit-character modification of the
pigmentation seen in mice produces a series
of dilute or pale pigmented varieties, but dif-
ferent in character from the pink-eyed series,
since their eyes may be dark, not pink. The
pale modification of gray is known to fan-
ciers as " blue-gray," that of black is known
as " blue," and that of brown is known as
" silver fawn." The pale quality is inter-
changeable between black, brown, and yellow
pigmentation, so that if one has a pale gray
variety he may by crosses obtain also pale
black, pale cinnamon, pale brown, and pale
yellow varieties. Or if one starts with pale
yellow, he may by crosses with a perfectly

80

EVOLUTION BY LOSS

wild mouse obtain also pale gray, pale black,
pale cinnamon, and pale brown varieties, all
within two generations from the cross.

Now the pale modification is distinct from
the pink-eyed modification, and independent of
it in transmission. Accordingly, it is possible
to have the two modifications combined in the
same race. Thus arises a series of pale pink-
eyed grays, blacks, cinnamons, browns, and
yellows. Since paleness is in crosses recessive
to intense pigmentation, and pink eyes are re-
cessive to dark ones, it follows that a variety
which is both pale and pink-eyed will breed
true to those characteristics without fixation.

The lightest colored of the pale pink-eyed
varieties develop very little pigment indeed,
yet the modifications to which they are due
are wholly different in nature from the albino
variation, as a very simple experiment will
show. Cross together an albino of variety (1),
page 74, — which is a snow-white animal with
pink eyes, — and a pale pink-eyed, brown ani-
mal, whose coat is pale straw color, and whose
eyes, like those of the albino, are pink. Although
both parents are pink-eyed, and one develops no

81

pigment whatever in its fur, while the other
develops very little, nevertheless the offspring
are as dark as the darkest wild mice, eyes,
fur, and all. They look just like common house-
mice. This result shows that the albino varia-
tion is something very different in nature from
the modifications found in the pink-eyed brown
parent, since each parent contains those con-
stituents of the wild gray coat which the other
parent lacks.

I can think of no more instructive labora-
tory experiment illustrative of Mendelian in-
heritance than to follow through two genera-
tions the cross just described, and to analyze
critically the results obtained. One who does
this can never be sceptical about the value of
crossing as an agency in the production of new
varieties. For in the second generation from
the cross he will obtain (1) ordinary gray,
black, cinnamon, and brown varieties; (2) pale
gray, black, cinnamon, and brown varieties;
(3) pink-eyed gray, black, cinnamon, and brown
varieties; (4) pink-eyed and pale gray, black,
cinnamon, and brown varieties; and lastly,
albinos, which, if he has the patience to test

82

EVOLUTION BY LOSS

them one by one, will prove to be of sixteen
different homozygous kinds, to say nothing of
the much more numerous heterozygous sorts.

No mention has thus far been made of spotted
races, in which a unit-character modification has
occurred which results in a distribution of pig-
ment to part of the coat only, the remainder
being unpigmented. Although this modifica-
tion apparently regulates the distribution of
pigment over the body, it is independent of the
general color factor C, since it is transmitted
through albinos, which by hypothesis lack C.

Spotting is also independent of all the other
unit-character modifications which have been
described. Consequently we have in mice four
different series of spotted varieties, — the in-
tense spotted, the dilute spotted, the intense
pink-eyed spotted, and the dilute pink-eyed
spotted. In each of these series are gray,
black, cinnamon, brown, and yellow individuals,
making a total of twenty spotted sorts, all of
which may be obtained from crossing a single
pair of properly selected parents, such, for
example, as an albino and a wild house-mouse
of the kind every barn contains.

83

HEREDITY

The color variations of guinea-pigs are simi-
lar to those of mice; the same series of unit-
character changes has produced them with one
exception. The pink-eyed modification is want-
ing in guinea-pigs. We are therefore limited
here to the intense series, the pale series, the
intense spotted series, and the pale spotted
series. In each of these occur gray (or agouti)
individuals, black ones, cinnamon ones, and
brown ones.

The parallelism between the color variations
in guinea-pigs and in mice received an inter-
esting demonstration in a particular case. The
brown pigmented series in mice has been
known for some time, but in guinea-pigs the
brown variety is of comparatively recent origin,
and the cinnamon variety was wholly unknown
until some three years ago. After an analysis
had been made in terms of unit-characters of
the color varieties of the mouse, it became clear
that if the color variation of guinea-pigs fol-
lowed a like course, a then unknown variety
of guinea-pig, cinnamon, should be capable of
production by crossing an agouti animal with
a brown one. In 1907 a statement of the sci-

84

EVOLUTION BY LOSS

entific expectation in the case was published,
and a few months later I had the satisfaction
of announcing its fulfillment in the second gen-
eration (F2) from the cross in question.

The experiment progressed as follows: The
parents were an agouti and a black, their Fx
offspring were agoutis in character ; but the F2
offspring were of four sorts, — agouti, black,
cinnamon, and brown. The cross thus produced
two varieties new to the experiment, viz. black
and cinnamon, the latter being a variety at
that time new among guinea-pigs.

The subsequent behavior, too, of the newly
produced cinnamon variety is in harmony
with expectation based on Mendelian prin-
ciples. The cinnamon variety has not pro-
duced agouti or black individuals, which from
the formulae it will be seen it may not be ex-
pected to produce, since it lacks the factor B.
But it has in some cases produced brown in-
dividuals, as it clearly could in case both par-
ents to a mating were heterozygous (single)
in factor A.

On the whole the evidence seems very clear
that the numerous color varieties of animals
7 85

HEREDITY

kept in captivity arise chiefly from loss or
modification of Mendelian unit-characters.
Loss of a unit-character might easily come
about by an irregular cell-division in which
the material basis of a character failed to
split, as normally. The consequence would be
that the character in question would be trans-
mitted by one only of the two cell-products
produced. The cell lacking a character might
be the starting-point of a race lacking the char-
acter, as of a black race, derived from a gray
one. On the other hand a modified condition
of a unit-character might possibly result from
unequal division of the material basis of a
character, so that one of the cell-products
would transmit the character in weakened in-
tensity, the other in increased intensity.

CASTLE, W. E.

1907. "Color Varieties of the Rabbit and of Other Rod-
ents: Their Origin and Inheritance." Science, N. S.,
26, pp. 287-291.

1908. "A New Color Variety of the Guinea-pig." Science,
N. S., 28, pp. 250-252.

1909. "Studies of Inheritance in Rabbits." Carnegie
Institution of Washington, Publication No. 114, 70 pp.,
4 pi.

86

CHAPTER VI

EVOLUTION OF NEW KACES BY VAKIATIONS IN THE
POTENCY OF CHAEACTEKS

IN the last chapter we discussed the color
variations of mammals, and we concluded
that these result largely from the loss or
modification of some half-dozen independent
Mendelian unit-characters. As to the material
basis of these unit-characters some interesting
evidence has recently been collected by Riddle.
Melanin pigment has been for some time known
to be formed by oxidation. A variety of or-
ganic compounds may undergo oxidation into
melanin pigments ranging in intensity from
light yellow to black; the greater the oxida-
tion, the darker the product. But it is not
certain, as assumed by Riddle, that the chemi-
cal method of oxidation is the same in all cases
or that the substance to be oxidized is the same.
The results obtained from breeding experiments

87

HEKEDITY

show that the capacity to form pigment of all
sorts may be lost by a single variation, which
we have called loss of the color factor, C. We
do not know whether it consists in the loss of
a substance capable of oxidation, or of the
power to take some indispensable first step in
the process of oxidation, perhaps due to loss
of an enzyme; but we do know that when this
particular variation has occurred, the power
to produce other than albino individuals can-
not be recovered by any known means except
a cross with colored animals. We know also
that the capacity to form specific kinds of pig-
ment (yellow, brown, or black) is independent
of the general color-factor, C, for albinos may
transmit those specific powers without them-
selves being able to form any kind of pigment
at all, i. e. without possessing C. Any animal
which forms pigment of one of the higher
grades has the capacity apparently to form
pigment also of the lower grades. Thus a
black animal can form also brown and yellow
pigment granules. Brown (chocolate) animals,
however, lack the capacity to form black pig-
ment. The oxidation, it would seem, can in

88

VAEIATIONS IN POTENCY

this case be carried no further than the brown
stage, because of the lack of some oxidizing
agency necessary to the last stage in pigment
production. The production of yellow is prob-
ably a first or early step in the oxidation
process preliminary to the production of brown
or black, yet all yellow animals, so far as
known, are able to take the further steps ; they
retain the capacity to form either brown or
black pigment to some extent, if only in the
eye.

The variations thus far described are what
De Vries has called retrogressive, i. e. due to
loss or modification. A much rarer sort of
variation has been called by De Vries progres-
sive, i. e. due to gain, acquisition of some char-
acter not before possessed by the race. I can
call to mind very few cases which certainly
fall in this category. One which it would seem
must belong here is the rough or resetted con-
dition of the hair in guinea-pigs, a variation
similar in nature to the reversed plumage of
birds, seen, for example, in the Jacobin pigeon.
The rough coat of guinea-pigs is surely not
an ancestral condition, yet it behaves as a

89

HEREDITY

dominant character in crosses. It can scarcely
be explained by loss; the only alternative is
to consider it an acquisition, unless we choose
to consider it a modification of the normal
condition.

Aside from the sorts of variations already
discussed, which consisted either in the loss or
modification of existing unit-characters or in
the gain of new ones, we must also recognize,
as a cause of permanent and heritable varia-
tion, changes in the potency of unit-characters,
i. e. their tendency to dominate in crosses.

When a gamete containing a particular unit-
character unites with a gamete not containing
it, the zygote formed will ordinarily show the
character in question fully developed. This re-
sult following Mendel's terminology we call
dominance. But dominance is frequently im-
perfect and may even be reversed. The zygote
in which a character is doubly represented fre-
quently develops the character more fully than
the zygote in which it is represented but once.
If a black guinea-pig is crossed with a yellow
one the offspring are black, but oftentimes of
a slightly yellowish shade. Likewise if black

90

VARIATIONS IN POTENCY

is crossed with brown, the crossbreds are apt
to develop in their coats more brown pigment
granules than do homozygous or pure blacks.
Nevertheless, we have no reason to question the
entire purity of the gametes, both dominant
arid recessive, formed by such cross-bred black
animals. It is the dominance, not the segrega-
tion, which is imperfect.

In other cases still the dominance may be
entirely reversed in character, owing to varia-
tion in the potency of a unit-character. Thus
in most rodents the gray or agouti pattern-
factor of the hair, A, is dominant. A cross of
black with homozygous gray, in rats, mice, or
rabbits, produces only gray offspring, which
in F2 produce three grays to one black. But
the so-called black rat, Mus rattus, a species
distinct from the one which has given rise to
the varieties kept in captivity, behaves in a
different way, as shown by Morgan ('09).
When crossed with its gray variety, the roof
rat, Mus alexandrinus , it produces only black
offspring, and in F2, three blacks to one gray.
If we suppose the gray coat in this case to
be due to the same factor as in other rodents,

91

HEREDITY

we must assign to it a different potency, or
power of dominance, so that it produces a
visible effect only when doubly represented in
the zygote.

In guinea-pigs, rabbits, and mice we have
seen that the presence together in the same
zygote of two factors, A and B, in any com-
bination whatever, produces the gray or agouti
coat. The two factors are A, the agouti or
gray marking of the hair, and B, black pig-
ment in the fur. If A is lacking, the coat is
black; if B is lacking, it is brown, cinnamon,
or yellow. If both are lacking, it is either
brown or yellow. But if both are present, the
wild or agouti type is produced. So far as
the production of the agouti coat is concerned,
it makes no difference whether either factor is
singly or doubly represented in the zygote.
Each factor has potency enough to produce the
full effect either in a single or in a double
dose. Accordingly, as we noticed in an earlier
chapter, we can distinguish by their breeding
capacity, though not by their looks, four types
of agouti guinea-pigs or gray rabbits, viz. :

92

VARIATIONS IN POTENCY

1. A A B B, which breeds true, since it forms game-

tes all AB;

2. ABB, which produces agouti young and black

ones in the ratio 3:1, since it forms
gametes A B and B ;

3. A A B, which produces agouti young and yellow

ones in the ratio 3:1, since it forms
gametes A B and A;

4. A B, which produces agouti, black, and yel-

low young in the ratio 9:3:4. For
the gametes formed by this sort are of
four kinds, A B, A, B, and neither
A nor B.

