BREEDING

AND THE

MEN DELIAN DISCOVERY

A. D. DARBISHIRE.

22102195726

Med

K2549

■l|

\

A

. f ■

f

: »* *

■ i

.. ■

m

\

A'

r

'.1

I

• /

y .

' r

DIAGRAM TO ILLUSTRATE THE MENDELIAN INHERITANCE OF TALLNESS AND DWARFNESS IN THE CULINARY PEA

BREEDING

AND

THE MENDELIAN DISCOVERY

BY

A. D. DARBISHIRE, M.A.

Demoattrator of Zoology, and Lecturer on Genetica at the University of
Edinburgh, and late Demonstrator of Zoology and Lecturer
on Heredity at the Imperial College of Science
and Technology, South Keusington

WITH ILLUSTRATIONS IN COLOUR AND
BLACK-AND-WHITE

SECOND EDITION

CASSELL AND COMPANY, LTD.
London, New York, Toronto and Melbourne
1912

First published September 1911,
Second Edition January igi2.

WELLCOME INSTITUTE
LIBRARY

Co(L

weilVIOmec

Cali

No.

Q[\

AIL RIGHTS RESERVEP

Preface

This book deals, as its title indicates, with the
facts discovered by Mendel, and with the bearing
of these facts, and of the theory put forward to ex-
plain them, on the science of heredity and the practice
of breeding. It is intended to serve merely as an
introduction to the subject. My conception of the
most serviceable form of such an introduction has
not been to place before the reader a sketch which
takes in the whole range of discovery and speculation
in this sphere of inquiry, but, rather, to open the
door to an intimate familiarity with a few instances
of the Mendelian phenomenon, and especially with
those studied by Mendel himself. To this end I have
given a fuller account of the phenomena observed
by Mendel than has yet appeared in popular form :
the seven pairs of characters studied by him are all
figured for the first time ; and other results of his
are illustrated by photographs from specimens which
I have bred myself. But, in case the reader takes
up the position, with which I heartily sympathise,
of refusing to be satisfied with anything short of
actually seeing the things which Mendel discovered,
I have given full instructions as to how such an

IV

PREFACE

experiment as Mendel’s is carried out, together with
the addresses of the seedsmen from whom peas
bearing one or more of the fourteen characters studied
by Mendel can be procured.

But although I regard a thorough knowledge of
the Mendelian discovery as the proper foundation on
which the attempt to breed, in the light of it, should
be based, I have not limited myself to an attempt
to impart this knowledge. I have also indicated
the more important lessons which the practical
breeder can learn from this discovery, and dealt with
some of the more interesting biological questions
which it raises, or helps to answer.

My endeavour, throughout, has been to eliminate
from my picture of the things described everything
with which these things have been invested by the
human imagination ; in other words, to make my
picture not the finished work of the post-impressionist,
but the untouched proof of the photographer. And
I hope that, in this way, I may do something, in my
own sphere of interest, to counteract the prevalent
tendency of that which is written, on any topic, to
bear but a remote relation to what can actually be
observed. This remoteness of words from actuality
is especially characteristic of much of the literature
on the subject of heredity which has accumulated
during recent years. The rational imagination, in its

PREFACE

V

flights in this sphere, has left the ground of facts
below, and attained to altitudes which put the feats
of airmen in the shade. I am not arguing that it
is desirable that those writers to whom the ground
is distasteful should cease tb fly, but merely pointing
out that, inasmuch as so much energy is at present
devoted to enterprise of this kind, there is room for
a book that will enable anyone to look closely at
some of the facts which, though they constitute the
ground over which a great deal of recent discussion
has raged, are seldom known at first hand by those
who witness, or even in some cases by those who
take part in, these battles of the air.

My thanks are due, and are readily tendered to
Professor de Vries for having taken the photographs
which constitute Figs. 1 to 5 inclusive expressly for
this book ; and to the trustees of the British Museum
for permission to photograph the case of my mice
which is reproduced in Plate II.

I should claim more credit than is properly mine
if I did not acknowledge the help which I have received
in the actual carrying out of the breeding experi-
ments to which such value as this book may possess
is in great part due. In connection with the breed-
ing experiments with mice, now concluded, my
thanks are especially due to Mr. Charles Biddolph

VI

PREFACE

for much help which he very generously gave me.
The task of harvesting and recording the peas, on
the scale on which the experiment is now carried on,
is made possible, and pleasant, by the assistance of
Mr. Frank Sherlock ; whilst the fullness of the harvest
itself is due to the energy and skill of my gardener,
Mr. George Reeves,

NOTE TO SECOND EDITION

The issue of a Second Edition gives me oppor-
tunity to correct a few mistakes due to my own
inadvertence. For discovering and pointing out to
me these mistakes I am especially grateful to
my cousin, Dr. 0. V. Darbishire, and to my
friend Mr. C. F. Meade.

A. D. D.

©0 ploba {tt stufible, for note is tfie Sfason
for soloing of fitcfits, of teanes, anb of peason
(Sobor runcibals timelie, anti all that be gras,
but soloe not tbe tobite till S. ffiregeries bag,

^olo peason anb beanes in tbe loane of tbe moone
tobo sotoetb them sooner, be sotoetb too soone
®bat tbeg toitb tbe planet mag rest anb arise,
anb flourish toitb bearing most plentifulltoise.

ISotb peason anb beanes some afore se boo plots,
tbe sooner ge baeroto, tbe better for goto.

White peason so goob for the purse anb the pot:
let tbrm be toell useb else toell boo ge not.

cSticfe plentg of botos among runeibal pease
to elimber tbereon, anb to branch at their ease,
<So booing more tenber anb greater tbeg toei,
if peaebocb anb turbeg leabe jobbing their bei.

From Thomas Timer's “Five Hundred Points of Good Husbandbie."
[Edition of 1680 collated with those of 1573 and 1577.]

I

Contents

CHAPTER PAGE

1. Breeding and Heredity ..... 1

2. The Mendelian Phenomena of Inheritance ; the

Result of Crossing Tall with Dwarf Peas . 12

3. The Inheritance of Four other Characters of the

Culinary Pea studied by Mendel . . .21

4. General Characteristics of the Mendelian Phe-

nomena Illustrated by the Case of the An-
dalusian Fowl and the Inheritance of Eye-

Colour IN Man 33

6. The Inheritance of the Characters of the Seed

IN the Culinary Pea 50

6. Throwing Back, or Reversion .... 72

7. The Relation between Characters belonging to

Distinct Pairs 93

8. The Interpretation of Reversion . . .114

9. The Nature of the Mendelian Pair of Characters 124

10. The carrying out of a Breeding Experiment with

A Plant 141

11. Mendel’s Explanation of His Results. . .160

12. Antecedents and Fate of Mendel’s Discovery . 186

13. Theory to Account for the Results which follow

A Cross involving Two Pairs of Characters . 203

X

CONTENTS

CHAPTER PAOB

14. The Theory of Keversion 220

15. Some New Principles of Breeding based on

Mendel’s Theory 233

16. The Inheritance of Sex 247

17. The Origin of the Mendelian Mode of Inheritance 267

Glossary . . 275

Index 277

Colour Plates

PLATE FACING PAO»

1. MENDELIAN INHERITANCE OF THE COLOUR OF THE SEED IN

THE CULINARY PEA 58

II. MENDELIAN INHERITANCE IN MICE 74

III. THE RESULT OF CROSSING A YELLOW WRINKLED WITH A

GREEN ROUND PEA 94

IV. MENDELIAN INHERITANCE OF THE COLOUR OF THE FLOWER

IN THE CULINARY PEA . e . . . r .118

List of Illustrations

DIAGRAM TO ILLUSTRATE THE MENDEUAN INHERITANCE OP

TALLNESS AND DWARFNESS IN THE CUUNARY PEA Frontispiece

FIGURE

1. (ENOTHERA LAMARCKIANA

FACING PAGE

4

2. (ENOTHERA GIGAS

3. (ENOTHERA LAMARCKIANA ; (E. GIGAS

4. (ENOTHERA RUBRINERVIS .

6. 02N0THERA NANELLA

6. SEEDLING OF TALL PEA ; SEEDLING OF DWARF PEA

7. TOP OF NORMAL STEM OF THE CULINARY PEA .

8. TOP OP FASCIATED STEM OF CULINARY PEA

9. MENDELIAN CHARACTERS IN THE TEXTURE OF THE POD

10. MENDELIAN CHARACTERS IN THE COLOUR OF THE POD IN THE

CULINARY PEA

11. “grey” pea ....

12. “ WHITE ” PEA

13. ANDALUSIAN FOWLS

14. DIAGRAM TO ILLUSTRATE THE MENDELIAN INHERITANCE OF

COLOUR IN THE ANDALUSIAN FOWL

15. VERTICAL SECTION OF HUMAN EYE OF SIMPLEX TYPE .

16. SECTION OF IRIS OF SIMPLEX EYE ....

17. SECTION OF IRIS OF DUPLEX EYE ....

18. THE SEVEN SEEDS WHICH GAVE RISE TO THE SEVEN SEEDLINGS

IN FIG. 19, IN THEIR NATURAL POSITIONS IN THE POD.

19. THE SEVEN SEEDLINGS RAISED FROM THE SEVEN SEEDS SHOWN

IN THE POD IN FIG. 18

20. DIAGRAM TO ILLUSTRATE THE DISTRIBUTION OF YELLOW AND

GREEN SEEDS IN PLANT NO. 12 IN TABLE ON PAGE 62 .

6

6

8

10

12

20

22

24

26
28
‘ 30
32

36

40

42

42

60

60

64

xii LIST OF ILLUSTRATIONS

FIGURE FACING

21. MENDELIAN INHERITANCE OP THE SHAPE OP THE SEED IN THE

CULINARY PEA .... ....

22. SINGLE COMB

23. ROSE COMB

24. PEA COMB

25. WALNUT COMB. .

26. THE RESULT OP CROSSING MAPLE AND PURPLE-SPOTTED SEED

COAT IN PISUM

27. THE RESULT OP CROSSING GREY WITH WHITE SEED-COAT IN

PISUM

28. STARCH GRAINS OP ROUND PEA

29. STARCH GRAINS OP WRINKLED PEA

30. STARCH GRAINS OP A CROSS BETWEEN A ROUND AND A WRINKLED

PEA

31. STAGE IN THE DEVELOPMENT OP A PLOWER WHEN PERTILISA-

TION OCCURS Page

32. POLLEN GRAINS OP THE CULINARY PEA (MAGNIPIED ABOUT 50

times) ; POLLEN GRAINS OP THE CULINARY PEA (MAGNIFIED
ABOUT 160 TIMES) .......