Now in rats we have no evidence that the
factor B has ever been lost, a matter to which
we shall presently return ; but the agouti factor
is apparently frequently wanting in ordinary
rats, which are then black. For ordinary rats,
then, the known combinations of A and B seem
to be three, viz. :

A A B B = the pure gray (wild type) ;

ABB = heterozygous gray, which produces off-
spring 3 gray : 1 black. This type is ob-
tained by crossing black with wild gray;

B B = pure black.

Now in Mus rattus, as we have seen, the
middle or heterozygous type is black, not gray

93

HEREDITY

in appearance, but it produces both the gray
and the black types. So the same gametic
formulas will account for both sets of facts,
if we suppose merely that the potency of A is
different in the two cases. In ordinary rats
(Mus norvegicus) A produces the gray coat in
a single dose; but in Mus rattus its potency
is less, two doses are required to produce the
gray coat. I am unable to frame any hypothe-
sis other than this which will account for the
reversal of dominance in one case as compared
with the other.

Yellow color in mammals affords another
illustration of this same thing, — reversal of
dominance. Black and brown are in most
mammals dominant over yellow in crosses, but
in mice the reverse is true. The differential
factor between black and yellow, if it is the
same in mice as in other rodents, must be in
one case potent enough to show itself if singly
represented in the zygote, whereas in the other
case it produces no visible effect unless doubly
represented in the zygote. Yellow certainly
seems to be a retrogressive variation from
gray, black, or brown. The pigment granules

94

VAKIATIONS IN POTENCY

remain in a lower oxidation stage in yellow than
in black or brown. We suppose that in the
yellow animal something is wanting which
makes that further oxidation possible. This
hypothesis would fully account for the observed
recessive nature of yellow in the case of all
mammals except mice. But here the capacity
to form black or brown pigment is regularly
present in the yellow individual but is held in
check. "We may suppose, therefore, that the
differential factor, that which converts yellow
into brown or black, must in this case be doubly
represented in the zygote in order to produce
brown or black fur, whereas in most mammals
a single dose is effective. Accordingly, if the
unmodified black or brown factor is represented
only once in the zygote, and the yellow modi-
fication is represented once, the latter will
show, since the former is singly ineffective.
The animal accordingly is a heterozygous yel-
low, capable of producing also black or brown
offspring. But mice are peculiar in that they
cannot exist in the doubly deficient condition
of a pure yellow zygote, consequently all
yellow mice are heterozygous dominants,

95

HEREDITY

whereas other yellow mammals are homozy-
gous recessives.

In connection with this same case may pos-
sibly be found the explanation of the complete
absence of the yellow variation in rats. In
nearly all mammals kept in captivity yellow as
well as black varieties occur; this is true of
horses, cattle, swine, dogs, cats, rabbits, guinea-
pigs, and mice. In rats, however, a yellow
variety is unknown. We know that rats are
able to form yellow pigment, for all wild rats
do form yellow pigment in their agouti fur, yet
singularly enough no all-yellow rat has ever
been observed, so far as we have any record,
either wild or in captivity. A rat of this sort
would command a high price at the hands of
any fancier. Suppose the variation did occur
in a single gamete. If, as in most mammals,
it behaved as a recessive in crosses, it would
not become visible, and might be carried along
for untold generations without ever becoming
visible unless two yellow gametes met. But if,
as in mice, the yellow-yellow combination when
formed quickly perished, then the character
might never become visible. So the yellow

96

variation may have occurred many times in
rats, as it has in so many other mammals, but
failed to become visible simply because it has
the same potency as in most mammals, but is
subject to the same physiological limitations
as in mice, so that it cannot exist in a homo-
zygous state. In that case the only evidence
of its existence in a race would lie in a
slightly diminished fecundity under inbreeding,
as is found to be the case in yellow mice.

Such sharply contrasted variations in the
potency of characters as we have been discuss-
ing are evidently of prime importance in evo-
lution, making all the difference between a
dominant and a recessive condition of a char-
acter, or between the occurrence and the per-
manent suppression of a particular variation.
The character which is potent enough to show
itself in a single dose will behave as a domi-
nant character in crosses. We might call it
nnipotent. That which must be present in a
double dose to produce a visible result will
behave as a recessive character in crosses. We
might call it semi-potent. It is not impossible
that the same character may as regards domi-

97

HEREDITY

nance behave in different ways under different
circumstances, at one time dominating com-
pletely, at another only feebly, and at other
times not at all.

Undoubtedly the chief condition affecting
dominance is the nature of the gamete with
which a union is made in fertilization. In 1905
(Carnegie Inst. PubL No. 23) I described a
case in which a particular guinea-pig (male
2002, shown in Fig. 32) having a rough or
resetted coat gave a varying result in crosses.
In crosses with most smooth animals his rough
character dominated completely (see Fig. 24,
which shows a son of the male 2002 by a smooth
mother), but with one particular smooth ani-
mal the dominance was very imperfect in all
the young (Fig. 36), while with a second it
was imperfect in half the young. The conclu-
sion was drawn that gametes vary in potency,
and that parents, too, differ as regards the
potency of the gametes which they produce,
some individuals producing gametes all of which
are relatively potent, others producing gametes
only half of which are potent, while still others
produce gametes none of which are potent.

98

VARIATIONS IN POTENCY

Relative potency would, therefore, seem to be
a character inherited in Mendelian fashion.1

Observations of Coutagne on silk-moths may
be cited in support of this idea. Coutagne
made crosses between races of silk-moths dif-
fering in cocoon color, viz. between a race which
spun yellow cocoons and another one which
spun white cocoons. He found that some of
the Fj offspring spun yellow cocoons, others
white ones. The Fj yellow cocooned animals
when bred together produced F2 progeny which
spun some yellow, others white cocoons, the
two sorts being as 3:1. In other words, yel-
low in such cases behaved consistently as a
dominant character. And the white-cocooned
Fj moths produced in F2 cocoons of both colors,
but in this case the white cocoons were to the
yellow ones as 3:1. In other words, when
yellow behaved as a dominant in Fj it behaved
as a dominant also in F2; and the same was
true of white. Each retained throughout the
two generations the relative potency with which

1 It is of course possible to interpret such a case as due to the
separate inheritance of a factor which inhibits the development
of the character, but it is doubtful whether this line of explana-
tion can be successfully applied to cases presently to be described.

99

HEREDITY

it started. C. B. Davenport has also produced
much evidence favoring the idea of varying
potency of characters in recent papers based
on his extensive studies on poultry.

The case which I described in 1905 was one
in which unusual potency seemed to inhere in
the gametes of a recessive individual, — one
which apparently did not possess the character
whose dominance was affected. But there occur
also cases in which the varying gametic potency
is associated directly with the character af-
fected. One such I was able to describe in
1906, — that of an extra toe in guinea-pigs. It
was found while building up a polydactylous
race by selection and crossing it with other
races that individuals varied in the potency
which the character had in their gametes. In
general the better developed the character was
in an individual the more strongly was it trans-
mitted, i. e. the larger was the proportion of
polydactylous individuals produced in crosses.
In no case, however, was this a recognizable
Mendelian proportion, though both dominance
and segregation seemed to be taking place.
Variation in potency was, however, unmistak-

100

FIG. 36. — An imperfectly rough guinea-pig. Produced by mating
the guinea-pig, shown in Fig. 32, with a particular smooth animal;
female, 2005.

FIG. 37. — A silvered guinea-pig. One in whose coat occur white
hairs interspersed with pigmented ones. The amount of the silver-
ing has been greatly increased by selection.

FIG. 38. — A. Front feet of an ordinary guinea-pig. B. Its hind feet.
D. Hind feet of a race four-toed on all the feet. C. Ordinary con-
dition of the hind feet of young obtained by crossing B with D.

FIG. 39. — Diagram showing variation in the color-pattern of hooded
rats.

VARIATIONS IX POTENCY

able and was transmitted from generation to
generation.1 See Fig. 36.

It is an important question whether potency
is a property of the unit-character or of the
gamete, i. e. whether it affects all the charac-
ters transmitted by a gamete or only a par-
ticular one. Practical breeders as a rule favor
the idea of gametic rather than of unit-character
potency, but this is probably due to a failure
to discriminate between the two. They desig-
nate as " prepotent " an individual supposed
to impress all its characters upon the offspring,
but it is very doubtful whether such individuals
exist. It is easy to mistake for an animal
potent in all respects one which is potent in
one or two important respects only, especially
if the observer is unaware, as every one has
been until quite recently, that one character is
independent of another in transmission.

Conditions other than the character of the
gametes themselves may determine the extent

1 An alternative explanation is possible, viz. that the develop-
ment of the fourth toe depends upon the inheritance of several
independent factors, and that the more of these there are present,
the better will the structure be developed. The correctness of
such an interpretation must be tested by further investigations.

8 101

HEREDITY

to which a character develops in the zygote,
i. e. the completeness or incompleteness of its
dominance in a particular case. For example,
in salamanders, which apparently, like mam-
mals, form skin-pigments of different sorts,
such as yellow, brown, and black, Tornier has
found that by feeding one may control the
proportions in which chromatophores of the
several sorts are formed in the skin. Abundant
feeding causes preponderance of pigment of
one sort, scanty feeding causes preponderance
of pigment of another sort. Here external
conditions determine the degree of development
of characters. In other cases internal condi-
tions may exercise a controlling influence. Thus
in cattle the capacity to develop horns is a
semi-potent unit-character, behaving as a re-
cessive in crosses, heterozygotes developing
only " scurs," that is, mere thickenings of the
skin, or else no trace of horns at all. In sheep,
moreover, horns are more strongly developed
in males than in females, the presence of the
male sex-gland in the body, or rather probably
some substance given off into the blood from
the sex-gland, favoring growth of the horns.

102

In merino sheep the male has well-developed
horns but the female is hornless ; yet if the male
is castrated early in life no horns are formed.

When a breed of sheep horned in both sexes,
such as the Dorset, is crossed with one horn-
less in both sexes, such as the Shropshire, horns
are borne by the male but not by the female
offspring. Both sexes, however, are heterozy-
gous in horns, as is shown by their breeding
capacity. For in F2 occur both horned and
hornless individuals in both sexes. The horn-
less males and the horned females proye to
be homozygous, but the horned males and the
hornless females may be either heterozygous
or homozygous. Accordingly the character,
horns, behaves consistently as a dominant char-
acter in one sex, but as a recessive in the
other. Further, the presence of the male sex-
gland in the heterozygote raises the potency of
the character, horns, from semi-potent to uni-
potent, as the result of castration shows.

It is impossible to be certain that in a horn-
less race the character horns has been wholly
lost. It may merely have fallen so low in
potency that under ordinary conditions it pro-

103

HEEEDITY

duces no visible structures. The occasional
occurrence of an imperfectly horned animal as
a sport within a hornless race need not, then,
occasion surprise. It would be a variation of
the same sort as the extra toe in guinea-pigs
(see Fig. 38), which, from a single sport, was
built up by selection into a well-established race
within a very few generations. This character,
seemingly lost from the germ-plasm for an in-
definite period, had perhaps merely fallen so low
in potency that it no longer produced the fourth
toe on the hind foot, though this was still pres-
ent on the front foot. In the variant observed,
the first polydactylus guinea-pig of my stock,
the toe was imperfectly developed on one hind
foot, doubtless as the result of an unusually
potent condition of the character in one of the
gametes which produced the individual. This
manifestation of the character, though feeble,
was sufficient to afford a guide for selection of
those individuals which formed the most potent
gametes, and so a polydactylous race was
formed by selection and inbreeding.

Great as has been the contribution of Men-
delian principles to our knowledge of heredity,

104

VARIATIONS IN POTENCY

they do not reduce the whole art of breeding
to the production of new combinations of unit
characters through crossing. Selection is re-
quired also, not merely among different combi-
nations of unit-characters, but also among in-
dividuals representing the same combinations
selection is required of those possessing the
desired characters in greatest potency. The
further rdle of selection in evolution we shall
need to consider in a subsequent chapter.

BIBLIOGEAPHY

CASTLE, W. E., and LITTLE, C. C.

1909. "The Peculiar Inheritance of Pink Eyes Among

Colored Mice." Science, N. S., 30, pp. 312-314.
COUTAGNE, G.

1902. "Recherches expe'rimentales sur 1'he're'dite' chez les

vers-a-soie." Bull. Sci., 37, pp. 1-194, 9 pi.
MORGAN, T. H.

1909. "Breeding Experiments with Rats." American

Naturalist, 43, pp. 182-185.
DE VRIES, H.

1901-03. "Die Mutationstheorie." Leipzig, Veit and Co.

See also the Bibliographies to Chapters III., IV., and V.

CHAPTER VII

CAN MENDELIAN TJNIT-CHAEACTERS BE MODIFIED
BY SELECTION?

IF, as suggested in the last chapter, the po-
tency of a character in crosses may be
modified by selection, why may not the
character itself be modified by selection, or are
not the two things perhaps identical, viz. modi-
fication of the potency of a character and modi-
fication of the character itself? Darwin firmly
believed that the characters of organisms can
be modified by selection, and he made this the
foundation stone of his theory of evolution.
De Vries and Johannsen, however, have taught
us a different doctrine, maintaining that selec-
tion is able to affect characters in superficial
and transitory ways only, that the slight varia-
tions in characters which we see everywhere
among organisms have no evolutionary signifi-
cance or permanent value ; that they come and

106

MENDELISM AND SELECTION

go like the wavelets on the ocean beach, but
have no more relation to evolution than the
waves have to the tides. The brilliancy of the
Mutation theory of De Vries, coupled with his
great service to biology in rediscovering the
Mendelian laws, has somewhat dazzled our eyes
and led us, I think, to accept too readily his
views concerning the efficacy of selection also.
Ten years' continuous work in selection con-
vinces me that much can be accomplished by
this means quite apart from the process of mu-
tation. The work of De Vries himself argues
strongly in favor of this idea. To be sure, his
interpretation of it is adverse to selection, and
has seemed to most of us at times overwhelm-
ingly convincing; but from his interpretation
we may fairly appeal to the record of the work
itself, and with this compare the record of our
own work.

One of the most extensive selection experi-
ments conducted by De Vries was made on the
common buttercup, Ranunculus bulbosus, which
occurs as a weed in pastures and meadows in
this country as well as in Europe. It has, as
is known, regular 5-petaled flowers. An ex-

107

HEREDITY

animation of 717 flowers in the field made by
De Vries in 1887 showed the rather frequent
occurrence of 6 and 7 petaled flowers also, the
average number of petals in the entire collec-
tion being 5.13. De Vries set himself the task
to see if the proportion of many petaled flowers
could be increased or the number of petals to
a flower be further increased. In both these
respects he succeeded surprisingly well. As a
result of five successive selections the average
number of petals was raised from 5.6 to 8.6,
the upper limit of variation from 8 to 31, and
the mode (or commonest condition) from 5 to 9.
Singularly enough De Vries concludes, in ac-
cordance with general ideas which he had
adopted, that selection had in this case done
practically all that it could accomplish, that
further selection, while it might advance the
average somewhat farther, would have no per-
manent effect in modifying the type. This be-
lief seems to have rested on considerations
such as these. De Vries had found, as had
others, that variations which are heritable have
their origin in the germ-cells only. He recog-
nized that the tendency to produce double

108

MENDELISM AXD SELECTION

flowers in the buttercup is a heritable varia-
tion and supposed it to be a unit-character,
and so to conform with Mendel's law.

Now, if the tendency to produce double
flowers were a simple Mendelian character it
could exist in only three conditions, — that of
a recessive, that of a homozygous dominant,
or that of a heterozygous dominant. But re-
cessives and homozygous dominants are pure,
that is, they form only one type of gamete,
and selection therefore from among their pro-
geny could produce no new type. As regards
the heterozygous dominant type, this would
itself be unfixable, and selection could accom-
plish nothing permanent except by isolating a
homozygous type. But such types should all
be in evidence within two generations; there-
fore, if a completely and permanently double
type had not been discovered within the five
generations covered by the experiment, such a
type was not to be expected at all from the
material in hand, unless either a wholly new
unit-character were introduced or an existing
one were profoundly modified. De Vries con-
siders changes of both these sorts possible,

109

HEREDITY

He calls them mutations, and regards them as
the sole means of evolutionary progress. But
it is a peculiarity of his mutation theory that
it regards only large changes in unit-characters
as having any permanency, namely, such
changes as mean a practical making over of
the character. To borrow a figure from Bate-
son, just as the gas carbon monoxide, C 0,
may change into a very different gas, — carbon
dioxide, C O2, — by taking up a single atom
of oxygen, but can make no less extensive
change, since oxygen atoms do not split; so,
according to De Vries, a unit-character may
not change unless it changes profoundly. Vari-
ous circumstances may modify the degree of
its expression, but these are without perma-
nent effect, since the character itself remains
unchanged.

But there are both a priori and experimental
grounds for questioning the correctness of
De Vries' conclusions. It is known that the
chemical compounds within the germ-cells are
not so simple in composition as C 0 and C 02.
They are very complex substances, made up,
it is thought, of very many atoms, often hun-

110

MENDELISM AND SELECTION

dreds in a single molecule. If so, it is quite
possible that an atom or two might be trans-
posed in position within the molecule without
wholly altering its chemical nature, and that
thus slight changes in the germ-plasm might
result, which, however, would be as permanent
as more profound changes.

The argument of De Vries against any per-
manent effect of selection in modifying unit-
characters has been greatly strengthened by
the subsequent work of Johannsen and Jennings.
Johannsen has found that if one selects from a
handful of ordinary beans the largest seeds and
the smallest seeds, and plants these separately,
the former will produce beans of larger average
size than the latter. Selection here has effect.

But if the selection is made, not from a
general field crop of beans, but from those
beans borne on one and the same homozygous
mother plant, then the progeny of the selected
large seed will be no larger than that of the
selected small seed. Selection here is without
effect.

The different result in the two cases may
be explained, according to Johannsen, on

111

105

FIG. 40. — Diagram showing the variations in size of eight different races of
paramecium. Each horizontal row represents a race derived from a
single parent individual. The individual showing the mean size in
each race is indicated by a cross placed above it. The mean of the
entire lot is shown at X — X. The numbers show the measure-
ments in microns. (After Jennings.)

11?

the principle of the " pure line." The pro-
geny of a single self-fertilized homozygous
bean plant constitute a pure line. They are
all alike, so far as the hereditary transmission
of size is concerned, for they are all derived
from like gametes. The differences in size
which occur among them are due to differences
in nutrition, not to germinal differences, and
they are not transmitted. But in a mixed
population of beans, such as is represented by
a field crop, differences of size occur which
are due to heredity as well as those which are
due to the environment. In the case of the
former, selection naturally has effect; in the
case of the latter, it does not.

Jennings has obtained similar results in his
studies of paramecium, — a one-celled animal
which multiplies asexually by dividing into two
similar parts. It lives in stagnant water and
may be reared in great numbers in a hay-infu-
sion, for it multiplies with great rapidity, divid-
ing two or three times within twenty-four hours.
The variations in size which occur in parame-
cium are shown in Fig. 40.

When from an ordinary culture of parame-
113

HEREDITY

cium Jennings selects the largest and the small-
est individuals respectively, he finds that the
descendants of the one lot will be of larger size
than the other. This looks like an effect of selec-
tion upon racial size. But if selection is made not
within a mixed population but among the de-
scendants of a single individual, it is found
that the descendants of large individuals are
of no greater average size than those of small
individuals.

The explanation of this fact is to be
found in the existence of what Johannsen
has called pure lines. Jennings has been able
to isolate eight distinct pure lines of parame-
cium differing in average size, as shown in
Fig. 40. The range of variation in size within
one of these races is great, but if one selects
extremely large or extremely small individuals
within the same pure line, i. e. among the
asexually produced descendants of the same
animal, no change in the average size of the
race is brought about.

A very different result is obtained, however,
if one mixes together several pure lines and then
selects from the mixed race on the basis of size.

114

MENDELISM AND SELECTION

The larger animals then produce larger average
offspring and vice versa. An examination of
Fig. 40 will show why. Animals of the same ab-
solute size are there placed in the same vertical
row. If, now, one selects from the mixed popu-
lation only the largest individuals, he will nat-
urally secure representatives of only two or
three pure lines, viz. of those lines which are
characterized by the largest average size, and
which, therefore, will produce large average
offspring. If on the other hand he selects
extremely small individuals, he will secure
representatives of only the smallest races,
which naturally will produce small offspring,
so that selection seems to be effective in modi-
fying racial size, but in reality it does this by
sorting out the elementary constituents of the
race.

It is impossible to deny the soundness of
the reasoning of Johannsen and Jennings. It
is perfectly clear that the effects of selection
should be more immediate and much greater
in the case of a mixed race than in that of a
pure line, but is it certain, as assumed by
them, that selection is wholly without effect in

115

HEREDITY

the case of a pure line? We know the effects
should be less, but are they nil? Concerning
this matter we are perhaps justified in await-
ing further evidence. For in the case of beans
and of paramecium alike size is subject to very
great variation through the influence of nutri-
tion. Variations due to this cause are natu-
rally not inherited, since the germ-cells are not
affected by them, but only the body. But is
it not possible that along with the striking size
differences due to nutrition there may occur
also slight size differences due to germinal
variation within the pure line, that is owing
to variations in the potency of the same unit-
character or combination of unit-characters?
To be sure, Johannsen and Jennings have not
observed these, but this does not prove their
non-existence. Others may yet be able to do
so; indeed one case is already on record in
which such observations have been made in the
case of a small crustacean (or water-flea),
Daphnia.

Daphnia is a small transparent animal, about
the size of a pin-head, which occurs in enor-
mous numbers in fresh-water lakes and pools,

116

MENDELISM AND SELECTION"

forming a large part of the food supply of
fresh-water fishes. It multiplies chiefly by the
production of unfertilized eggs, — those which
undergo no reduction and which develop with-
out fertilization into an individual like the
parent. The germinal composition, therefore,
of all descendants produced in this way by the
same mother should be identical, unless germi-
nal composition can be modified in other ways
than by reduction and recombination of unit-
characters. Now the German zoologist, Wol-
tereck, has shown that, among the offspring
developed from the unfertilized eggs of the
same mother Daphnia, variations do occur
which are heritable, so that if one selects ex-
treme variants he obtains a modified race.
Systematic zoologists recognize as a generic
distinction between Daphnia and Hyalodaphnia
absence from the latter of the rudimentary eye
found in Daphnia. Woltereck observed that
in a pure line of Hyalodaphnia the rudi-
mentary eye, usually wanting, may occur in
individual cases. He found further that it
occurred in varying degrees of development,
which ranged all the way from a group of
9 117

HEEEDITY

pigmented cells outside the brain, through
stages in which cells were present without pig-
ment, and others in which pigment was visible
within the brain but no cells outside it were
developed, and finally to those in which all
traces of the eye had vanished, cells and pig-
ment alike. By selection in three successive
generations of the mother having the rudi-
mentary eye best developed offspring were
obtained, 90 % of which had the pigmented
eye, and which would therefore pass for ani-
mals of a wholly different genus. The degree
of development of the organ in the last genera-
tion was also greater than in the previous
generations. Here within a pure line produced
by parthenogenesis selection served to augment
both the degree of development of an organ
and the frequency of its occurrence within the
race, a result precisely parallel to that which
I obtained some years ago by selection in the
case of a rudimentary fourth toe in the guinea-
pig. The experiment with Daphnia is not open
to the objection that may be offered to the
guinea-pig experiment, that it is possibly a
result of gametic segregation and recombina-

118

MENDELISM AND SELECTIOX

tion, for in Daphnia the reproduction was ex-
clusively by unreduced and unfertilized eggs.

The rudimentary eye of Daphnia is an organ
the development of which, so far as observed,
is wholly independent of environmental influ-
ence; but the case is different with another
structure of Daphnia, upon which also Wol-
tereck made observations, namely, a projection
or spine borne on the head of the animal.
This is not a constant structure, but is some-
times present, sometimes wanting altogether, in
the same pure line. In extreme cases it forms
a great angular extension of the head forward.
To a considerable extent its development is
subject to control through the temperature of
the surrounding water, but independently of
such influence the degree of its development
varies and is heritable. Although in general,
just as in the experiments of Johannsen and
Jennings, selection of animals with the best-
developed spine did not increase the degree of
development of the organ or the frequency of
its occurrence, yet in individual cases such
increase was observed, so that the structure
occurred in over 50 % of the offspring. In

119

HEEEDITY

such cases, then, it would seem that along with
the cases due to environmental influence oc-
curred others due to germinal variation. Al-
though selection of the former would not in-
fluence the race permanently, there is every
reason to think that the latter would so influ-
ence it, and did in the experiment.

Accordingly the results of Johannsen and
Jennings on the one hand, and of Woltereck
on the other, are not necessarily in opposition
to each other. Woltereck 's conclusions agree
with those of Johannsen and Jennings so far
as concerns the great bulk of the variations,
those caused by external influences. All agree
that they are not inherited. Woltereck, how-
ever, observes also, what the others have failed
to observe, that along with the non-inherited
variations occur other similar but less numer-
ous ones which are inherited.