33. TO ILLUSTRATE METHOD OP LABELLING MALE AND FEMALE

PARENTS OF CROSS .......

34. JOHN GOSS’s GARDEN ; WHERE HE MADE HIS CROSSES IN 1820

PA.OB

68

104

104

106

106

no

114

128

128

134

148

152

154

194

BREEDING:

AND

THE MENDELIAN DISCOVERY
CHAPTER I

BREEDING AND HEREDITY

The practical breeder and the student of heredity
both deal with the same natural process, but for very
different reasons. The breeder’s object is to main-
tain or increase the value of the animals and plants
which minister to the needs or please the fancy of
mankind. The student of heredity aims at finding
out how the characters of animals and plants are
handed on from generation to generation. The
breeder is not concerned with the interpretation of
what he achieves ; he does not care how the charac-
ters of his stock are handed on from generation to
generation, so long as the changes which he effects
tend in the direction of an improvement. The stu-
dent of heredity, on the other hand, is not concerned
with the application of what he finds out ; he does
not care whether the characters of his material are
tending in the direction of an improvement, so long
as he finds out how they are handed on from genera-
tion to generation. But there cannot, I think, be
any doubt that both the practical breeder and the

B

2

BREEDING

student of heredity would profit much by a know-
ledge of the methods employed and the results ob-
tained by the other. It is a question which of the
two would gain most : the man of science by the
knowledge of the facts relating both to the successes
and the failures of the breeder ; or the breeder by
an acquaintance with the principles elucidated, and
the precision of the records kept and the methods
employed by the man of science. It is desirable
that the student of heredity should, for the sake of
his science, become interested in breeding. It is
desirable that the breeder should, for the sake of his
art, become interested in heredity. And, in general,
it is desirable that each should recognise that he has
much to learn from the other ; for though, as we
have seen, the objects of their work are different,
the problem which they are both investigating is the
same.

My object in writing this book is to place within
reach of the practical breeder some of the things
found out by the scientific student of heredity, which
are likely to be of service to him.

The methods of the breeder are two. They are,
as a matter of fact, nearly always used in conjunc-
tion, and not separately ; but they are neverthe-
less essentially distinct. They are selection and
crossing. Pure selection, operating on material which
is not the immediate result of a cross, modifies the
form of an animal or plant, and leaves it different
from what it was when the selection began ; some-
thing has been added or taken away ; something

BREEDING AND HEREDITY 3

created or destroyed. The primary object of cross-
ing is to combine within one strain two desirable
qualities existing in distinct strains. But an appar-
ent result of crossing is very often a much more rapid
production of something new than is brought about
by selection {see Plate II.). The novelty in such a
case, however, is an illusion, and due to the recombina-
tion of characters which have been long hidden and
long separated. Selection, then, creates something
new ; crossing merely recombines characters which
already exist. Very little — it may almost be said
that nothing — is known of the causes which determine
the origin of new characters, either in the domes-
ticated state or in a state of nature. At any rate,'
there is not enough known to base a scientific prac-
tice of breeding upon. But the art of breeding by
the recombination of already existing characters —
and it is a question whether in its ultimate analysis
this may not be all that the breeder can do — has been
brought to a point of extreme precision by the dis-
covery made by Mendel in 1865. And the object
of this book is to make perfectly clear what that
discovery was, and how it affects the practice of
breeding. The manner of origin of new forms will
be dealt with because a knowledge of Prof, de Vries’
observations on this process is necessary to an appre-
ciation of the significance of Mendel’s discovery.

Ever since 1859, when Darwin effected the general
acceptance of the doctrine of evolution, it has been
believed that new forms have originated by the
accumulation of the almost imperceptible differ-

4

BREEDING

ences between individuals, whether this has been
brought about by the stress of circumstance or the
skill of the breeder. In other words, the material
upon which selection, both natural and artificial,
has been supposed to operate has been assumed to
be these imperceptible differences, the occurrence of
which is referred to as the phenomenon of continuous
variation.

This belief was attacked by Mr. Bateson in his
“ Materials for the Study of Variation ” in 1894, and
was severely shaken by the publication of Prof, de
Vries’ “ Mutationstheorie ” in 1901. The belief em-
bodied in both these works is that evolution has
not proceeded by the accumulation of the differ-
ences presented by continuous variation, but by the
differences which are furnished from time to time by
what is called “ discontinuous variation ” by Mr.
Bateson, and “ mutation ” by Prof, de Vries.

The point at issue between the Selectionist, who
holds that evolution has been brought about by
the selection of continuous variations, and the Muta-
tionist, is this. It is admitted that species, as we
see them, constitute a discontinuous series ; they
are, with rare exceptions, distinct from and do not
merge into the species most closely allied to them.
The question is whether, when one species arises
from another, the new species arises at one bound,
i.e. in one generation, or whether it only comes
into existence after all the imperceptibly small
gradations, between it and the old species from
which it sprang, have been traversed over the

FIG. l.—(ENOTHERA LAMARCKIANA

BREEDING AND HEREDITY 5

course of many generations. According to the
former view, species cannot be said to arise by
selection. If, when they have once arisen, they
have defects which hamper them in the struggle for
existence, where this occurs, they will doubtless
drop out. But this does not constitute origin by
selection. According to the latter view, on the
other hand, the character of the new species is only
attained by the traversing of the gradational stages
between it and the parent one ; and inasmuch as it
may be supposed that these gradational forms would
not have left offspring if they had any defects which
prevented them doing so, it may be said that the
new species arises by a process of natural selection.

The view that species have originated by muta-
tion is based on Prof, de Vries’ observations on the
Evening Primrose {CEnothera Lamarchiana) (Fig. .1).
Working with this form, he was able to witness, for
the first time, the actual process of the origin of new
species.

For some time he had been searching in the neigh-
bourhood of’ Amsterdam for plants which were
giving off new forms ; but as he failed to find any
he concluded that they were all in an immutable
condition. At last he found a plant which appeared
to be in a mutable state. This was CEnothera
Lamarchiana, It was growing in a disused potato
field near Hilversum, in Holland. This plant had
escaped into the field from a bed in a park close by,
where some annuals were grown every year. The
CEnotheras were thickest at the corner of the field

6

BREEDING

where they had invaded it ; and at this spot they
formed a dense thicket of vigorous plants as high as
a man. From this point outwards the plants became
sparser, and the extreme edge of the area which they
covered was occupied by an advance guard of seed-
lingSt and young rosettes.

Two new species were found in the field — (Enothera
IcBvifolia, characterised by its smooth, as opposed
to crumpled, leaves and by its narrow petals, which
do not overlap one another as do those of 0. Lamarck-
iana ; and 0. brevistylis, characterised by the short-
ness of its style, which is so short that, when the
flower is first looked into, it looks as if there were no
style or stigma.

Seed was collected from the plants in the potato
field at Hilversum and sown in the Botanic Garden
at Amsterdam ; and rosettes — i.e. plants in their
first year, which have not yet developed a stem —
were also transplanted there. In this way there
arose a series of pedigree cultures in which a number
of other new species originated. Of these I will
only refer to those which are figured in this book.

(Enothera gigas is the rarest and the finest of the
new species which has arisen in the Garden at
Amsterdam. It has only appeared three times ; and
it is in every respect more vigorous and more heavily
built than its parent 0. Lamarckiana,

The flowering spike is shown in Fig. 2 ; and its
fruits, alongside those of 0. Lamarckiana, in Fig. 3.
The following characteristics of 0. gigas can be made
out from these figures. Its leaves are broader ; its

Fig. 2—(EN0THERA GIGAS

fig. 3.—(EN0THERA
LAMARCKIANA

(E. GIGAS

BREEDING AND HEREDITY 7

fruits shorter but stouter ; its stem thicker ; and its
petals broader, and consequently its buds fatter, than
those of the parent species. The top of the stem is
densely clothed with foliage ; and its appearance
forms a striking contrast to the naked look of the
top of the stem of 0. Lamarchiana, This feature
is due to the greater breadth of the leaves, to the
fact that they are bent downwards, and to the short-
ness of the internodes, especially at the extreme
top of the stem {see Fig. 3) in 0. gigas.

Its first appearance was as follows. In 1895
there was a crop of about 14,000 plants, constituting
the fourth generation of the Lamarekiana culture.
All the mutational forms had been transplanted
from the crop, and most of the plants of 0. Lamarck^
iana had been thrown away. In the beginning of
August about 1,000 of these plants were in flower,
but many were still in the rosette stage. Thirty-two
of the strongest of these rosettes were selected and
planted out ; they flowered in the July and August
of the following year. One of them attracted Prof, de
Vries’ attention by its thick stem, and by its large
cup-shaped flowers. On August 10 all of its flowers
were picked off, both the open and the withered
ones, and the whole top of the plant enclosed in a
paper bag to ensure self-fertilisation. It set a quan-
tity of good seed. This plant was the parent of the
new species 0. gigas. Its ancestors, for at least three
generations back, were ordinary 0. Lamarekiana,
The numbers of seed-bearing plants in these three
generations were successively only nine, six, and ten,

8

BREEDING

and the characters of 0. gigas are so striking that if
one of these twenty-five plants had been a gigas it
could not have been overlooked, as the seed-bearing
plants were under continual and close observation.

The self-fertilised seeds of the original gigas were
sown in 1897. They produced somewhat over 450
plants. ' All of them were exactly like their parent,
with one exception, to be mentioned later ; and they
constitute a type distinct, from the very outset,
from 0. Lamarckiana, They bred true on their first
appearance, and have continued to do so ever
since.

(Enoihera ruhrinervis, which, unlike 0. gigas, is
one of the commonest of the new species which
arose from 0. Lamarckiana, is characterised, as its
name implies, by the possession of bright red veins
on the leaves and broad red stripes on the calyx and
fruit. The flowering spike is shown in Fig. 4. The
leaves are narrower than those of Lamarckiana ;
but perhaps the most remarkable feature of this
new species is the brittleness of the stem. If the stem
of the flowering plant is hit, it breaks in two or three
places with perfectly smooth surfaces of fracture.
The cause of this brittleness is the extremely slight
development of the bast fibres in the vascular
bundles, in the stem. In all other respects ruhrinervis
is a healthy species, and it is the only one of the new
species which produces as much pollen and seed as
the parent species.