My own observations are entirely in har-
mony with those of Woltereck. Like him, I
find that selection may modify characters. In
several cases I have observed characters at
first feebly manifested gradually improve under
selection until they became established racial

120

MENDELISM AND SELECTION

traits. Thus the extra toe of polydactylous
guinea-pigs made its appearance as a poorly
developed fourth toe on the left foot only.
Only 6 % of the offspring of this animal by
normal unrelated mothers were polydactylous,
but among his offspring were some with better
developed fourth toes than the father pos-
sessed. Such individuals were selected through-
out five successive generations, at the end of
which time a good four-toed race had been
established. It was found in general that those
animals which had best-developed fourth toes
transmitted the character most strongly in
crosses with unrelated normal animals. The
percentage of polydactylous individuals pro-
duced in such crosses varied all the way from
0 to 100 %. By selection this percentage was
increased, as was also the degree of develop-
ment of the fourth toe in crosses.

Another character which made its appear-
ance among our guinea-pigs, at first feebly
expressed, was a silvering of the colored fur,
due to interspersing of white hairs with the
colored ones (see Fig. 37). The first indi-
viduals observed to have this character bore

121

HEREDITY

white hairs on the under surface of the body
only. By inbreeding, a homozygous strain of
the silvered animals was soon obtained, one in

43

— 1

gen.

_.--":

6

FIG. 41. — Chart showing effects of selection in eight successive
generations upon the color-pattern of hooded rats. A,
average condition of the selected parents in the plus series;
B, average condition of their offspring. A1, average condi-
tion of the selected parents in the minus series; B1, average
condition of their offspring.

which all the offspring were silvered to a
greater or less extent. Selection was now
directed toward two ends, — (1) to secure ani-
mals which were free from spots of red or
white, a condition which was present in the

122

MENDELISM AND SELECTION"

original stock, and (2) to secure extensive and
uniform silvering on a black background. In
both these objects good progress has been
made. We have animals which are silvered all
over the body except on a part of the head,
and the percentage of such well-silvered indi-
viduals is relatively high.

But the most extensive selection experiment
which I have personally observed is one in
which I have been assisted by Dr. John C.
Phillips (see Figs. 39 and 41). Selection in
this case has been directed toward a modifica-
tion of the color pattern of hooded rats, — a
pattern which is known to behave as a reces-
sive Mendelian character in crosses with either
the self (totally pigmented) condition or the
so-called Irish (white-bellied) condition found
in some other rats. The extreme range of
variation among our hooded rats at the outset
of this experiment is indicated by the grades
- 2 and + 3 of Fig. 39. Selection was now
made of the extreme variates in either direc-
tion and these were bred separately. Two
series of animals were thus established, — one
of narrow striped animals, minus series; the

123

HEKEDITY

other of wide striped, plus series. In eaclr
generation the most extreme individuals were
selected as parents ; in the narrow series, those
with narrowest stripe ; in the wide series, those
with widest stripe.

TABLE I

Results of Selection for Modification of the Color-pattern
of Hooded Rats.

Plus series.

Minus series.

GENERA-
TION.

1
2
3
4
5
6
7
8

AVERAGE

GRADE,

PARENTS.

2.50
2.51
2.73
3.09
3.33
3.51
3.53
3.65

AVERAGE NUMBER
GRADE, OF OFF-
OFFSPRING. SPRING.

2.05
1.92
2.51
2.72
2.90
3.09
3.14
3.30

150

471

341

444

610

834

874

91

3,815

1.46

1.00

55

1.41

1.07

132

1.56

1.18

195

1.69

1.28

329

1.73

1.41

701

1.86

1.56

1252

2.00

1.70

1544

2.03

1.78

713

124

4,921

MENDELISM' AND SELECTION

The result of the selection is shown graph-
ically in Fig. 41 (compare Table I). The
offspring in the narrow series became with
each generation narrower; those in the wide
series became with each generation wider, with
a single exception. In generation two the wide
stock was enlarged by the addition of a new
strain of animals. This caused a temporary
falling off in the average grade of the young,
the two series overlapping for that generation.
No new stock was at any other time intro-
duced in either series, the two remaining dis-
tinct at all times except in generation two.
It will be observed that a change in the aver-
age grade of the parents is attended by a
corresponding change in the average grade
of the offspring. The amount of variabil-
ity of the offspring is not materially • affected
by the selection, but the average about which
variation occurs is steadily changed, as are
also the limits of the range of variation.

The interesting feature of this experiment
is the production, as a result of selection, of
wholly new grades ; in the narrow series, of ani-
mals having less pigment than any known type

125

HEREDITY

other than the albino; in the wide series, of
animals so extensively pigmented that they
would readily pass for the ' ' Irish type, ' ' which
has white on the belly only, but which is known
to be in crosses a Mendelian alternative to the
hooded type. By selection we have practically
obliterated the gap which originally separated
these types, though selected animals still give
regression toward the respective types from
which they came. But this regression grows
less with each successive selection and ulti-
mately should vanish, if the story told by these
statistics is to be trusted. As yet there is no
indication that a limit to the effects of selec-
tion has been reached.

From the evidence in hand we conclude that
Darwin was right in assigning great importance
to selection in evolution; that progress results
not merely from sorting out particular com-
binations of large and striking unit-characters,
but also from the selection of slight differences
in the potentiality of gametes representing the
same unit-character combinations. It is pos-
sible to ascribe such differences to little units
additional to the recognized larger ones, but

126

MENDELISM AND SELECTION

if such little units exist, they are indeed very
little as well as numerous, and by adding to the
effect of the larger ones they produce what
amounts to modification of them.

BIBLIOGEAPHY

CASTLE, W. E.

1906. "The Origin of a Polydactylous Race of Guinea-
pigs." Carnegie Institution of Washington, Publication
No. 49, pp. 17-29.
JENNINGS, H. S.

1909. " Heredity and Variation in the Simplest Organ-
isms." The American Naturalist, 43, pp. 321-337.

JOHANNSEN, W.

1909. "Elemente der exakten Erblichkeitslehre." G.

Fisher, Jena, 516 pp.
DE VRIES, H.

(See Bibliography to Chapter VI.)

WOLTERECK, R.

1909. " Weitere experimentelle Untersuchungen iiber Art-
veranderungen, speciell iiber das Wesen quantitativer
Artunterschiede bei Daphniden." Verh. Deutsch. Zool.
Gesellsch., pp. 110-172.

CHAPTER VIII

MENDELIAN INHERITANCE WITHOUT DOMINANCE,
" BLENDING " INHERITANCE

WE shall now discuss a seemingly
different type of inheritance from
that discovered by Mendel, — one
in which the offspring are a true intermediate
or blend between the parents, and in which the
occurrence of segregation has not in all cases
been certainly established.

Differences in size between parents have been
found to behave in this blending fashion. Rabbits
are apparently favorable material in which to
study size inheritance, for some races are fully
twice as large as others. If a large rabbit is
crossed with a small one the young are of inter-
mediate size, and the F2 offspring show no such
segregation into large, small, and intermediate
individuals as a simple Mendelian system would
demand. For this reason size has been de-

FIG. 42. — Skulls of three rabbits. Father (1 and la),
mother (3 and 3a), and son (2 and 2a).

" BLENDING " INHEEITANCE

scribed as a non-Mendelian, non-segregating
type of inheritance, but recent discoveries place
this interpretation in doubt. Let us first con-
sider what are the observed facts and after-
ward the interpretation.

Fig. 42 shows the skulls of three rabbits, -
of the father at the left, of the mother at the
right, and that of the son between. Notice the
fully intermediate or blended character of the
son's skull as regards both absolute dimensions
and proportions. The intermediate character
was possessed also by the next generation of
offspring. Now this same cross, while pro-
ducing a blend in size and ear-length, was
yielding dominance and segregation of coat-
characters. Fig. 43 shows a picture of the
rabbit with the small skull in the cross just
described. He was an albino and his fur was
long. The mother, which had the large skull,
was a sooty-yellow rabbit, with short fur and
long ears (see Fig. 44). The son is shown
in Fig. 45. His fur was black and short, the
albinism and long fur of his father having
become recessive in the cross in accordance
with Mendel's law. The pigmentation is also

129

HEREDITY

intensified in the son, black having been re-
ceived through the albino parent as a latent
factor, which became fully active in the son.
The excluded albinism, recessive in the son and
his brothers and sisters, all seven of which
were similar in character, reappeared among
the grandchildren, as, for example, in the one
shown in Fig. 46, which was short-haired.
Other F2 offspring were long-haired, some of
them being albinos, others being pigmented.
But the size and ear-length of the son were
intermediate between the sizes and ear-lengths
of his parents, and this intermediate character
persisted without apparent segregation among
the F2 offspring. The animals in the pictures
are unfortunately not all shown on the same
scale, but the relative ear-lengths are sufficiently
clear.

A Mendelian interpretation of blending in-
heritance, illustrated in the inheritance of skull-
size and ear-length among rabbits, has been sug-
gested by my colleague Dr. East, and by others,
an interpretation in which Mendelian dominance
is indeed wanting but segregation nevertheless
occurs, yet not of a simple kind, involving one

130

"BLENDING" INHERITANCE

or two segregating factors, but involving sev-
eral such factors. Before entering into this ex-
planation it will be necessary to discuss a fur-
ther extension of Mendelian principles recently
made.

Some modified Mendelian ratios of particular
interest have lately been obtained by the Swedish
plant-breeder, Nilsson-Ehle (1909, Lunds Uni-
versitets Arsskrift) in crossing varieties of
wheat of different color. When a variety hav-
ing brown chaff is crossed with one which has
white chaff, the hybrid plants are regularly
brown in F± and 3 brown : 1 white in F2, but
a particular variety of brown-chaffed wheat
gave a different result. In 15 different crosses
it gave uniformly a close approximation to
the ratio 15 : 1 instead of 3 : 1. The totals are
sufficiently large to leave no doubt of this.
They are 1410 brown to 94 white, exactly 15 : 1.
This is clearly a dihybrid Mendelian ratio, and
Nilsson-Ehle interprets it to mean that there
exist in this case two independent factors, each
of which is able by itself to produce the brown
coloration, though no qualitative difference can
be detected between them.

131

A still more remarkable case was observed
in crosses between varieties of wheat of dif-
ferent grain-color. Red crossed with white
gave ordinarily all red in Fa and 3 red to 1
white in F2, but a certain native Swedish sort
gave only red (several hundred seeds) in F2.
This result was so surprising that one cross
which had yielded 78 grains of wheat in F2 was
followed into F3, with the following result:

50 plants gave only red seed; expected 37

5 " " approximately 63 R: 1 W; " 8

15 " 15R:1W; " 12

8 "• " 3R:1W; " 6

0 " all white; " 1

The interpretation given by Nilsson-Ehle is
this. The red variety used in this cross bears
three independent factors, each of which by
itself is able to produce the red character.
Their joint action is not different in kind from
their action separately, though possibly quanti-
tatively greater. The F2 generation should
contain 1 white seed in 64. It happens that
none were obtained in this generation. The
next generation should contain in a total of
64 individuals, the sorts actually observed as

132

£- •>•-*

•*-*-> I

£ ,ja 5 3

" BLENDING " INHERITANCE

well as a sort which would produce only white
seed, the progeny namely of the expected
white seed of F2, but as that was not obtained,
the all-white plant of F3 could not be obtained
either. The expected proportions of the sev-
eral classes in F3 are given for comparison with
those actually obtained. The agreement be-
tween expected and observed is so good as to
make it seem highly probable that Nilsson-
Ehle's explanation is correct. Corroborative
evidence in the case of maize has been obtained
by Dr. E. M. East (Am. Naturalist, Feb., 1910).
This work introduces us to a new principle
which may have important theoretical conse-
quences. If a character ordinarily represented
by a single unit in the germ-plasm may become
represented by two or more such units identi-
cal in character, then we may expect it to domi-
nate more persistently in crosses, fewer reces-
sives being formed in F2 and subsequent gen-
erations. Further, if duplication of a unit tends
to increase its intensity, as seems probable,
then we have in this process a possible expla-
nation of quantitative variation in characters
which are non-Mendelian, or at any rate do not
10 133

HEREDITY

conform with a simple Mendelian system. Con-
sider, for example, the matter of size and
skeletal proportions in rabbits. It is perfectly
clear from the experiments described that in
such cases no dominance occurs, and also that
no segregation of a simple Mendelian character
takes place, but it is not certain that the ob-
served facts may not be explained by the com-
bined action of several similar but independent
factors, the new principle which Nilsson-Ehle
has brought to our attention. Let us apply
such a hypothesis to the case in hand.