0. ruhrinervis arose from 0. Lamarckiana sixty-
six times in all, and not only did it always possess

/ Ah

FIG. 4.—(EN0THERA RUBRINERVIS

(Slightly Enlarged.)

V.

■ •V^ -^» * ^ ' ' , ' • ‘f • sj

r-- .'-r

• . .1 V

i •

1

■ ’ i** • ?5I

^ /*

6 • >^
i .I

//*

: •^M

BREEDING AND HEREDITY 9

the same set of characters — the red colouring, the
brittleness, the narrow leaves — but none of these
characters ever appeared separately. It also, like
0. gigas, produces offspring, all of which are indis-
tinguishable from itself.

CEnothera nanella, — This is the dwarf CEnothem,
A whole plant is shown in Fig. 5. It does not
attain to half the stature of the parent form. It
arose 158 times during the whole course of the
experiments, during which some 50,000 plants were
raised, i.e. it appeared in a ratio of about 3 per
cent. It also, like the other two new species already
mentioned, came perfectly true to seed directly it
had arisen.

0, nanella is distinguished from the other new
species which I have described, first in the fact that
it differs from its parent species by only one salient
feature — dwarfness ; and secondly in the fact that this
characteristic is commonly found in other genera of
plants, as for instance the pea (see Fig. 6), to mention
one of a large number of plants which present this
peculiarity. Professor de Vries has proposed to give
the name “ varietal ” to characters like dwarfness,
which are widely distributed throughout the vege-
table kingdom, to distinguish them from specific
characters which, according to him, do not recur in
this way. Varieties art forms characterised by the
possession of varietal characters, and species are
forms characterised by specific characters thus
defined. For instance, the complex of characters
possessed by 0. gigas or by 0. ruhrinervis do not occur

10

BREEDING

elsewhere in the vegetable kingdom, and these two
forms are regarded as true species ; whilst the cha-
racter of 0, nanella, which is regarded as a variety,
not only occurs in the pea and in other plants, but
also amongst the new forms which have arisen from
0. Lamarckiana, In fact the single exception referred
to amongst the 450 plants raised from the first 0.
gigas was a plant which had all the characters of
0. gigas, but possessed the dwarf habit of 0. nandla,
and should be designated 0. gigas-nanella.

Professor de Vries has drawn this further ex-
tremely important distinction between specific and
varietal characters. Specific characters, according
to him, blend in inheritance; whilst varietal ones
behave in Mendelian fashion when crossed. Indeed,
he makes the behaviour of a particular character
when crossed the test whether it is a varietal or a
specific one. Professor de Vries has made a number
of crosses between 0. gigas and other species of
(Enotliera, For instance, 0. gigas, crossed with 0.
Lamarckiana, gives rise to forms which are a blend
between their two parents, and moreover breed true
to this intermediate character. The conclusion to
which he comes in the paper in which he describes
these results is stated as follows by him : “In all
these respects, and in others as well, 0. gigas
behaves as a true species when crossed, and not as
a variety, as 0. nanella does when crossed with the
same forms.” There are certain complications in
the case of 0. nanella, but these need not detain us
now. Full details can be found in the paper referred

FIG. 5.—(EN0THERA NANELLA

BREEDING AND HEREDITY ii

to below * by those who wish to follow the matter
further. Having thus indicated the mode of origin
and possible nature of characters which are inherited
in Mendelian fashion, we may now proceed to a
consideration of the Mendelian phenomenon itself.

♦ “ Berichte der Deutschen Botanischen Gesellschaft.” Jahrgang 1908.
Band XXXU, Heft 10.

CHAPTER II

THE MENDELIAN PHENOMENA OF INHERITANCE;

THE RESULT OF CROSSING TALL WITH DWARF PEAS

Mendel published his results in the journal of
a local scientific society in Briinn, in Austria,
in 1865; but his paper remained unnoticed* until
the year 1900, when attention was directed to it
almost simultaneously by Correns, Tschermak and
de Vries, the former two of whom had repeated
some of Mendel’s experiments, and were thus able
to attest the accuracy of his observations. I shall
deal in Chapter XII. with the circumstances in
which Mendel’s discovery was made, and with the
probable causes which led to its being forgotten for
so long ; and shall pass on now to a description of the
Mendelian phenomena themselves.

The reader is probably familiar with two kinds
of the edible pea (Pisum sativum)^ the tall and the
dwarf. The tall peas grow to any height from three
to six, or even eight, feet, and need to be supported
with sticks ; the dwarf peas do not rise more than
a foot or two above the level of the ground, and
need not be staked.

The difference between a tall and a dwarf pea is

♦With the single exception of a reference in Focke’s Pflanzenmisch-
linge^ 1881.

12

Fig. 6.— seedling of tall pea seedling of dwarf pea

The seeds were sown at the same time.

THE MENDELIAN PHENOMENA 13

not the same as the difference between a tall and a
dwarf man. In a human dwarf everything is on a
smaller scale than in the normal man. But a dwarf
pea is not simply a miniature edition, as it were, of
a tall one : it differs from a tall pea in one single
characteristic, the length of the internodes, i.e. the
sections of the stem between two nodes, or joints,
where the leaves are given off. This is shown in
Fig. 6, which represents a seedling of a tall, and a
seedling of a dwarf pea sown on the same day. The
number of nodes between the seed and the point
of insertion of the two leaves which are face to face
and enclose the growing point of the stem, is the
same in both cases, namely five. They can all be
clearly seen in the tall seedling, but the lowest one
in the dwarf is difficult to make out on account of the
shadow. The shortness of the dwarf pea is thus seen
to be due solely to the shortness of its internodes.

If seed is saved from a dwarf plant it will give
rise, when sown, to dwarf plants only ; and seed
saved from tall plants will only produce tall plants.
In other words, both tall and dwarf peas, when left
to themselves, breed true.

It is desirable to state here that the pea is normally
self-fertilised. The male and female organs and pro-
ducts are both present in every flower ; and what
happens when a plant is “ left to itself ” is that the
female products (or egg-cells) of a single flower are
fertilised by the male products (the pollen grains) of
the same flower. When one kind, such as tall, is
crossed with another kind, such as dwarf, the pollen

14

BREEDING

grains are transferred from the flowers of a tall
plant to a particular part of the female organ of a
dwarf plant, or vice versa. The details of this
process will be described in Chapter X. The main
point which I wish to be noted here is that the
pea is always self-fertilised unless it is artificially
cross-fertilised by the agency of man.

We have seen that dwarf plants, when left to
themselves, set seed which produces dwarf plants
only ; and that tails similarly breed true.

We will now consider the result of crossing a tall
with a dwarf.

The plants which result from a cross between a
tall and a dwarf are always tall. These tails are no
shorter than their pure tall parents ; in fact, as a
rule, they are a little taller, because, like most hybrids,
they are very vigorous. They constitute the first
hybrid generation.

These hybrid tails are allowed to self-fertilise. If
the seed saved from them is sown, about three-
quarters of the plants produced are tall, and about
a quarter are dwarf. That is to say, in every four
plants, three are tall and one is dwarf, on the average.
This average result is shown in the genealogical
table which forms the frontispiece, where it will also
be seen that this generation — namely, the one which
consists of three tails and one dwarf in every four —
is called the second hybrid generation.

Let us now consider the result of sowing the seed
saved from the plants composing the second hybrid
generation. These plants, like those of the first

THE MENDELIAN PHENOMENA 15

hybrid generation, are allowed to self-fertilise. The
seed set by the dwarfs produces only dwarfs. This
is what we should expect, if we look at the tails and
dwarfs in the genealogical table from the point of
view of their stability in breeding, that is to say
according as to whether they produce their own kind
only, or both kinds, namely tall and dwarf. If we
look back through the pedigree of the dwarfs of the
second hybrid generation, we do not come to a dwarf
until we reach the pure dwarf parent of the cross.
This pure dwarf, as we know, breeds true. We
might, therefore, expect that the dwarfs will breed
true when they reappear in the second hybrid genera-
tion. And we see that they do.

Before we consider the offspring of the tails in
the second hybrid generation, let us look back in
their pedigree and see if we can form any idea as to
what their offspring will be. In the case of the
dwarfs of the second hybrid generation we had to go
back two generations, namely, to the parents of the
cross, before we met with an ancestor bearing the
dwarf character. But in the case of the tails we have
only to go back one generation — to the first hybrid
generation — to find a tall. But this hybrid, although
it resembles the pure tall externally, differs from it
in its breeding properties. The pure tall produces
only tails ; from the point of view of its breeding
properties, it is stable. But the hybrid tall pro-
duces both tails and dwarfs, in the proportion
of three tails to one dwarf in every four ; it is
unstable. That is to say, if we look back through

i6

BREEDING

the ancestry of the second hybrid generation, we find
only one kind of dwarf — a pure, true-breeding or
stable kind ; but we find two kinds of tails, a pure
stable kind which produces tall offspring only, and
a hybrid or unstable one which produces both tall
and dwarf.

What, then, is the nature of the three tails of
the second hybrid generation ? There are three
possibilities. They may be all pure ; or all hybrid ; or,
lastly, some of them may be hybrid and some pure.
This latter is actually the case. Two of the three
are hybrid, and produce tails and dwarfs in the
ratio of three tall to one dwarf in every four. The
remaining tall is pure, and produces only tails.

In the genealogical table we have supposed that
each plant of the second hybrid generation produces
only four offspring, because this enables us to put
the average proportions of tails and dwarfs amongst
these offspring most conveniently in diagrammatic
form. The pure tall which produces only tails is
seen at the extreme left of the second hybrid genera-
tion. In the middle are the two hybrid tails, each
producing three tails and one dwarf on the average ;
and at the extreme right is the dwarf, which produces
only dwarfs.

Let us now glance at the actual proportions
observed by Mendel himself. He does not say how
many plants of the first hybrid generation he raised ;
he only states that he made thirty-seven crosses
between tall and dwarf. We do not, therefore, know
how many of these were successful. His second

li THE MENDELIAN PHENOMENA 17

hybrid generation consisted of 1,064 plants, of which
I 787 were tall, and 277 were dwarf, a very close
approximation to a ratio of 3 to 1.

Ten seeds from each plant of a hundred only
I of these 787 tails were sown. Twenty-eight of

I these produced tails ; seventy-two produced both

tails and dwarfs. This result departs consider-
ably from the expected one. Let us examine
it more closely. Of the three tails in the second
j hybrid generation, two should produce tails and

j dwarfs in the proportion of 3 to 1 ; one should

produce tails only. This is one way of expressing
I the average result. Another way is to say that

two-thirds of the tails in the second hybrid generation
i produce tails and dwarfs in the ratio of 3 to 1 ; and
one-third tails only. Of a hundred tails in the second
hybrid generation we should, therefore, expect sixty-
six or sixty-seven to produce tails and dwarfs in the
ratio of 3 to 1, and thirty-four or thirty-three to pro-
Ip duce only tails, if the general statement I have made as

I to these ratios is true. It will, therefore, be seen that

amongst the hundred tails of the second hybrid
I generation which Mendel tested, the number of pure
tails was too low (twenty-eight as against thirty-four
j or thirty-three), and the number of hybrid tails too
high (seventy-two as against sixty-six or sixty-seven).
This brings us up against a very important question,
namely, how close a fit between the so-called Men-
delian expectation and the actual result must we
I expect ? The general rule is that the greater the

I numbers of individuals dealt with, in the experiment,

c

i8

BREEDING

the closer will the actual result approximate to the
Mendelian expectation. This is a very important
question, which will be dealt with later.

In the course of the description of the Mendelian
phenomenon, which is diagrammatically represented
in the frontispiece, it was said that certain of the
results were what we should have expected them to
be if we looked at the whole pedigree from a certain
point of view, namely, that of stability in breeding.
They are what they would be expected to be by anyone
who had no knowledge of the Mendelian phenomenon,
and who looked at them from the point of view of
stability. For instance, the true breeding of the
dwarfs in the second hybrid generation was in
harmony with the fact that only one kind of dwarf,
a pure breeding one, occurred in the ancestry of
the cross. And, similarly, the fact that there are
two kinds of tails, a hybrid and a pure, in the second
hybrid generation, is in harmony with the fact that
there are two kinds of tails, a hybrid and a pure,
in this ancestry.

But, beyond this, most of the results we have
described differ widely from what current notions of
heredity would lead us to expect them to be. In
the first place the general expectation, as to the
result of crossing a tall with a dwarf would probably
have been that the resultant hybrid would have
been intermediate in height between its two parents.
But the individuals composing the first hybrid
generation are all tall, as we have already seen.
Intermediates between tall and dwarf have never been

THE MENDELIAN PHENOMENA 19

observed to occur in the first hybrid generation, nor
do they occur in any subsequent generation derived
from the cross. It is perhaps not surprising that
the dwarf which disappears altogether in the first
hybrid generation should reappear in the second.
And although we have said that the true breeding
of the dwarf when it does reappear is roughly what
we should expect from one point of view, there is a
point of view from which it is the very opposite of
what we should expect. The fact of the dwarfs in the
second hybrid generation breeding true means that
they are entirely free from the taint of tallness in spite
of the fact that both of their parents and half of the
rest of their ancestors were tall. A dwarf occurring in
the second hybrid generation — an ‘‘ extracted ” dwarf
as it is called — is as pure a dwarf as the pure parental
dwarf with which the cross was made, not only in
its external characteristics, but also, what is more
remarkable, in its breeding properties. The true
breeding of one-third of the tails in the second hybrid
generation is not quite such a remarkable fact, because
in this case their parents at least bore the tall
character ; but it is remarkable that these tails
should exhibit no trace of the dwarf character of
half the grand-parental generation and of half the
ancestry behind that point.

When we come to consider the fourth and further
hybrid generations we shall find that the results come
to differ still more widely from what current
notions of heredity would lead us to expect them
to be.

20

BREEDING

Suffice it for the present that the actual results,
so far as we have considered them, of crossing a
tall with a dwarf pea stand out in unexpected sim-
plicity and orderliness against the vague and, as we
have seen, often contradictory expectations based
on the notions of heredity which were current when
Mendel’s results were discovered.

Fig. 7.— top of NORMAL STEM OF THE CULINARY PEA

CHAPTER m

THE INHERITANCE OF FOUR OTHER CHARACTERS OF
THE CULINARY PEA STUDIED BY MENDEL

Mendel also made crosses between peas which
differed from one another in other characters than
those of tallness or dwarfness.

Another character with which he dealt was the
position of the flowers on the stem. These either
arise, in the ordinary way, from the whole length of
the stem, above the point, on the stem, where they
begin; or they are all bunched together at the top.
Fig. 7 shows the former, ordinary, arrangement of
the flowers, which is characteristic of all the varieties
of peas commonly grown for the kitchen. Fig. 8
shows the latter, in which they are bunched together
at the top. The photographs were taken late in the
year, so that pods are shown instead of flowers.

I have described these two types of pea in terms
of the position of their flowers. But the difference
between the two lies deeper than this. In a normal
stem the flowers are given off at successive nodes
up the stem. The node at which the first one is
given off varies greatly in the different kinds of peas.
But the essential point is that, in the normal stem, the
nodes follow one another at fairly regular intervals ;

the actual distance between them depending usually

21

22

BREEDING

on whether the plant is tall or dwarf, that is to say,
on whether the internode is long or short.

If a normal growing stem is examined, from the
ground upwards, it will be found that the nodes
succeed one another in this regular way, and that
the arrangement of the flowers which arise from the
nodes is correspondingly regular.

But if a plant on which the flowers are bunched
together at the top is examined from the ground
upwards, it will be found that, about three quarters
of the way up the plant, there are two nodes where
there should be only one ; so that at this point there
are two leaves, and sometimes two flower-stalks,
where there should be only one of each. Beyond this
point, as we proceed upwards, abnormal nodes of
this kind follow one another in rapid succession until
we reach the top of the plant, from which the flower-
stalks arise in such profusion that they form a sort
of crown of flowerg. The name “ fasciated ” is given
to this kind of stem ; and in the pea, at any rate,
the fasciation seems to be due to the loss, by the
plant, of something which brings about the production
of nodes at regular intervals, as in the normal stem.

The fasciated pea was known in this country in
the sixteenth century, when it was described by
Gerarde in his “Herbal,” published in 1597.* He
calls it “ Pisum umbellatum (Tufted or Scottish
Pease),” and says that these peas “ carry their

* For directing my attention to this and further references to different
kinds of peas I am much indebted to a very interesting paper by Mr. N. N.
Sherwood, on “ Garden Peas ” in the Journal of the Royal Horticultural
Society Vol. XXII., Part 3, 1898.

Fig. 8.— top of FASGIATED STEM OF CULINARY PEA

Wjf* , ,

■: ■■ - “'fiVN*'- ' - ' “

■i

:• ^

>

ll

H er ^ X^.m

'■j .

— * 'sSksSt

i ,

- . '■ »

.&*’ .-'ti

^ '■ . ■'•» '
4^'-'

I

OTHER MENDELIAN CHARACTERS 23

fruit on the top of the branches, and they are
esteemed and taken for Scottish Peason, which is
not very common.” The fasciated variety now
exists under the name Mummy Pea. What this
name means I do not know. Many people imagine
that these peas are found in Egyptian tombs.

Mendel crossed a variety characterised by a
fasciated stem, with a variety the stem of which was
normal; and found that the stem of the resultant
hybrid was normal.* The second hybrid generation
consisted of 651 normal, and 207 fasciated plants.
Only a hundred of these 651 normal plants were
tested to see if they were pure or hybrid normals.
Ten seeds of each of the hundred were sown, as in
the case of the tall and dwarf character. The
offspring of thirty-three of the hundred were normal
without exception. Of the offspring of the remain-
ing sixty-seven, on the other hand, some were
normal and some fasciated in the case of every
plant — as close an approximation to the expected
Mendelian ratio as can be obtained. In other
words, of every three normals in the second hybrid
generation two on the average were hybrid and one
was pure.

We have so far dealt with characters of the stem
of the plant ; first, with the differences in the length
of the internodes which produce tall and dwarf stems,
and then with the differences in the arrangement of

* Particulars as to how to procure varieties which exhibit the characters
studied by Mendel are given on page 158.

24

BREEDING

the nodes which produce normal and fasciated stems.
We shall now deal with characters of the pods ; first
with differences in texture, and then with differences
in colour.

The third character, with which Mendel experi-
mented, which we shall consider is therefore the
shape or, more strictly, the texture of the pod. The
pods of all the varieties of culinary pea commonly
grown possess a tough, parchment-like lining, which
makes the immature pod inedible and the axis of
the ripe pod straight (in the sense that a horse’s
back is straight), and its surface even, as shown
in the right-hand pair of pods in Fig. 9. For
brevity in reference, this type of pod may be called
“ hard.” There is also a variety of pea, the pods of
which lack this tough lining, the absence of which
renders it possible for the immature pods to be eaten
— and very good they are. They are called sugar
peas in England and Germany, and the variety is
described as sans parchemin in France. The absence
of this parchment-like lining has a very noticeable
effect on the shape of the pod throughout its develop-
ment. The main axis of the pod, unsupported by
this tough tissue, becomes much bent and contorted,
and the two flanks of the pod collapse on to the
seeds, so that the contour of the seeds can be clearly
seen on the outside of the pod, and the surface of
the ripe pod is generally uneven, as is well shown
in the left-hand pair of pods in Fig. 9. We may
conveniently refer to this type of pod as
“ soft.”

Two “ Soft ” Pods. Two “ Hard ” Pods.

Fig. 9.-MENDELIAN CHARACTERS IN THE TEXTURE OF

THE POD

OTHER MENDELIAN CHARACTERS 25

The hard- and soft-podded varieties were also
known to Gerarde. He recognised ‘‘ Some with
tough skins or membranes on the cods, and others
have none at all, whose cods are to be eaten with
the Peason, when they are young, as those of Kidney
Beans.”

Mendel crossed these two forms and found
that the hybrid resulting from a cross between
a hard- and a soft-podded pea was hard-podded.
It has since been found that the pod of the hybrid
is never so hard as that of the hard-podded parent.
But this point requires further investigation, and
need not concern us here. The second hybrid
generation raised by Mendel from these hybrids
consisted of 882 plants with hard pods and 299
plants with soft pods. A hundred of the hard-
podded plants were tested by sowing ten seeds of
each of them. Twenty-nine of them produced hard-
podded plants only ; seventy- one produced families
in which both hard-podded and soft-podded plants
occurred — a tolerably close approximation to the
Mendelian expectation (twenty-nine instead of thirty-
three or thirty-four, and seventy-one instead of
sixty-seven or sixty-six).