Suppose a cross be made involving ear-
lengths of approximately 4 and 8 inches respec-
tively, as in one of the crosses made. The Ft
young are found to have ears about 6 inches
long, the mean of the parental conditions, and
the F2 young vary about the same mean con-
dition. If a single Mendelian unit-character
made the difference between a 4 inch and an 8
inch ear, the F2 young should be of three
classes as follows:

Classes 4 in. 6 in. 8 in.

Frequencies 121

134

" BLENDING » INHERITANCE

(Compare Fig. 47, bottom left.) The grand-
parental conditions should in this case reappear

.1

I.

.ill.

i

i

FIG. 47. — Diagrams to show the number and size of the classes
of individuals to be expected from a cross involving Mende-
lian segregation without dominance. One Mendelian unit
involved, bottom left; two units, middle left; three units,
top left; four units, right.

in half the young. This clearly does not oc-
cur in the rabbit experiment. But if two unit-

135

HEREDITY

characters were involved, Fj would be un-
changed, all 6 inches, yet the F2 classes would
be more numerous, viz., 4, 5, 6, 7, and 8 inches,
and their relative frequencies as shown by the
height of the columns in Fig. 47, middle left, 1, 4,
6, 4, 1. The grandparental states would now re-
appear in i/s °f the F2 young, while % would be
intermediate. It is certain, however, that in rab-
bits the grandparental conditions, if they re-
appear at all, do not reappear with any such
frequency as this.

If three independent size-factors were in-
volved in the cross, the F1 individuals should all
fall in the same middle group, as before, viz.
6 inches, but the F2 classes should number
seven, and their relative frequencies would
be as shown in Fig. 47, top left. For 4 independ-
ent size-factors, the F2 classes would be more
numerous still, viz., 9 (Fig. 47, right), and the
extreme ear-size of either grandparent would
be expected to reappear in only one out of 256
offspring, while considerably more than half
of them would fall within the closely inter-
mediate classes included between 5^ and 6l/o
inches, the three middle classes of the diagram.

136

With six size-characters, the extreme size of a
grandparent would reappear no oftener than
once in 4000 times, while with a dozen such
independent characters it would recur only
once in some 17,000,000 times. It would be
remarkable if under such conditions the ex-
treme size were ever recovered from an ordi-
nary cross.

There is one means by which we can deter-
mine with certainty whether in a particular
case of seemingly blending inheritance segre-
gation does or does not occur, namely, by com-
paring the variability of the Fl and the F2
generations. If segregation does not occur, F2
should be no more variable than Fx, whereas
if segregation does occur, F2 should be more
variable. For, in a segregating system, the F±
individuals should all fall in a middle, inter-
mediate group, but the F2 individuals should
be distributed also in classes more remote from
a strictly intermediate position, that is, they
should be more variable. But, in a non-segre-
gating system, Ft and F2 individuals alike
should fall in the same intermediate group,
that is, they should have the same variability.

137

HEREDITY

The matter should be easy of determination
by observation of considerable numbers of i\
and F2 offspring. Investigations are now in
progress to test this matter.

My colleague, Dr. East, has found clear evi-
dence that, in maize, size-characters, although
they give a blending result in Fj, nevertheless
give segregation in F2. The character to be con-
sidered relates to length of ear in corn. A single
illustration will suffice. The variation in two
pure varieties is shown in the two upper rows of
Fig. 48. The " Length " of each class is given
in centimetres, its frequency just below at ' ' No.
Var., ' ' abbreviation for number of variates. The
variation in the Fx offspring obtained by cross-
ing the two pure varieties is shown in the third
row, and that of the F2 offspring in the lowest
row. Note that the variability in the Fx gen-
eration is not increased; its range is interme-
diate between the range in the parental varie-
ties. In the F2 generation, however, the varia-
bility is so increased that it includes almost
the entire range of both parental varieties, to-
gether with the intervening region.

In the light of this evidence it is clear that
138

3 :r ». AS 16 ,'f 10 7 1

. X-' I 10 II It *7 73 6S (S 39 2S (f 9 f .

FIG. 48. — Photographs to show variation in ear length of two varie-
ties of maize (upper row), of their F! offspring (second row), and of
their F2 offspring (third row). (After East.)

" BLENDING " INHERITANCE

in maize, seemingly blending is really segre-
gating inheritance, but with entire absence of
dominance, and it seems probable that the same
will be found to be true among rabbits and
other mammals; failure to observe it hitherto
is probably due to the fact that the factors
concerned are numerous. For the greater the
number of factors concerned, the more nearly
will the result obtained approximate a com-
plete and permanent blend. As the number
of factors approaches infinity, the result will
become identical with a permanent blend.

Theoretically it is important to know whether
segregating units are involved in inheritance
which we call blending; practically it does not
matter much, since if these units are only as
numerous as six or eight it will be practically
impossible to undo the effects of a cross and to
recover again the conditions obtaining previous
to the cross. The great majority of the offspring
both in the first and in subsequent generations
following the cross will be strictly intermediate
between the conditions crossed whether several
units, an infinite number of units, or no units
at all are involved.

139

HEREDITY

A practical question of some importance is
how to manipulate simultaneously blending (or
seemingly blending) and Mendelian inheritance.
This must be by a system of line-breeding in
alternate generations, not in successive genera-
tions. To test the practicability of this matter
I several years ago set myself the task of com-
bining in one race the large size of some lop-
eared, yellow rabbits which I had, with the
albino character of some small white rabbits
of common race. A first cross produced gray
rabbits of intermediate size, but no white ones.
On inbreeding the gray animals, there were
obtained in F2 white young of intermediate
size. These were now crossed again with the
original yellow stock, and again colored young
were obtained, but now with % of the desired
increase in size. These bred inter se again
produced albinos, this time of the % size. A
third cross with the original large stock brought
the size up to % of that desired, and combined
it in F2 with the desired albinism. Having
satisfied myself of the correctness of the
method, the experiment was now discontinued.
By further crosses, especially with a fresh lop-

140

" BLENDING " INHEEITANCE

eared stock, to avoid ill-effects of inbreeding,
the size could have been still further increased,
with judicious selection doubtless up to the ex-
treme size of colored lop-eared rabbits.

The general conclusion to be drawn is that
in attempting to combine in one race by cross-
breeding characters which exist separately in
different races, one should first inquire very
carefully how each character, in which the races
differ, behaves in transmission, for on the an-
swer to this question should depend the mode
of procedure to be chosen.

If simple Mendelian characters only are con-
cerned, nothing is required but to cross the two
races and select from the second generation
offspring the desired combination. If blending
characters only are concerned and F± yields
the desired blend, this is secure without fur-
ther procedure, except possibly selection to re-
duce its variability; but if the desired blend
is not yet secured, further back-crossing with
one race or the other may be necessary. If,
finally, both blending and Mendelian characters
are simultaneously involved in a cross, then the
method of combined line-breeding and selec-

141

HEREDITY

tion in alternate generations, already described,
should be adopted.

BIBLIOGRAPHY

CASTLE, W. E.

1909. (See Bibliography to Chapter V.)
EAST, E. M.

1910. "A Mendelian Interpretation of Variation that is
Apparently Continuous." American Naturalist, 44,
pp. 65-82.

EMERSON, R. A.

1910. " The Inheritance of Sizes and Shapes in Plants."

American Naturalist, 44, pp. 739-746.
NILSSON-EHLE, H.

1909. " Kreuzungsuntersuchungen an Hafer und Weizen."
Lunds Universitets Arsskrift, 5, No. 2, 122 pp.

CHAPTER IX

THE EFFECTS OP INBREEDING

WHAT is the probable source of the
evil effects which have been fre-
quently observed to follow in-
breeding?

By inbreeding we mean the mating of closely \
related individuals. As there are different de-
grees of relationship between individuals, so
there are different degrees of inbreeding. The
closest possible inbreeding occurs among plants
in what we call self-pollination, in which the
egg-cells of the plant are fertilized by pollen-
cells produced by the same individual. A simi-
lar phenomenon occurs among some of the
lower animals, notably among parasites. But
in all the higher animals, including the domes-
ticated ones, such a thing is impossible because
of the separateness of the sexes. For here no
individual produces both eggs and sperm. The

143

HEKEDITY

nearest possible approach to self-pollination is
in such cases the mating of brother with sister,
or of parent with child. But this is less close
inbreeding than occurs in self-pollination, for
the individuals mated are not in this case iden-
tical zygotes, though they may be similar ones.

It has long been known that in many plants
self-pollination is habitual and is attended by
no recognizable ill-effects. This fortunate cir-
cumstance allowed Mendel to make his remark-
able discovery by studies of garden-peas, in
which the flower is regularly self-fertilized, and
never opens at all unless made to do so by
some outside agency. Self-pollination is also
the rule in wheat, oats, and the majority of the
other cereal crops, the most important econom-
ically of cultivated plants. Crossing can in such
plants be brought about only by a difficult
technical process, so habitual is self-pollina-
tion. And crossing, too, in such plants is of
no particular benefit, unless by it one desires
to secure new combinations of unit-characters.

In maize, or Indian corn, however, among
the cereals, the case is quite different. Here
enforced self-pollination results in small un-

144

EFFECTS OF INBREEDING

productive plants, lacking in vigor. But racial
vigor is fully restored by a cross between two
depauperate unproductive individuals obtained
by self-fertilization, as has recently been shown
by Shull. This result is entirely in harmony
with those obtained by Darwin, who showed
by long-continued and elaborate experiments
that while some plants do not habitually cross
and are not even benefited by crossing, yet in
many other plants crossing results in more
vigorous and more productive offspring; that
further, the advantage of crossing in such cases
has resulted in the evolution in many plants
of floral structures, which insure crossing
through the agency of insects or of the wind.

In animals the facts as regards close fer-
tilization are similar to those just described
for plants. Some animals seem to be indiffer-
ent to close breeding, others will not tolerate
it. Some hermaphroditic animals (those which
produce both eggs and sperm) are regularly
self -fertilized. Such is the case, for example,
with many parasitic flat- worms. In other cases
self-fertilization is disadvantageous. One such
case I was able to point out some fifteen years

145

ago, in the case of a sea-squirt or tunicate,
Ciona. The same individual of Ciona produces
and discharges simultaneously both eggs and
sperm, yet the eggs are rarely self-fertilized,
for if self-fertilization is enforced by isolation
of an individual, or if self-fertilization is
brought about artificially by removing the eggs
and sperm from the body of the parent and
mixing them in sea-water, very few of the
eggs develop, — less than 10%. But if the
eggs of one individual be mingled with the
sperm of any other individual whatever, prac-
tically all of the eggs are fertilized and
develop.

In the great majority of animals, as in many
plants, self-fertilization is rendered wholly im-
possible by separation of the sexes. The same
individual does not produce both eggs and
sperm, but only one sort of sexual product.
But among sexually separate animals the same
degree of inbreeding varies in its effects. The
closest degree, mating of brother with sister,
has in some cases no observable ill-effects.
Thus, in the case of a small fly, Drosophila,
my pupils and I bred brother with sister for

146

EFFECTS OF INBREEDING

fifty-nine generations in succession without ob-
taining a diminution in either the vigor or the
fecundity of the race, which could with cer-
tainty be attributed to that cause. A slight
diminution was observed in some cases, but
this was wholly obviated when parents were
chosen from the more vigorous broods in each
generation. Nevertheless crossing of two in-
bred strains of Drosophila, both of which were
doing well under inbreeding, produced off-
spring superior in productiveness to either
inbred strain. Even in this case, therefore,
though inbreeding is tolerated, cross-breeding
has advantages.

In the case of many domesticated animals,
it is the opinion of experienced breeders, sup-
ported by such scientific observations as we
possess, that decidedly bad effects follow con-
tinuous inbreeding. Bos ( '94) practiced con-
tinuous inbreeding with a family of rats for
six years. No ill-effects were observed during
the first half of the experiment, but after that
a rapid decline occurred in the vigor and fer-
tility of the race. The average-sized litter in
the first half of the experiment was about 7.5,

147

HEREDITY

but in the last year of the experiment it had
fallen to 3.2, and many pairs were found to
be completely sterile. Diminution in size also
attended the inbreeding, at the end amounting
in the case of males to between 8 and 20 %.

Experiments made by Weismann confirm
those of Bos as regards the falling off in fer-
tility due to inbreeding. For eight years Weis-
mann bred a colony of mice started from nine
individuals, — six females and three males.
The experiment covered 29 generations. In
the first 10 generations the average number of
young to a litter was 6.1; in the next 10 gen-
erations, it was 5.6; and in the last 9 genera-
tions, it had fallen to 4.2. But sweeping
generalizations cannot be drawn from these
cases. Each species of animal must probably
be tested for itself before we shall know what
the exact effects of inbreeding are in that case.
In guinea-pigs, a polydactylous race built up
by the closest inbreeding out of individuals all
descended from one and the same individual
has now been in existence for ten years. It
consists of one of the largest and most vigor-
ous strains of guinea-pigs that I have ever

148

seen, and has shown no indications of dimin-
ished fertility.