The fourth characteristic of the pea with which
Mendel dealt was, as I have already said, the colour
of the unripe pod. This is, of course, usually green,
and, in ordinary kinds, the green only disappears
as the pod becomes dry and ripe. But there is
a variety of the culinary pea in which the pod goes

26

BREEDING

yellow before it ripens. This change in colour
is not the same as that which ripening pods undergo ;
or, if it is, it is the same change shifted back in
development, because it takes place long before the
pods begin to dry and ripen. This turning yellow,
whilst the plant is still juicy and growing, affects the
stem and, afterwards, the leaves, as well as the
pods ; and if seed of yellow-podded and ordinary
green-podded plants are sown together the numbers
of the two kinds of plants can be counted more
conveniently by the stems than by the pods, so
pronounced is the difference between the colour of
the green and that of the yellow stems. In Fig. 10
there is seen, to the left, part of a plant of the yellow-
podded variety ; and, to the right, a specimen of
the ordinary green-podded kind. In the yellow-
podded specimen, the pods themselves have gone
yellow, but the leaves are still green ; the contrast
between the pale colour of the pod and the deeper
colour of the leaf is clearly shown. The pod looks
as if it were made of yellow wax ; and the analogous
yellow-podded dwarf bean is sometimes called the
wax-pod bean.

Mendel crossed a green-podded with a yellow-
podded variety of pea, and found that the offspring
were green-podded. The second hybrid generation
consisted of 428 plants with green pods and 152 with
yellow ones. A hundred of the green-podded plants
were tested, and forty of them gave rise to green-
podded plants only, and sixty to both green- and
yellow-podded ones.

I

The Yellow-podded Variety. The ordinary Green-podded Variety.

Fig. 10.— MENDELIAN CHARACTERS IN THE COLOUR OF
THE POD IN THE CULINARY PEA

OTHER MENDELIAN CHARACTERS 27

We now leave the characters of the pods and pass
on to the colour of the seed coats, which is the last
of the characters dealt with by Mendel which we
shall consider at present. The two colours that
Mendel worked with were white and ‘‘ grey ” — or,
more strictly speaking, a grey-green that becomes
brown with age. They were recognised in the
sixteenth century by Thomas Tusser, who refers to
them in his remarks on February in his “ Hundred
Points of Good Husbandrie,” which is printed on
p. hi of this book.

“Sowe runcival* timelie and all that be gray,
but sowe not the white till S. Gregories day.”

The white seed-coat is characteristic of the majority
of peas commonly grown in the kitchen-garden. A
plant that has white seed- coats always has white
flowers. The grey colour of the seed-coat is charac-
teristic of many but by no means of all field peas. In
addition to the plain grey of the seeds there is very
often a greater or lesser amount of purple colour in
the form of minute spots distributed uniformly over
the seed-coat. This purple colour is left out of
account in considering the result of crossing white
with grey. A plant with a grey seed-coat always
has purple flowers.

A grey-skinned pea is shown in Fig. 11. This
particular race, which was given to me by Mr. R. H.
Lock, does not exhibit the purple spots. An ordinary
white-skinned culinary pea is illustrated in Fig. 12.

♦ A kind of pea, figured by Gerarde.

28

BREEDING

I have shown the peas in the pods, in order to
emphasise the fact that all the seed-coats on a plant
are the same — all grey on the plant in Fig. 11,
and all white on the plant in Fig. 12. This is a
very important point to bear in mind.

The hybrid produced by crossing a white-skinned
pea with a grey-skinned one was found by Mendel
to be grey. The second hybrid generation consisted
of 705 grey-skinned plants, and 224 white-skinned
ones. A hundred of the former were tested, and
thirty-six gave grey-skinned ones only ; the remain-
ing sixty-four gave both grey-skinned and white-
skinned plants.

Let us now take a retrospective glance at the
five characters of the pea, the manner of inherit-
ance of which has just been described.

Let them be considered first from the point of
view of the time at which they appear in the life-
history of the plant. The difference between tall
and dwarf appears in the very young seedling long
before it needs to be supported with a stick. The
difference between normal and fasciated cannot be
detected until the plant has grown to some height
and needs the support of a stick ; but it can be seen
before the appearance of the flower buds. The
difference between hard and soft pods cannot, of
course, be detected until after the flower is “ over ”
and the pod has begun to develop, but it can be seen
very soon after the pod projects beyond the withered
petals, and long before the peas begin to fill out the

Fig. 11.— “GREY” PEA

OTHER MENDELIAN CHARACTERS 29

pod. It is not, however, until this stage in the
development of the pod is reached that the difference
between the yellow and the green pod can be observed.
And, lastly, the colour of the seed-coat cannot be
seen, because it is not developed, until the pod, con-
taining it, is ripe and dry.

It will be noted that in the above recapitulation
the five characters have been placed in the order in
which we have considered them in detail, and also
in the order in which they appear on the plant.

There remain two characters of the pea dealt with
by Mendel ; but they will be taken out of their order.
They are the earliest of all to appear ; so early,
indeed, that they can be seen before the seed is
planted — i.e. as soon as it is taken out of the pod.
It will be remembered that the earliest character
which we have yet considered (the distinction
between tallness and dwarfness) was visible in the
young seedlings. The two characters still to be dealt
with belong to the first two leaves of the plant which
remain in the seed, and never come above the ground.
These first two leaves, or cotyledons, are so important
economically, because they contain the food-material
(starch and sugar) which renders the pea valuable
as food, and the characters in respect of which they
differ have played so prominent a part in estab-
lishing the reality of the Mendelian phenomenon, and
in testing the theories put forward to explain it, that
a special chapter will be devoted to their consideration.

Let us now take a general survey of the five

30

BREEDING

instances of the Mendelian phenomenon with which
we have dealt, and note the points which they
exhibit in common. In each of the five instances
the characters were considered in pairs — tall and
dwarf, normal and fasciated, and so on. The justi-
fication for this is that in every case the two characters
of a pair always appertain to the same part of the
plant — to the stem, or pod, or seed-coats. And not
only so, but to the same feature of a particular part
of the plant. Thus tallness of stem and softness of
pod do not constitute a pair, because they relate to
different parts of the plant. Nor even do tallness of
stem and fasciation of stem constitute a pair, because,
though both relate to the stem, one relates to the
length of the internode and the other to the arrange-
ment of the nodes. Thus the two characters which
constitute a pair appertain to the same feature of a
particular part of the plant.

A further characteristic of the two members of a
pair is that when an individual bearing one of them
is mated with an individual bearing the other, the
resulting hybrid usually bears one of the pair of
characters to the complete exclusion of the other.
In the case of tallness and dwarfness, for instance,
the hybrid exhibits no trace of the dwarf character ;
but in the case of hardness and softness of pods, the
pod of the hybrid is a little softer than that of
its hard-podded parent. That character of a pair
which completely or partially excludes the other of
the pair, in the hybrid, is called the Dominant member
of the pair ; the other character is called the Recessive

FIG. 12.— “WHITE” PEA

<

4^ ,V '

. 4' ■ *

. -r; ■?■; ,

-d .

OTHER MENDELIAN CHARACTERS 31

one because it recedes, as it were, for one generation,
but reappears in the next.

The dominant and recessive members of the five
pairs of characters which we have dealt with are
shown in the following table : —

Feature of the Plant a^ecied
hy the two characters of
the Pair.

Dominant
Member of
Pair.

Recessive
Member of
Pair.

1st pair

Length of intemode

Long

(producing

tallness)

Short

(producing

dwarfness)

2nd „

Arrangement of nodes

Normal

Abnormal

(producing

fasciation)

3rd „

Texture of pod

Hard

Soft

4th „

Colour of unripe pod

Green

Yellow

5th „

Colour of ripe seed-coat

Grey

White

The Mendelian phenomenon {see the Frontispiece)
can now be described in general terms.

The result of crossing an individual bearing a
dominant character with one bearing a recessive
character are hybrids bearing the dominant character.
These constitute the first hybrid generation. The
second hybrid generation consists both of individuals*
bearing the dominant character and of individualsf
bearing the recessive character, in the ratio of three
dominants to one recessive in every four, on the
average. Of these dominants one is a pure
dominant, and will produce only dominants. The

* Sometimes called dominants for brevity,
t Sometimes called receesives for brevity.

32

BREEDING

other two are hybrids, and will produce dominants
and recessives in the proportion three to one,
just as the hybrids of the first hybrid genera-
tion did.

We see, therefore, that when the recessive
reappears in the second hybrid generation (an
extracted recessive, as it is then called), we may
count on its being pure, and be confident that
it will breed true ; but we cannot tell whether
the dominants which occur in the second hybrid
generation are pure dominants or hybrids until
we have bred from them. This can be done in
the case of self- fertilising plants by allowing self-
fertilisation to take place and by sowing the seed
thus set. If dominants only are produced, it is
pure ; if dominants and recessives, it is hybrid. But
there is another way which is much more suitable
in the case of animals, in which the sexes are separate.
It has been found that if a hybrid of the first hybrid
generation is mated with a recessive, half of the
offspring are hybrids (bearing the dominant character,
of course), and half are recessives.

Fig. 13.— ANDALUSIAN FOWLS

(Reproduced /7-oi)i W7 ight's Book of Poult7y.'\

CHAPTER IV

GENERAL CHARACTERISTICS OF THE MENDELIAN PHENO-
MENON ILLUSTRATED BY THE CASH OF THE
ANDALUSIAN FOWL AND THE INHERITANCE OF
EYE-COLOUR IN MAN

I PROPOSE to give now some further instances of the
Mendelian hereditary phenomenon to illustrate some
of its go’^.eral characteristics. One of the most
striking instances of the Mendelian phenomenon is
that which is exhibited by the breeding properties
of the Andalusian fowl. In perhaps no other case
is the contrast between what actually occurs
and what we expect to happen, seen in bolder
relief.

The blue Andalusian fowl owes its popularity
with the public, and its interest to the student of
heredity, to its colour. This is a slaty blue-grey.

If two birds of this strain are mated, it will
be found that they do not breed true. Besides
the blue birds which hatch out, there will also
be a certain number of blacks, and of whites with
occasional black points, hereafter to be referred to
simply as “ whites.” If the blues thus produced are
mated together, blacks and “ whites ” will again
appear amongst their offspring, and no amount of
breeding from the blues alone will rid them of the

D

33

34

BREEDING

black and white blood which will crop out at every
generation, although blue birds only are mated
every time.

I gave an account of the above facts in an article
on Mendelian principles which I wrote for the
Country Home, An acquaintance, who kept fowls,
read the article, and said to me that though the
article as a whole was very interesting, what I had
said about Andalusian fowls was, of course, untrue.
“ It stands to reason,” she said, “ that if you continue
to breed from the Andalusians alone you will ulti-
mately fix the strain. It is common sense.” Let us
see what actually happens. When blue Andalusians
are mated together, birds of three types of colour are
produced in the following proportions : one black,
two blue Andalusians, and one white in every four
birds, on the average. The blacks are found to
breed true ; so are the whites. The Andalusians
when mated together again produce these three
types in the same proportions. The result may be
expressed in the form of a genealogical table, as
follows ; —

Blue X Blue

1. Black 2. Blue 1. White

Blacks only 1. Black 2. Blue 1. White Whites only

It is impossible to look at this table without
suspecting that the blue Andalusian is a hybrid, like

CASE OF THE ANDALUSIAN FOWL 35

the hybrid tall pea, and that the two forms which
correspond to the pure tall and pure dwarf in the
peas are the black and white birds respectively. Or,
rather, this case more closely resembles that of the
texture of the pods in the pea where, it will be remem-
bered, the result of crossing a hard-podded with a
soft-podded pea was a pea whose pods were inter-
mediate in texture — ^between those of its two parents,
but closer to those of the former.

Be this as it may, the result of crossing a black
and a white is a blue Andalusian, and blue birds
obtained in this way, when mated together, produce
a generation of fowls consisting of about one black,
two blues, and one white in every four.

The Mendelian phenomenon exhibited by these
fowls may be set forth in the form of a genealogical
table (Fig. 14) corresponding to that representing
the result of crossing a tall with a dwarf pea. A
comparison between the table of the fowls and that
of the peas brings out one important point of differ-
ence between them. In the case of the peas, the
hybrid tall, whether it occurs in the first or second
hybrid generation, only differs from the pure tall
by its breeding properties and not by any outward
characteristics of its own — or, more strictly, not by
any characteristics of its own so far as we know at
present. It is not improbable that accurate measure-
ment may reveal a difference. For the present,
however, at any rate, the hybrid tall and pure tall
are identical so far as we know them, except in
their breeding properties. In other words, the tall

BREEDING

36

character behaves as a simple dominant when it
meets the dwarf character, in a cross.

But the case exhibited by the Andalusian fowl
is different. The so-called blue colour of this bird
may, or may not, be described as an intermediate
between the black and white of the fowls which
produce it when they are crossed ; but, at any rate,
it is different from either of them, and the important
point to note is that the hybrid, whether of the first
or second or third hybrid generation, is distinguished
from both of the parents of the cross, not only by
its breeding properties, but by easily observable
external characteristics.

The reader will probably admit the reasonableness
of referring to the black fowl as the dominant, and the
white as the recessh'^, although the black is not
absolutely dominant ■ ' /er the white in the result of
the cross. It is exped-ent to use these terms in this
case, because it is convenient to have a general term
to apply to the characters of the two forms crossed.
It is also legitimate to use them because the difference
between the result r crossing the black and the
white fowl and t suit of crossing the tall and
dwarf pea is only one of degree. In the case of the
pea, tallness is completely dominant over dwarf ; in
the case of the fowl blacks .ss may be said to be only
partially dominant over whiteness. Moreover, it may
be that the appearance of complete dominance is
merely due to the fact that we are as yet unable to
detect the difference between the hybrid and the
dominant in the case of stature in peas. This is no

fig. 14.-DIAGRAM TO ILLUSTRATE THE MENDELIAN INHERITANCE OF COLOUR IN THE ANDALUSIAN FOWL

CASE OF THE ANDALUSIAN FOWL 37

*11>3 prophecy ; for a difference between the hybrid
and the dominant parent has recently been detected
by the aid of the microscope in a stock instance of
simple dominance — roundness and wrinkledness in
peas. This will be described in detail in Chapter IX.

If, therefore, we compare the two genealogical
tables (Frontispiece and Fig. 14), we see that the
Mendelian phenomena, as exhibited by these two
cases, are perfectly analogous. The only difference
between the two is that in one case (that of the peas)
the completeness of dominance of one character
makes it impossible to distinguish between dominants
and hybrids, so that the second hybrid generation
appears at first sight to consist of three individuals
bearing the dominant character, and one bearing the
recessive in every four, on the average ; and it is
not until we breed from these apparent dominants
that we discover that one of the three is a pure
dominant and the other two are hybrids. In fact/
it is only when we have raised the third hybrid
generation that we discover that the second hybrid
generation consists of one pure dominant, two
hybrids, and one recessive in every four.

But in the case of the fowls, where the dominance
of black over white is incomplete, or, at any rate,
where the hybrid bears a colour which is peculiar to
itself, the real constitution of the second hybrid
generation (one pure dominant, two hybrids, and one
recessive) is apparent directly the individuals com-
posing it have been raised.

If the Frontispiece and the Andalusian pedigree

BREEDING

38

are folded out and compared it will be seen that the
two are fundamentally the same, and correspond with
one another individual for individual. The point in
which they differ is an unessential feature of the
Mendelian phenomenon — namely, the matter of
dominance. The fact that the dominant and hybrid
tall appear to us identical is probably no more than
a measure of the crudeness of the means which have
hitherto been adopted to distinguish between them
— observation by the naked eye.

The point in which the two tables agree is the
essential feature in Mendelian inheritance, and con-
sists in the orderly reappearance of the characters
of the two varieties crossed, amongst the progeny of
the cross, in definite numerical ratios and in a state
as pure as that in which they existed in the two
parental varieties before the cross was made. The
term Segregation is commonly used to denote this
splitting out of the parental characters amongst the
offspring of hybrids.

We may obtain some idea of the value of a know-
ledge of the Mendelian phenomenon to the practical
breeder by comparing the fate of a would-be breeder
of Andalusians who possessed this knowledge with that
of one who did not. The breeder who had not heard
of the Mendelian discovery, or had heard of it and
had rejected it as contrary to common sense, would
continue to breed from his Andalusians from genera-
tion to generation in the hope that ultimately he
would eradicate the black and also the white taint.
But he would never obtain more than 50 per cent.

CASE OF THE ANDALUSIAN FOWL 39

Andalusians amongst his chickens, and he would not
be able to sell the blacks and whites, which would
make up the remaining 50 per cent., at anything like
the price which the Andalusians fetch. Mr. Sturges,
in his “ Poultry Manual,” states (p. 244) that these
so-called mismarked chicks are usually killed at an
early age or sold as wasters.

The breeder familiar with the Mendelian pheno-
menon would never mate Andalusians together, but
keep a small stock of blacks and whites, and always
pair black cocks with white hens, or vice versa. In
this way he would be relieved of the expense of
rearing, and the trouble of getting rid of the black
and the white wasters ; and would be certain of
raising 100 per cent. Andalusians, all of which he
would sell at good prices.

The case of the Andalusian fowl places the whole
question of the fixation of varieties in an entirely
new light.

I Breeders have long been familiar with certain
varieties which were tolerably easily fixed, and with
others which seemed to withstand every effort of
the breeder to fix their qualities. The Andalusian
fowl has for long been a flagrant example of such
an unfixable variety. For fifty years it has been
bred, and it comes true to type no more now than
when the attempt to fix it began. The reason is
simply that the Andalusian is a hybrid : the par-
ticular blue colour, which characterises it, is peculiar
to and only manifested by the hybrid bird, whether
this be produced by the union of such hybrids or

40

BREEDING

by the union of a black and a white. Mr. Sturges,
indeed, seems to incline to the view that the Anda-
lusian actually arose from such a cross. There is no
doubt that the Andalusian is a very near relation of
the Black Spanish, and that both of them came from
Spain about the middle of the nineteenth century.
In 1846 a Mr. Barber imported black and also white
fowls from Spain ; and Mr. Sturges suggests (“ Poultry
Manual,” p. 437) that “ these may have been the
foundation of the Andalusian. This,” he continues,
“ is borne out by a further note : Another gentleman
says : ‘ I have a few chickens out from Mr. Barber’s
Andalusian hens (called Andalusian because they
came from thence), some of which seem to be the
true old Black Spanish, and some a grizzly white . . .
some of these are of a blue or grey or slaty colour.’ ”
Here we catch a glimpse of the Mendelian phenomenon
being unfolded before the eyes of a man who cannot
have apprehended its significance, and before Mendel
had begun his experiments. Such glimpses are com-
forting to those who, like myself, have a horror that
this kind of orderliness in natural processes is not
seen or, at any rate, not so easily seen until we have
been led by some portent, like Mendel’s discovery,
to expect to see it.

HUMAN EYE COLOUR

In the case of the Andalusian fowl we considered
an instance of the Mendelian phenomenon in which
the hybrid bears a character peculiar to itself — the
blue colour we have described.

fig. 15.— vertical section of human eye of

SIMPLEX TYPE

Bulging out to the left is the tough transparent cornea. At the points where this
joins the main wall of the eye there are two projections, each bordered on their
sides remote from the cornea by a dark line ; this is the iris seen in section.
The space between the free ends of the projections corresponds to the pupil of
the eye. The large oval body behind the iris is the lens. The branching pro-
jections just above and below the lens and the long ones farther to the right
do not concern us.

HUMAN EYE COLOUR

41

The character which I propose to consider now
resembles the case of stature in peas, in that the
hybrid does not differ externally, so far as we
can see, from the pure dominant. It is also of
especial interest because it is one of the few
Mendelian characters which have been discovered
in man.