In the production of pure breeds of sheep,
cattle, hogs, and horses inbreeding has fre-
quently been practiced extensively, and where
in such cases selection has been made of the
more vigorous offspring as parents, it is doubt-
ful whether any diminution in size, vigor, or
fertility has resulted. Nevertheless it very
frequently happens that when two pure breeds
are crossed, the offspring surpass either pure
race in size and vigor. This is the reason for
much cross-breeding in economic practice, the
object of which is not the production of a new
breed, but the production for the market of
an animal maturing quickly or of superior size
and vigor. The inbreeding practiced in form-
ing a pure breed has not of necessity dimin-
ished vigor, but a cross does temporarily (that
is in the Fj generation) increase vigor above
the normal. Now why should inbreeding un-
attended by selection decrease vigor, and cross-
breeding increase it? We know that inbreed-
ing tends to the production of homozygous
conditions, whereas cross-breeding tends to
11 149

HEREDITY

produce heterozygous conditions. Under self-
pollination for 1 generation following a cross,
half the offspring become homozygous; after
2 generations, % of the offspring are homo-
zygous ; after 3 generations % are homozygous,
and so on. So if the closest inbreeding is
practiced there is a speedy return to homo-
zygous, pure racial conditions. We know fur-
ther that in some cases at least heterozygotes
are more vigorous than homozygotes. The
heterozygous yellow mouse is a vigorous lively
animal; the homozygous yellow mouse is so
feeble that it perishes as soon as produced,
never attaining maturity. Cross-breeding has,
then, the same advantage over close-breeding
that fertilization has over parthenogenesis. It
brings together differentiated gametes, which,
reacting on each other, produce greater meta-
bolic activity.

Inbreeding, also, by its tendency to secure
homozygous combinations, tends to bring to
the surface latent or hidden recessive charac-
ters. If these are in nature defects or weak-
nesses of the organism, such as albinism and
feeble-mindedness in man, then inbreeding is

150

EFFECTS OF INBREEDING

distinctly bad. Existing legislation against
the marriage of near-of-Mn is, therefore, on
the whole, biologically justified. On the other
hand, continual crossing only tends to hide
inherent defects, not to exterminate them; and
inbreeding only tends to bring them to the
surface, not to create them. We may not,
therefore, lightly ascribe to inbreeding or in-
termarriage the creation of bad racial traits,
but only their manifestation. Further, any
racial stock which maintains a high standard
of excellence under inbreeding is certainly one
of great vigor, and free from inherent defects.
The animal breeder is therefore amply jus-
tified in doing what human society at present
is probably not warranted in doing, — viz. in
practicing close inbreeding in building up
families of superior excellence and then keep-
ing these pure, while using them in crosses
with other stocks. For an animal of such a
superior race should have only vigorous, strong
offspring if mated with a healthy individual
of any family whatever, within the same spe-
cies. For this reason the production of
" thoroughbred ': animals and their use in

151

HEREDITY

crosses is both scientifically correct and com-
mercially remunerative.

BIBLIOGRAPHY

BOS, RlTZEMA.

1894. "Untersuchungen ueber die Folgen der Zucht in
engster Blutverwandtschaft." Biol. Centrbl., 14, pp. 75-
81.
CASTLE, W. E., CARPENTER, F. W., CLARK, A. H., MAST,

S. O., and BARROWS, W. M.

"The Effects of Inbreeding, Cross-breeding and Selec-
tion upon the Fertility and Variability of Drosophila."
Proc. Amer. Acad. Arts and Sci., 41, pp. 731-786.
GUAITA, G. VON.

1898. "Versuche mit Kreuzungen von verschiedenen
Rassen des Hausmaus." Ber. naturf. Gesellsch. zu
Freiburg, 10, pp. 317-332. [Contains observations of
Weismann.]

CHAPTER X

HEEEDITY AND SEX

THE value of a domesticated animal often
depends in considerable measure on its
sex. Therefore, if a means could be
devised for controlling the sex of offspring, it
would be of great economic value to the breeder.
Endless attempts have been made to do this,
and occasionally a claim of success has been
made, but none of these claims has withstood
the test of critical analysis or experiment. The
hypotheses advanced to explain how sex may
be controlled have been of the most varied
character. In some the determination of sex
has been supposed to inhere in the nature of
the parents, in others it is referred to condi-
tions of the gametes themselves.

Relative age or vigor of the parents have
been supposed to influence sex in various ways.
The same idea has been advanced regarding

153

HEREDITY

the gametes themselves, it being supposed that
early or late fertilization of the egg might
influence its sex. Experimental evidence, how-
ever, as to these several hypotheses is wholly
negative, when one eliminates other possible
factors from the experiment. Everything
points to the conclusion that sex rests in the
last analysis upon gametic differentiation, just
as the color of a guinea-pig in a mixed race
of blacks and whites depends upon whether the
gametes which unite to produce it carry black
or white. As the heterozygous black guinea-
pig forms black-producing and white-producing
gametes in equal numbers, so there is reason
to think male-producing and female-producing
gametes are formed in equal numbers by the
parent, in many cases at least. But is it not
possible that there may exist individuals which
produce the two sorts of gametes in unequal
numbers, and so would have a tendency to
produce more offspring of one sex than of the
other? Perhaps so, though we have no evi-
dence that such a condition, if it does exist, is
transmitted from one generation to another.
On this point I made experimental observa-

154

HEREDITY AND SEX

tions upon guinea-pigs extending over a series
of years. Oftentimes I found an individual
that produced more offspring of one sex than
of the other, but this was probably due merely
to chance deviations from equality. I could
get no evidence that the condition was inher-
ited, though the experiment was continued
through as many as seven generations, includ-
ing several hundred offspring.

The essential difference between a female
and a male individual is that one produces
eggs, the other sperm. All other differences
are secondary and dependent largely upon the
differences mentioned. If in the higher ani-
mals (birds and mammals) the sex-glands (i. e.
the egg-producing and sperm-producing tis-
sues) are removed from the body, the super-
ficial differences between the sexes largely dis-
appear. In insects, however, the secondary
sex-characters seem to be for the most part
uninfluenced by presence or absence of the
sex-glands. Their differentiation occurs in-
dependently though simultaneously with that
of the sex-glands.

The egg or larger gamete (the so-called
155

HEREDITY

macro-gamete) in all animals is non-motile and
contains a relatively large amount of reserve-
food material for the maintenance of the de-
veloping embryo. This reserve-food material
it is the function of the mother to supply. In
the case of some animals, for example flat-
worms and mollusks, the food-supply of the
embryo is not stored in the egg-cell itself, but
in other cells associated with it, and which
break down and supply nourishment to the
developing embryo derived from the fertilized
egg. Again, as in the mammals, the embryo
may derive its nourishment largely from the
maternal tissues, the embryo remaining like a
parasite within the maternal body during its
growth, feeding by absorption. But in all cases
alike the mother supplies the larger gamete
and the food-material necessary to carry the
zygote through its embryonic stages. The
father, on the other hand, furnishes the bare
hereditary equipment of a gamete, with the
motor apparatus necessary to bring it into
contact with the egg-cell, but without food for
the developing embryo produced by fertiliza-
tion. The gamete furnished by the father is

156

HEREDITY AND SEX

therefore the smaller gamete, the so-called
micro-gamete.

From the standpoint of metabolism, the
female is the more advanced condition; the
female performs the larger function, doing all
that the male does in furnishing the material
basis of heredity (a gamete), and in addition
supplying food for the embryo. As regards
the reproductive function, the female is the
equivalent of the male organism, plus an ad-
ditional function, — that of supplying the em-
bryo with food. When we come to consider the
structural basis of sex, we find reasons for
thinking that here, too, the female individual
is the equivalent of the male plus an addi-
tional element. The conclusion has very natu-
rally been drawn that if a means could be
devised for increasing the nourishment of the
egg or embryo, its development into a female
should be thereby insured, while the reverse
treatment should lead to the production of a
male. But in practice this a priori expecta-
tion is not fulfilled. Better nourishment of the
mother may lead to the production of more
eggs, but not of more female offspring, as has

15?

HEREDITY

been repeatedly demonstrated by experiment.
Also poor nutrition of the mother may diminish
the number of eggs which she liberates, but will
not increase the proportion of males among the
offspring produced.

An excellent summary of evidence on this
point was made by Cuenot in 1900. Attempts
to influence the sex of an embryo or larva by
altered nutrition of the embryo or larva itself
have proved equally futile. Practically the
only experimental evidence of value in favor
of this idea has been derived from the study
of insects, and this is capable of explanation
on quite different grounds from those which
first suggest themselves. It has sometimes been
observed, as by Mary Treat for example, that
a lot of insects poorly fed produce an excess
of males. In such lots, however, the mortality
is commonly high, and more females die than
males, because the female is usually larger and
requires more food to complete its development.
The fallacy in concluding from such evidence
that scanty nutrition causes individuals which
would otherwise become females to develop
into males was indicated years ago by Eiley.

158

HEREDITY AND SEX

Nevertheless an argument for the artificial con-
trol of sex based on such evidence is from time
to time brought forward, as, for example, a
few years since by Schenk. The latest advocate
of sex-control by artificial means is an Italian,
Russo (1909). He claims in the case of rabbits
that by feeding the mother on lecithin or by
injections of lecithin, the proportion of female
births may be increased. His evidence in sup-
port of this claim is, however, wholly inade-
quate, and two independent repetitions of his
experiments, made by Basile in Italy and by
Punnett in England, have given entirely nega-
tive results.

An alternative hypothesis concerning the de-
termination of sex has been steadily gaining
ground during the last ten years, that sex has
its beginning in gametic differentiation and is
finally determined beyond recall in the ferti-
lized egg by the nature of the uniting gametes.
Instructive in this connection is a study of
parthenogenesis, — reproduction by unfertilized
eggs. But before entering upon this, it may
be well to review briefly the changes which
regularly take place in the egg which is to be

159

HEREDITY

fertilized, and compare with this the changes
which occur in eggs not to Be fertilized.

In each cell of the ordinary animal there
occurs a characteristic number of bodies called
chromosomes. We do not know that they are
any more important than other cell constitu-
ents, but we know their history better. These
are contained in the nucleus of the cell, and at
the time of nuclear division they are found at
the equator of the division spindle. For ex-
ample, in the egg of the mouse (Fig. 4, A), the
nucleus is seen to be in the spindle stage, and
its chromosomes are gathered together at the
equator of the spindle. There each of them
regularly splits in two, and one derivative goes
to either end of the spindle, and so into one of
the daughter-nuclei. Thus each new nucleus
has, as a rule, the same chromosome composi-
tion as the nucleus from which it was derived.

But the egg which is to be fertilized under-
goes two nuclear divisions in succession, in only
one of which do the chromosomes split (see
Fig. 4, A-D). In the other division the chromo-
somes separate into two groups without split-
ting, and each group goes into a different cell

160

HEREDITY AND SEX

product. Consequently, in each of these prod-
ucts the number of chromosomes is reduced to
half what it is in the cells of the parental body.
Thus in the egg of the mouse, by maturation,
the number of chromosomes becomes reduced
from about twenty-four to about twelve.

Similar changes occur in the developing
sperm-cell (see Fig. 5). Starting with the
double or 2 N chromosome number, there are
formed by two nuclear divisions, with only one
splitting of chromosomes, four cells, each with
the reduced or simplex number of chromosomes,
N. Consequently, when the sperm enters the
egg at fertilization it brings in a group of N
chromosomes (in the mouse apparently twelve),
which, added to the egg-contribution of N
chromosomes, brings the number in the new
organism again up to 2 N (in the mouse twenty-
four).

Now, as regards the maturation of partheno-
genetic eggs, those which are to develop with-
out having been fertilized, three categories of
cases deserve separate discussion. The simplest
of these in many respects is found among the
social hymenoptera (ants, bees, and wasps).

161

HEEEDITY

See Fig. 49, left column. The eggs are, so far
as we can discover, all of a single type. They

BEE KOTIFER APHID

?

\/ ^ — " / v, —
/» \ in \

FIG. 49. — Diagram of sex-determination in parthenogenesis.
First row, nuclear condition of the parthenogenetic mother;
second row, of her eggs when they develop without reduction,
after forming a single polar cell; third row, condition of the
eggs after complete maturation — the fertilized egg in each
case produces a male; fourth row, nuclear condition of the
fertilized egg, always a female.

undergo maturation in the manner already de-
scribed, the chromosomes being reduced to the

162

HEREDITY AND SEX

N or simplex number. The eggs of most ani-
mals, after they have undergone reduction, are
incapable of development unless fertilized, but
those of the hymenoptera may develop either
fertilized or unfertilized. In the former case
a female is produced, in the latter a male. The
simplex, or N condition is in this case the male,
the duplex or 2 N condition is the female, natu-
rally the one of higher metabolic activity, the
one which forms the macro-gametes.

In an earlier chapter I explained how the
development of the sperm-cells in a male having
the reduced or simplex number of chromosomes
differs from that in the ordinary male. Refer-
ence to Fig. 8 may help to recall this. The
cells of the male are in this case already in the
reduced or simplex condition, N. In the pro-
duction of the sperms the reducing division is
omitted so far as nuclear components are con-
cerned, so that each sperm formed contains the
full simplex chromosome number, N. If it were
less, the gamete formed would perhaps not be
capable of transmitting all the hereditary char-
acteristics of an individual.

A second category of cases (Fig. 49, middle
163

HEREDITY

column) is represented by such simple aquatic
organisms as rotifers and small Crustacea, like
Daphnia. In these parthenogenesis occurs ex-
clusively, when the food supply is very abun-
dant and conditions otherwise favorable, whereas
reproduction by fertilized eggs occurs only
when external conditions, including food-supply,
are not good. Under favorable conditions only
female offspring are produced. The conclusion
has naturally but erroneously been drawn that
good nutrition in itself favors the production
of females in animals generally, which is not
true. The egg produced by Daphnia, or by a
rotifer, under optimum conditions does not un-
dergo reduction (see Fig. 49, second row). It
remains in the 2 N condition, forming but a
single polar cell. It is therefore unprepared
for fertilization, and in fact it is not fertilized.
Its sex is like that of the animal which formed
it, female. Under unfavorable conditions, how-
ever, the eggs of the rotifer and of Daphnia
do not begin development until they have un-
dergone maturation. They are also of two
sizes (Fig. 49, third row), — small eggs, which
develop without fertilization and which form

164

HEREDITY AND SEX

males, and large eggs, which require fer-
tilization, and which form females. In this
category of cases, as in that of the hymenop-
tera, the egg which develops in the 2 N condi-
tion, either from failure of reduction to occur
in maturation or from fertilization following
reduction, forms a female; whereas the egg
which develops in the N condition forms a
male.

In a third category of cases there is a quan-
titative difference in chromatin between male
and female, just as in the foregoing cases, but
this does not amount to a whole set of chromo-
somes, N, but to only a partial set, one or two
chromosomes (see Fig. 49, right column). This
category of cases occurs in plant-lice (aphids
and phylloxerans) ; evidence of its existence
rests chiefly on recent observations made by
von Baehr and Morgan. Females are formed
by parthenogenesis without reduction, occurring
under favorable conditions, just as in the case
of rotifers. Females are also formed by fer-
tilization following reduction under unfavor-
able conditions, just as in rotifers. In both
cases the female is 2 N. Males arise only by
12 165

HEEEDITY

parthenogenesis under unfavorable conditions,
just as in rotifers, but the reduction which
occurs before development begins is partial
only. A whole set, N, of chromosomes is not
eliminated in maturation, but only 1 or 2 chro-
mosomes. Hence the male condition here is
2 N — 1 or — 2. The condition of the gametes
formed, however, is N in both sexes. In
spermatogenesis, division of the germ-cells
takes place into N and N -- 1 daughter cells,
but the latter degenerate (like the non-nucleated
cells of the bee and wasp), and only the former
produce spermatozoa. Hence in fertilization
only 2 N zygotes are produced, which are in-
variably female.

Summarizing the three categories described,
we may say that in all known cases of par-
thenogenesis, the female is in the duplex, 2 N
condition, the male in the simplex (N) or par-
tially duplex condition (2N--1, or 2 N — 2).
The female in all cases has the greater chro-
matin content.

In a great many insects and other arthro-
pods, which are not parthenogenetic, it is
known that, although the male, like the female,

166

HEEEDITY AND SEX

develops only from a fertilized egg, neverthe-
less the male possesses fewer chromosomes than

FIG. 50. — Diagram of sex-determination when the female is
homozygous, the male heterozygous.

the female. In such cases the female forms,
as in cases of parthenogenesis, only N gametes,
but the male forms gametes of two sorts, N and

167

HEREDITY

N — 1 or N — 2 (see Fig. 50). In consequence
zygotes of two sorts result, — those which are
2 N, female, and those which are 2 N -- 1 or
2 N — 2, male. Thus in the squash-bug, Anasa-
tristis, according to Wilson, the mature egg
contains 11 chromosomes, the spermatozoa
either 10 or 11 chromosomes, the two sorts
being equally numerous.

Egg 11 + sperm 11 produces a zygote 22 (2N), a female;
Egg 11+ " 10 " " " 21(2N-l),amale.

N in this species = 11 ; 2 N = 22, the female ;
2 N — 1 == 21, the male. Males and females are
therefore approximately equal in number, as in
most animals where the two sexes are not sub-
ject to unequal mortality. In the Mendelian
sense the female is in such cases a homozygote,
the male a heterozygote. The sex of an indi-
vidual in such cases depends upon which sort
of a sperm chances to enter the egg.

But the experimental evidence indicates
that both as regards sex and as regards
heritable characters correlated with sex, these
relations may in some cases be reversed, the
female being heterozygous, the male homozy-

168

HEREDITY AND SEX

gous. In such cases there is reason to think
that structurally the male is 2 N but the female
2 N -J-. That is, the female is still the equiva-
lent of the male plus some additional element
and function. A structural basis in the chromo-
somes for such a condition has been described
by Baltzer in the case of the sea-urchin. He
found the regular duplex number of chromo-
somes in the male; but in the female, while
the number was the same, one of the chromo-
somes was larger than its mate, having an extra
or odd element attached to it. In such a case
the gametes formed by the male would all be
N, but those formed by the female would be
of two sorts equally numerous, viz. N and N -f-
(see Fig. 51). Egg N fertilized by sperm N
would produce a zygote 2 N, a male ; egg N -(-
fertilized by sperm N would produce a zygote
2 N +, a female. Hence here, as in other ani-
mals, the sexes would be approximately equal,
but the sex of a particular individual would
depend upon which sort of egg gave rise to it.
Upon the existence, as in the foregoing cases,
of an unpaired or odd structural element in
the egg, may perhaps depend the explanation

169

HEEEDITY

of a curious sort of heredity known as sex-
limited heredity.

FIG. 51. — Diagram of sex-determination when the female is
heterozygous, the male homozygous.

Every one who knows anything about poultry
is acquainted with the popular American breed
known as the barred Plymouth Bock. In this

170

HEEEDITY AND SEX

breed the feathers are marked with alternate
bars of darker and lighter black. Pure barred
Eocks breed true, but when crossed with other
breeds, the male proves to be homozygous, the
female, heterozygous in barring. For the male
Rock crossed with a non-barred breed produces
only barred offspring in both sexes, but the
female Rock crossed with the same non-barred
breed produces offspring approximately half
of which are barred, the other half being non-
barred. Further, the barred individuals in this
cross are invariably males, the non-barred ones
being females. Accordingly, the distribution of
barring and non-barring in the cross is sex-
limited.

The barred offspring produced by a cross
between barred Plymouth Rocks and a non-
barred breed, whether those offspring are males
or females, prove to be heterozygous in barring,
as we should expect, the barring factor having
been received only from one parent, the barred
one. Further, the non-barred offspring pro-
duced by a barred Rock female crossed with
a non-barred breed, do not transmit barring,
hence they are pure recessives as regards bar-

171

HEREDITY

ring. Hence, also, we are forced to conclude,
as already suggested, the female of the pure
barred Rock breed is heterozygous as re-
gards barring, and transmits the character only
to her male offspring, her female offspring
(if the father is non-barred) neither being
themselves barred nor being able to transmit
barring.

A pure Plymouth Bock race breeds true to
barring merely because all its males are pure,
for the females are not pure. This is shown
by the following experiment. If a heterozygous
barred male, produced by a cross between a
Rock and a non-barred breed, is crossed with
barred females, either those of a pure Rock
race or those produced by a cross, the result
is the same. The male offspring are all barred ;
the females, half of them barred, half non-
barred. This result shows that all barred
females alike are heterozygous in barring.

Sex-limited inheritance such as this finds at
the present time its most probable explanation
in the existence in the egg of an extra or plus
element never found in the sperm, this element
pairing with the sex-limited character in the

172

HEBEDITY AND SEX

reduction division. Thus, in the barred Rock,
calling barring B, the male of pure race is
plainly B B and every sperm is B. But the
female clearly contains only one B and can-

FEMALE

MALE

(eg)

FIG. 52. — Diagram of sex-limited inheritance when the female
is a heterozygote, as in barred fowls. X, female sex deter-
miner; B, barring.

not be made to contain two. Perhaps a second
B is kept out by some structural element, X,
the distinctive structural element of the female
individual. Then the eggs will be of two
sorts: B and X. Since the sperms are all B,
the first type of egg when fertilized will con-
tain B B, a homozygous barred individual and

173

HEREDITY

a male, since it lacks X; the second type will
contain B X, a bird heterozygous in barring,
and a female, since it contains X. This agrees
with the experimental result (see Fig. 52).

A heterozygous barred male will form two
kinds of sperm, only one of which will contain
B. If such a male be mated with a barred
female, four sorts of zygotes should result, as
follows :

Gametes of heterozygous barred male = B and — ,
Gametes of barred female = B and X,

Zygotes = B B (homozygous barred male) ; B —

(heterozygous barred male), B'X (barred female), and

—•X (non-barred female).

The observed result of this cross accords
fully with the foregoing expectation.

The sex-limited inheritance of barring in
fowls may be explained, as we have just seen,
on the assumption that the female is the hetero-
zygous sex. The same is true of sex-limited
inheritance in canary-birds and in the moth,
Abraxas, according to Bateson and Doncaster.
But these relations are exactly reversed in the
pomace-fly, Drosophila ampelophila, according
to Morgan.

174

HEREDITY AND SEX

In Drosophila the female is apparently
homozygous as regards some cell-structure,
X, which in the male is never represented
more than once. Accordingly the formula of

FEMALE MALE

FIG. 53. — Diagram of sex-limited inheritance when the female
is a homozygote, as in the red-eyed Drosophila. X, sex-
determiner; R, red-eyes.

the female is in such cases XX; that of the
male, X — . Now the sex-limited characters in
Drosophila seem to be bound up with the X
structure, not repelled by it, as is barring in
fowls. Accordingly, a sex-limited character
may be represented twice in the female Droso-
phila, but only once in the male; or in other

175

HEREDITY

words, the female may be homozygous as re-
gards a sex-limited character, but the male can
only be heterozygous (see Fig. 53).

Drosophila normally has red eyes, but the
redness of the eye is a distinct unit-character,
sex-limited in heredity. Further males are
regularly heterozygous in this character, while
females are homozygous. For Morgan has ob-
tained a race in which the eyes are white,
owing to the loss of the red character; and
reciprocal crosses of this race with ordinary
red-eyed animals yield different results. The
red-eyed female crossed with a white-eyed male
produces only red-eyed offspring, but the red-
eyed male crossed with a white-eyed female
produces offspring only half of which are red-
eyed, viz. the females, whereas the males are
white-eyed.

These different results in the two cases ap-
parently come about as follows:

First case.

Gametes of red-eyed female = X-R and X-R,
Gametes of white-eyed male = X and — ,
Zygotes= X'X-R (red-eyed female), and = — -X-R
(red-eyed male).

176

HEREDITY AND SEX

Second case.

Gametes of white-eyed female = X and X,
Gametes of red-eyed male = X-R and — ,
Zygotes == X'X-R (red-eyed female), and — -X
(white-eyed male).

A short condition of the wings in Drosophila,
which renders the animal incapable of flight,
is likewise sex-limited in heredity, as has been
shown by Morgan. By crossing two races of
Drosophila, each of which possessed a different
sex-limited character, Morgan has been able to
combine the two characters in a single race.
Thus was obtained a race both white-eyed and
short-winged. The synthesis cannot be made
originally in a male individual, but only in a
female. For only in the female can the two
characters be brought together, each associated
with a different X, since in the male only one
X is present. Although each sex-limited char-
acter seems to be attached to or bound up with
an X structure, it evidently has a material basis
distinct from X. Otherwise it would not be
possible for the character to leave one X and
attach itself to the other, as apparently takes
place in the female when the combination of

177

HEREDITY

two sex-limited characters in the same gamete
is secured through a cross. The combination
is apparently secured in this way:

Gametes uniting, X-R and X— L,

Zygote formed, X-R'X-L,

Its gametes, X-R and X-L, or X-R-L and X.