When we speak of the colour of the human eye,
we are speaking only of the colour of a particular
part of it namely the iris, which is the only part
of the eye the colour of which varies to any great
extent. There is, of course, an almost infinite variety
in the colour of the eye ; but the various types of
colour can be grouped into two large classes according
as to whether there is brown pigment on the outer
surface of the iris or not. When the whole of this
surface, which is the one that we see when we look
at the eye, is covered with brown pigment, the eye
is brown ; and when there is no such brown pigment on
the outer surface of the iris the eye is generally blue.
The word duplex has been coined to denote those
eyes in which there is some brown pigment in front
of the iris ; and simplex to denote those eyes in
which there is none.* Speaking generally, duplex
eyes are brown, and simplex are blue. But as the
two classes, duplex and simplex, include the whole
range of human eye colour it is evident that duplex
does not simply mean brown and simplex blue,

* These terms were introduced by Mr. C. C. Hurst, who was the first to
discover, investigate and record the Mendelian inheritance of eye colour
in man.

42 BREEDING

because they also include the green, the grey, and
the hazel eyes.

A duplex eye is, as we have said, any eye in
which there is brown pigment in front of the iris.
This brown pigment exists on the top of the fibrous
tissue of the iris which appears blue ; so that if we
could dissolve away the brown pigment from a dark
brown eye the blue would appear underneath. There
is every degree in the amount of this brown pigment.
If there is very little indeed, it usually exists as a
thin ring round the pupil, all the rest of the iris
being blue. This has been called a ring-duplex, and
if there is a mere trace of the brown pigment, a duplex
eye of this kind would doubtless often pass as a
“ grey ” or even “ blue ” eye. If there is more of
the brown pigment, a “ hazel ” eye is very often the
result. In another form of duplex eye the brown
pigment is distributed in patches over the blue
ground colour which can be seen between the patches ;
“ hazel ” eyes are also often the result of this arrange-
ment. When there is a very thin uniform layer of
brown pigment, the colour of this (a sort of yellow
ochre) blends with the blue of the background, which
can be seen through, and a “ green ” eye is the result.
And, finally, there are the self-coloured duplex eyes
in which the brown pigment is spread uniformly
over, and completely concealing, the blue back-
ground, giving the various intensities of brown eye,
according to its abundance.

A simplex eye is one in which there is no brown
pigment on the anterior surface of the iris. All the

HUMAN EYE COLOUR

43

clear blue and grey eyes are of this class. The colour
of simplex eyes is not due to the colour of the fibrous
tissue of the iris itself, which consists of muscle
fibres, nerves, blood-vessels, and so forth, but to a
layer of dark purple pigment behind the iris, called
the uvea. In the new-born infant the fibrous tissue
of the iris is very delicate and transparent, so that
most of the colour of the uvea can be seen through
it : that is why the eye of the new-born infant is
dark blue. A clear blue eye is due to the delicacy
of the fibrous tissue of the iris, which permits the
colour of the uvea to shine through. A pale blue or
grey eye is due to the coarseness or “ stringiness ”
of the fibrous tissue, which prevents most of the
colour of the uvea from being seen. Evidence that
the colour of very pale blue and of grey eyes is due
to the colour of the uvea which can filter through the
fibrous tissue in front of it, and not to the colour of
the fibrous tissue itself, is afforded by the fact that in
albinos in whom the uvea is colourless the iris
appears pink, this colour being given by the blood in
the fibrous tissue. The various colours of simplex eyes,
namely, the different intensities of blue and grey eyes
and the intermediates between these two colours, are
therefore due to differences in the texture of the
fibrous portion of the iris, which allow different
amounts of the colour of the purple uvea behind to
filter through.

A simplex eye, therefore, is one in which there is
only one layer of pigment, the purple of the uvea;
a duplex eye is one in which there are two layers of

44

BREEDING

pigment, the purple of the uvea behind the iris
and a varying amount of brown pigment in front
of it.

If the reader prefers the use of familiar terms
duplex may be spoken of as brown and simplex as
blue, and in nine cases out of ten the application of
these common words will be correct ; but it must
be remembered that there are some eyes which would
be described as blue in common parlance, but which
really fall into the duplex class because they possess
a trace of brown pigment.

The result of a union between a person with a pure
duplex eye and one with a simplex eye is a child
with a duplex eye. Duplex, therefore, is dominant
and simplex is recessive. The generation produced
by the union of such hybrid duplexes, as we may
call them, consists of three duplex and one simplex
in every four. These simplexes breed true, i.e.
produce offspring with simplex eyes only, when mated
with their like. Of the three duplexes, one, the pure
dominant, will, if mated with another of the same
kind produce pure duplexes only. The other two
duplexes are hybrid, and when mated with similar
hybrid duplexes will produce duplexes and simplexes
in the ratio of three to one. The reader may make a
scheme of the mode of inheritance of human eye-
colour by substituting duplex for tall, and simplex
for dwarf in the frontispiece.

At the beginning of the last paragraph I spoke of
“ the result of a union between ... a fure duplex eye
...” By “ pure ” I do not mean anything which

HUMAN EYE COLOUR

45

relates to the colour of the eye. I mean pure from
the point of view of its breeding properties, as opposed
to hybrid, in order to distinguish it from the latter.
But the question will present itself to the reader :
How is a pure duplex to be distinguished from a
hybrid one ? The answer has been briefly given at
the conclusion of the third chapter. It was there
stated that the simplest way, in the case of animals, to
distinguish between the pure dominant and the
hybrid bearing the dominant character was to mate
the individual in question with one bearing the
recessive character. The reason that it is simpler to
find out whether an individual is hybrid or pure
by mating it with a recessive, than by mating it
with another hybrid (which was the means adopted
in the case of the peas) is that there is no means
of finding a hybrid to mate it with, other than
by breeding one for the purpose by mating a
dominant with a recessive — for this purpose it does
not matter whether “ the dominant ” is pure or
hybrid. And even if a hybrid could be found, if
the individual to be tested happened to be a pure
dominant, the result of mating it with a hybrid
would be equal numbers of pure dominants and
hybrids, all of which would be alike ; that is to say,
it would be the same result, so far as we could see
from the characters of the individuals themselves,
as that of mating two pure dominants, which would,
of course, be nothing but pure dominants.

The difficulty of distinguishing between pure
dominants and hybrids has not presented itself

BREEDING

46

before, because in the peas where self-fertilisation
always takes place, when it is not artificially prevented,
like always mates with like, so that there is no chance
of a hybrid mating with a dominant ; and because
in the case of the Andalusian fowl, the hybrid differed
from the dominant (which we called the black type
for convenience, and not without justification) in
visible characteristics. This difficulty presents itself
not only in the case of human eye colour, but in all
cases in which self- fertilisation does not occur, and
the hybrid is externally indistinguishable from the
dominant parent. The practical solution of the
difficulty is to mate the individual in question with
one bearing the recessive character. If all the off-
spring bear the dominant character, it was a pure
dominant ; if half bear the dominant and half the
recessive, it was a hybrid. Another reason why it
is better to test the nature of an individual bearing
the dominant character by mating it with a recessive,
than by mating it with a hybrid, is that if the
individual tested is a hybrid it will only produce
one recessive in every four in the latter case, but
will produce one in every two in the former. And
where only a small total number of offspring can be
raised this is obviously an advantage.

I have discussed this at some length, because it
is a question which is very likely to crop up in practice
and because I wish to lay emphasis on a general pro-
perty of the hybrid, namely, that when hybrids are
mated with recessives they produce hybrids and reces-
sives in equal numbers, and when mated with domi-

HUMAN EYE COLOUR

47

nants they produce dominants and hybrids in equal
numbers. Thus, if our tall peas of the first hybrid
generation had been crossed with dwarfs they would
have produced tails and dwarfs in equal numbers ;
and if Andalusian fowls had been mated with the
whites they would have produced equal numbers of
Andalusians and whites. And if hybrid tails had been
crossed with pure tails, equal numbers of these two
kinds of tails would be produced ; and if Andalusian
fowls were mated with the blacks equal numbers of
black and Andalusians would have been the result. It
is desirable to know this property of hybrids, because
if one does not, the existence of families composed
of both blue-eyed and brown-eyed children, one of
the parents of whom was blue-eyed whilst the other
was brown-eyed, does not seem reconcilable at first
with the statement that the result of mating duplex
(or brown) with simplex (or blue) is invariably duplex.
There is, of course, no contradiction, as the reader
will now readily see ; in such cases the duplex parent
has evidently been a hybrid duplex.

Moreover, this type of mating — hybrid by
dominant, or hybrid by recessive — plays an im-
portant part in the Mendelian theory of sex,
and in the theory of the origin of Mendelian
characters.

Before quitting the subject of the inheritance of
eye-colour in man, it may be useful to set forth the
breeding properties of simplex, and of pure and
hybrid duplex eyes, in a few short general state-
ments.

48

BREEDING

The nature and result of the union between two
persons, both with duplex eyes, may be any one of
the three following kinds : —

A. Pure duplex by pure duplex, giving all pure

duplex.

B. Pure duplex by hybrid duplex, giving these

two kinds in equal numbers.

c. Hybrid duplex by hybrid duplex, giving 25 per
cent, pure duplex, 50 per cent, hybrid duplex,
and 25 per cent, simplex.

Unions between duplex and simplex can only be
of two kinds : —

D. Pure duplex by simplex, giving all hybrid

duplex.

E. Hybrid duplex by simplex, giving hybrid

duplexes, and simplexes, in equal numbers.

Unions between two simplexes can only be of one
kind, inasmuch as there is only one kind of simplex,
the pure.

The above statements are true of the three kinds,
simplex, pure duplex, and hybrid duplex, whatever
' may have been their origin. Thus, if a simplex
occurring in the second hybrid generation mates
with another simplex of like extraction he or she
will produce nothing but simplexes as assuredly as
will two individuals both with simplex eyes who
have descended from, say, two generations of simplex
ancestry. Or, to express this in general terms, the
offspring of a man and a woman both possessing blue

HUMAN EYE COLOUR 49

eyes, but both descended from (of course, hybrid)
duplex parents on both sides in both cases, will
possess blue eyes as certainly as if their four grand-
parents all had blue eyes. The offspring of the
union of two persons with simplex eyes, whatever
their ancestry is, will never have brown eyes. At
any rate, no exceptions to this rule have yet been
observed and recorded.