One of the uniting gametes, X-R, is formed
by the red-eyed, short- winged parent ; the other,
X-L, is formed by the long-winged, white-
eyed parent. The zygote resulting is a red-
eyed individual, since it contains R ; it is long-
winged, since it contains L ; it is a female, since
it contains two Xs. Now, its gametes are of
four sorts, as indicated. The first two sorts
result from simple separation of the two Xs,
each with its associated character, R in one
case, L in the other. But the third sort could
result only from the attachment of R and L
to the same X, leaving the other X without
either R or L as the fourth kind of gamete.
This kind, which transmits neither red eyes nor
long wings, would represent the new gametic
combination,— white-eyed and with short wings.

The experimental evidence that gametes of
178

HEREDITY AND SEX

these four sorts are formed by females of the
origin described is as follows : — When such a
female is mated with a long- winged, white-eyed
male, there are obtained female offspring, all
of which are long- winged, but half of them are
red-eyed, half white-eyed. The male offspring,
however, are of four sorts, viz. red short, white
long, red long, and white short. This result
harmonizes with the hypothesis advanced. For
if the gametes of the female are X-R, X-L,
X-R-L, and X, and those of the male are X-L
and — , then the following combinations should
result :

X-L' X-R, red long female,
X-L' X-L, white long female
X-L' X-R-L, red long female,
X-L' X , white long female,

• X-R, red short male,

• X-L, white long male,

• X-R-L, red long male,

• X , white short male.

This expected result accords with that ac-
tually obtained by Morgan.

Color-blindness in man is a sex-limited char-
acter, the inheritance of which resembles that

179

HEBEDITY

of white eyes or short wings in Drosophila,
rather than of barring in poultry.

Color-blindness is much commoner in men
than in women. A color-blind man, however,
does not transmit color-blindness to his sons,
but only to his daughters, the daughters, how-
ever, are themselves normal provided the
mother was; yet they transmit color-blindness
to half their sons. A color-blind daughter
could be produced, apparently, only by the
marriage of a color-blind man with a woman
who transmitted color-blindness, since the
daughter to be color-blind must have received
the character from both parents, whereas the
color-blind son receives the character only
from his mother.

Color-blindness is apparently due to a defect
in the germ-cell, — absence of something nor-
mally associated there with an X-structure,
which is represented twice in woman, once in
man. Color-blindness follows, therefore, in
transmission the scheme shown in Fig 53.

If, as has been suggested, the determination
of sex in general depends upon the inheritance
of a Mendelian factor differentiating the sexes,

180

HEREDITY AXD SEX

it is highly improbable that the breeder will
ever be able to control sex. Male and female
zygotes should forever continue to be produced
in approximate equality, and consistent inequal-
ity of male and female births could result only
from greater mortality on the part of one sort
of zygote than of the other. Only in partheno-
genesis can man at will control sex, and until
he can produce artificial parthenogenesis in the
higher animals, he can scarcely hope to con-
trol sex in such animals.

Negative as are the results of our study of sex-
control, they are perhaps not wholly without
practical value. It is something to know our
limitations. We may thus save time from
useless attempts at controlling what is uncon-
trollable and devote it to more profitable em-
ployments.

BIBLIOGRAPHY

BATESON, W.

1909. (See Bibliography to Chapter IV.)

CASTLE, W. E.

1909. "A Mendelian View of Sex-heredity." Science,
N. S., vol. 29, pp. 395-400.

CUENOT, L.

1900. "Sur la determination du sexe chez les animaux."
Bull. Sci. de la France et de la Belgique.

13 181

MORGAN, T. H.

1909. "A Biological and Cytological Study of Sex Deter-
mination in Phylloxerans and Aphids." Journal of
Experimental Zoology, 7, pp. 239-352.

1910. "Sex-limited Inheritance in Drosophila." Science,
N. S., 32, pp. 120-122.

1911. "The Application of the Conception of Pure Lines
to Sex-limited Inheritance and to Sexual Dimorphism."
The American Naturalist, 45, pp. 65-78.

Russo, A.

1909. "Studien liber die Bestimmung des weiblichen

Geschlectes." G. Fischer, Jena.
WILSON, E. B.

1909. "Recent Researches on the Determination and
Heredity of Sex." Science, N. S., 29, pp. 53-70.

1910. "The Chromosomes in Relation to the Determina-
tion of Sex." Science Progress, 5, pp. 570-592.

For references to the earlier literature see CU^NOT and
BATESON.

INDEX

Abraxas, sex determination in,

174.
Atavism, 62.

von Baehr, 165.
Basile, 159.
Bateson, 37, 110, 174.
Baur, 61.
Bos, 147.
Buttercup, 107.

Cattle, polled, 39, 102.

Ciona, 146.

Color blindness, 179.

Correns, 34.

Coutagne, 99.

Cross-breeding, effects of, on

vigor, 149.
Cuenot, 158.

Daphnia, 116, 164.
Darwin, 7, 62, 106.
Davenport, 100.
Doncaster, 174.
Drosophila, 146.
sex limited inheritance in,
174.

East, 130, 138.

Egg, fertilization of, 11.
of Nereis, 12.
of mouse, 13.
of sea-urchin, 18.

Factors, inhibiting, 55.

multiple, 62, 131.
Farabee, 39.
Fern, 20.
prothallus of, 22.
spores of, 22.
Fingers, inheritance of short,

39.

Fixation of a reversionary char-
acter, 68.

Fowls, Andalusian, 55.
crosses of, 53.

sex limited inheritance in,
170.

Gamete, definition of, 25.
Guinea-pig, angora, 38, 42, 47.

black crossed with white, 34.

new variety of, 84.

polydactylous, 100, 121.

reversion in, 63.

rough crossed with smooth,

38, 41, 47, 98.
Hare, Belgian, 30.

183

INDEX

Heape, 30. Rats, 91.

Heredity, collateral, 6. inbreeding in, 147.

definition of, 6. selection in, 123.

Hyalodaphnia, 117. Reduction, in fern, 22.

Reversion, 62.

Jennings, 111. Riddle, 87.

Johannsen, 45, 106. Riley, 158.

Rodents, coloration of, 72.
Little, 59. Rotifers, sex determination in,

164.

Maize, ear-length in, 138. Russo, 159.

Maturation of egg, 15, 20.

of sperm-cells, 17. Schenck, 159.

Mendel, 7, 34. Self-fertilization, 145.

Mendelian ratios, 35, 45, 50, 59. Sheep, horns in, 102.
Mice, pale-colored, 81. Silk moths, 99.

pink-eyed, 79. Snapdragon, golden variety of,

spotted, 83. 61.

yellow, 57. Spermatogenesis, 17.

Morgan, 91, 165, 174. of wasp, 24.

Mus alexandrinus, 91. Squash bug, sex determination

Mus norvegicus, 94. in, 168.

Mus rattus, 91, 94.
Mutilations, inheritance of, 29. Tornier, 102.

Transplantation of egg, 30.
Nilsson-Ehle, 131. of ovary, 31.

Treat, 158.
Parthenogenesis, artificial, 18.

sex determination in, 162. Unit-characters, 38.

Pigeons, reversion in, 62.

Phenotype, 45. De Vries, 34, 89, 106.

Phillips, 31.

Prepotency, 101. Weismann, 29, 148.

Punnett, 159. Wilson, 168.

Woltereck, 117.
Rabbits, size inheritance in,

129. Zygote, definition of, 25.

184

(i)

TWENTIETH CENTURY TEXT-BOOKS.

TEXT-BOOKS OF ZOOLOGY.

By DAVID STARR JORDAN, President of Leland Stan-
ford Jr. University; VERNON LYMAN KELLOGG, Professor
of Entomology; HAROLD HEATH, Assistant Professor of
Invertebrate Zoology.

Evolution and Animal Life.

This is a popular discussion of the facts, processes, laws, and theories
relating to the life and evolution of animals. The reader of it will have
a very clear idea of the all-important theory of evolution as it has been
developed and as it is held to-day by scientists. 8vo. Cloth, with about
300 illustrations, $2.50 net ; postage 20 cents additional.

Animal Studies.

A compact but complete treatment of elementary zoology, especially
prepared for institutions of learning that prefer to find in a single book
an ecological as well as morphological survey of the animal world. I2mo.
Cloth, $1.25 net.

Animal Life.

An elementary account of animal ecology — that is, of the relations
of animals to their surroundings. It treats of animals from the stand-
point of the observer, and shows why the present conditions and habits
of animal life are as we find them. I2mo. Cloth, $1.20 net.

Animal Forms.

This book deals in an elementary way with animal morphology. It
describes the structure and life processes of animals, from the lowest
creations to the highest and most complex. I2mo. Cloth, $1.10 net.

Animals.

This consists of "Animal Life" and "Animal Forms" bound in one
volume. I2mo. Cloth, $1.80.

Animal Structures.

A laboratory guide in the teaching of elementary zoology. I2mo.
Cloth, 50 cents net.

D. APPLETON AND COMPANY,

NEW YORK. BOSTON. CHICAGO. LONDON.

TWENTIETH CENTURY TEXT-BOOKS.

TEXT-BOOKS OF ZOOLOGY.

By DAVID STARR JORDAN, President of Leland Stan-
ford Jr. University; VERNON LYMAN KELLOGG, Professor
of Entomology; HAROLD HEATH, Assistant Professor of
Invertebrate Zoology.

Evolution and Animal Life.

This is a popular discussion of the facts, processes, laws, and theories
relating to the life and evolution of animals. The reader of it will have
a very clear idea of the all-important theory of evolution as it has been
developed and as it is held to-day by scientists. 8vo. Cloth, with about
300 illustrations, $2.50 net ; postage 20 cents additional.

Animal Studies.

A compact but complete treatment of elementary zoology, especially
prepared for institutions of learning that prefer to find in a single book
an ecological as well as morphological survey of the animal world. I2mo.
Cloth, $1.25 net.

Animal Life.

An elementary account of animal ecology — that is, of the relations
of animals to their surroundings. It treats of animals from the stand-
point of the observer, and shows why the present conditions and habits
of animal life are as we find them. I2mo. Cloth, $1.20 net.

Animal Forms.

This book deals in an elementary way with animal morphology. It
describes the structure and life processes of animals, from the lowest
creations to the highest and most complex, ramo. Cloth, $1.10 net.

Animals.

This consists of " Animal Life " and " Animal Forms " bound in one
volume. I2mo. Cloth, $1.80.

Animal Structures.

A laboratory guide in the teaching of elementary zoology. I2mo.
Cloth, 50 cents net.

D. APPLETON AND COMPANY,

NEW YORK. BOSTON. CHICAGO. LONDON.

SCIENCE, RELIGION, EDUCATION.

Adolescence : Its Psychology and Its Relations
to Physiology, Anthropology, Sociology, Sex,
Crime, Religion, and Education.

By G. STANLEY HALL, Ph.D., LL.D. Two vols., royal
8vo, gilt top. Cloth, $7.50 net.

This work is the result of many years of study and teaching. It is
the first attempt in any language to bring together all the best that has
been ascertained about the critical period of life which begins with
puberty in the early teens and ends with maturity in the middle twenties,
and it is made by the one man whose experience and ability pre-emi-
nently qualify him for such a task. The work includes a summary of
the author's conclusions after twenty-five years of teaching and study
upon some of the most important themes in ' Philosophy, Psychology,
Religion, and Education.

The nature of the adolescent period is the best guide to education
from the upper grades of the grammar school through the high school
and college. Throughout, the statement of scientific facts is followed
systematically by a consideration of their application to education, pe-
nology, and other phases of life.

Juvenile diseases and crime have each special chapters. The changes
of each sense during this period are taken up. The study of normal
psychic life is introduced by a chapter describing both typical and excep-
tional adolescents, drawn from biography, literature, lives of the saints,
and other sources.

The practical applications of some of the conclusions of the scientific
part are found in separate chapters on the education of girls, coeduca-
tion and its relations to marriage, fecundity, and family life, as seen by
statistics in American colleges, with a sketch of an ideal education for
girls.

Another chapter treats with some detail and criticism the various
kinds and types of organization for adolescents from plays and games to
the Y. M. C. A., Epworth League, and other associations devised for
the young.

The problem of the High School, its chief topics and methods, is
considered from the standpoint of adolescence, and some very important
modifications are urged. It closes with the general consideration of the
relations of a higher to a lower civilization from this standpoint

D. APPLETON AND COMPANY, NEW YORK.