E

CHAPTER V

THE INHERITANCE OF THE CHARACTERS OF THE SEED
IN THE CULINARY PEA

We have so far dealt with only five of the seven
characters of the pea experimented with by Mendel ;
and the reader may remember that they were dealt
with in the order in which they appeared on the
plant : the first was the stature of the plant,

whether tall or dwarf; the last was the colour of
the ripe “ seed,” whether “ grey ” or “ white.”
Suppose that it is spring, and that you have
just sown the seed to produce the second hybrid
generation from a cross involving any of the five
pairs of characters already dealt with. If the pair
of characters is tallness and dwarfness, you will
be able to count the numbers of tails and dwarfs
as soon as the young plants are a few inches high,
that is to say, early in May ; but it will be necessary
to wait till the plant is ripe and dry, that is to say,
till the middle of August at the earliest,* before the
number of white-seeded and grey-seeded plants can
be counted. And the numbers of plants with the
three characters intervening in the date of their
appearance between these two extremes can be
counted at intervening dates.

* Except when there is a drought like that of the summer of 1911.

50

INHERITED SEED-GHARACTERS 51

The two characters, to be dealt with now, appear
earlier than the earliest of the five which we have
so far considered, namely stature. But these two
characters appear by the heading of the chapter
to be characters of “ the seed ” ; yet I have just
said that the difference between grey-seeded and
white-seeded plants was the latest to appear. How
can characters of “ the seed ” be both the earliest
and the latest to appear ? ^What is the meaning
of this apparent contradiction ? There is no con-
tradiction, and we can understand how a character
of “ the seed ” can be both the earliest and latest
to appear when we understand what a seed is.

Suppose that you decide to repeat Mendel’s
experiments with peas, and to cross tails with
dwarfs ; suppose that you decide to breed tails
and dwarfs separately for five generations to make
sure that they breed true to their characters before
you cross them ; and suppose, further, that you
buy some seed of a tall variety and some of a
dwarf, and sow it ; the plants raised may be de-
scribed as belonging to the 'first generation of the five
that you propose to breed. Now, consider one of
the seeds produced by one of these plants. Only the
seed- coat — the thin, outer skin of the seed — is a part
of this plant of the first generation ; all that is
inside this seed-coat is the embryo plant of the
next, or second, generation, together with its first
two relatively enormous ‘‘ leaves.” If the seed is
soaked and dissected the two hemispherical ‘‘ leaves,”
or cotyledons as they are called, are seen to fill up

52

BREEDING

almost the entire space within the seed-coat; and,
lodged between them — and, of course, attached to
them — are seen the little shoot and root, which
are all there is, as yet, of the embryo plant.

These two cotyledons (which are the hemi-
spheres which constitute “ split peas ”) are supposed
by some botanists to represent leaves which have
become greatly modified in accordance with their
having taken on a function not usually performed
by leaves, namely, that of acting as a storehouse
of food material for the young plant. Unlike
ordinary leaves, these cotyledons do not come
above the ground when the seed germinates. In
the case of many plants, however (as, for instance,
the sunflower), in which the cotyledons are less
specialised as storehouses of food material, they
do come above the ground ; and, although in the
latter case they behave physiologically like ordinary
leaves, they nearly always differ in shape from
the other leaves of the plant.

The whole of this store of food material is laid
down in the first two leaves (or cotyledons) of the
embryo plant before the embryo pea-plant has
become detached from its parent. In other words,
the food material in the first two leaves of our
plant of the second generation is being laid down
whilst that plant is undergoing its early develop-
ment within the seed-coat, which, as we have said,
is a part of our plant of the iirst generation,
just as the embryo of a mammal is nourished in
the womb. In fact, the womb with its contained

INHERITED SEED-CHARACTERS 53

embryo is roughly comparable to the seed : the
womb is a part of the mother, the embryo is the
next generation. The seed-coat is a part of the
parent-plant, the embryo with its two cotyledons
is the next generation.

Mendel experimented with two pairs of
characters of these cotyledons : one pair related
to their colour, which was either yellow or green ;
the other to their shape, which was either round
or wrinkled. Let us consider the colour first.

In the title to this chapter I have spoken of its
contents as being concerned with the characters
of “ the seed.” But, now that I have explained
that a seed is a composite structure, i.e. that
the seed-coat is a part of the parent plant, '
and that everything inside this seed-coat is the
embryo of the next generation, it is necessary
when we speak of the characters of “ the seed ”
to make it perfectly clear whether we are referring
to the seed-coat which belongs to one generation
or to the cotyledons which belong to the next.
Let it be clearly understood, therefore, that the
character with which we are now dealing is the
colour of the cotyledons, and not of the seed-coat.

Mendel crossed a plant the first two leaves or
cotyledons of which were yellow, with a plant
the cotyledons of which were green. The result
was a plant the cotyledons or first two leaves of
which were yellow.

This description of the cross made by Mendel
may sound unnecessarily explicit. But it is not so.

54

BREEDING

The bald statement that he crossed a yellow pea
with a green pea leaves one completely in the dark
as to the true nature of the character dealt with ;
and even the statement that he crossed a yellow-
seeded with a green-seeded pea does not make it
clear whether by yellow-seeded (for instance) is
meant a pea-plant grown from a yellow seed, or a
pea-plant which produces yellow seeds. What should
be meant by “ a yellow pea ” or ‘‘a yellow-seeded
pea ” is a pea-plant grown from a yellow seed, because
it is the yellow colour of its first two leaves or coty-
ledons, seen through the enveloping seed-coat, which
is the character with which we are concerned ; and
also because, as we shall see later on, a plant grown
from a yellow seed does not necessarily produce
yellow seeds.

When a tall is crossed with a dwarf pea the result
cannot be seen until the seed resulting from the
cross is sown. But the result of crossing a yellow-
seeded pea (as defined above) with a green-seeded
pea can be seen directly the pod, which has developed
from the flower on which the cross was made, is
opened. Let me briefly describe here how such a cross
is made. A detailed account is given onpp. 146-152.
To make a cross between a tall and a dwarf, pollen
is taken from a flower of a tall pea and placed upon
the pistil in the flower of a dwarf, or vice versa.
We will consider the former case. A label is tied
round the stalk of the flower of the dwarf plant
which has received the pollen from the tall. In due
course, the petals wither ; and, if all is well, the pod

INHERITED SEED-CHARACTERS 55

grows from within the withered petals. When the
pod is dry and ripe, which is not likely to be much
earlier than the beginning of August, it is opened,
the seeds are taken out, and sown in the following
spring. All the seeds will produce tall plants, tallness
being dominant over dwarfness. The result would
have been the same if the flower of a tall plant had
been pollinated with pollen from the flower of a
dwarf one. The point is that the result of a cross
made, say, in the summer of 1910 is not seen until
the spring of 1911 in the cases of tallness and dwarf -
ness. And the result of a cross between a grey-
seeded pea (or, more strictly, a pea-plant which
produces grey seed-coats) and a white-seeded pea
(i.e. a pea-plant which produces white seed-coats)
made in the summer of 1910 will not be seen till
the autumn of 1911.

But if, in the summer of 1910, pollen is taken
from the flower of a plant, the first two leaves or
cotyledons of which were yellow, and placed on the
pistil of the flower of a plant which had green
cotyledons, the result of the cross is seen when the
pod which develops from the latter flower is opened.
For instead of the pod containing “ green seeds ”
(as all the other pods of the plant will, if no other
crosses have been made on the plant), it will contain
yellow ones, because the first two leaves of the
plants of the first hybrid generation will be yellow,
inasmuch as yellowness in the cotyledons is dominant
over greenness. That is to say, whilst the result
of a cross made in the summer of 1910 between a

BREEDING

56

tall and a dwarf will not be seen till the spring of
1911, the result of a cross made on the same day
between a plant which had yellow cotyledons and
one which had green, will be seen in the autumn of
the same year, namely 1910.

Before we proceed to describe the reappearance
of the parental characters in the second hybrid
generation, let us pause to consider the Mendelian
phenomenon manifested by the colour of the coty-
ledons, and compare it with the other instances of
it with which we have dealt up to the present.

So far as we know, yellowness is a simple dominant
to greenness, in the case of the cotyledons in Pisum,
just as tallness is to dwarfness. That is to say, no
difference between the pure yellow and the hybrid
yellow has so far been detected ; but it does not
seem to me unlikely that a difference between them
may be revealed by spectral analysis.

It is curious that yellow is dominant over green
in the case of the cotyledons, in view of the fact that
in the case of the pods green is dominant over yellow,
according to Mendel’s account.

We will now follow up the results of the supposed
cross that was made in the summer of 1910. As a
matter of fact, I did make many such crosses during
that summer.

In our supposed cross a flower of a plant raised
from a green seed was pollinated in the summer of
1910 from a flower of a plant raised from a yellow
seed ; and in the autumn of the same year the result

INHERITED SEED-CHARACTERS 57

was seen as soon as tlie pod which developed from the
flower of the former plant was opened. The seeds
in it, instead of being green like those on all the other
pods on the plant, were yellow ; or, more accurately
stated, the cotyledons within the seed-coats in this
pod were yellow, instead of green as they would be in
all the other pods. Stated in the most general, but
also the most simple and the most correct, terms, the
result which we have hitherto observed is that when
a plant whose first two “ leaves ” are yellow is crossed
with a plant whose first two “ leaves ” are green
the result is a plant whose first two “ leaves ” are
yellow. Such a ‘‘ plant ” constitutes the first hybrid
generation, and though the fact that it bears the
dominant character can be observed in the year in
which the cross was made, the ‘‘plant” does not, of
course, become mature until sown in the following
year, 1911.

The yellow peas in the pod which developed from
the flower on which the cross between the yellow-
seeded and the green-seeded variety had been made,
would be taken from the pod in the autumn of the
same year in which the cross was made (1910), and
sown in the following spring. Let us suppose there
were five seeds in the pod, and that five plants were
produced from them during the summer of 1911.

These five plants, growing in 1911, belong, as
already indicated, to the first hybrid generation
produced by crossing, in 1910, a yellow-seeded (as
defined above) with a green-seeded pea ; just as the
tall plants growing in 1911 belonged to the first

BREEDING

58

hybrid generation produced by crossing, in 1910,
a tall with a dwarf pea. The only difference between
the two cases is that the character of the first hybrid
generation can be seen before the seed to produce
plants of the first hybrid generation is sown, in the
case of the colour of the cotyledons ; but cannot
be seen until after the plants of the first hybrid
generation have grown a few inches, in the case
of the stature of the plant.

It will be remembered that the colour of the first
two “ leaves,” or cotyledons, of the plants of the
-first hybrid generation could be seen in the pods of
one of the parent plants in the same year in which
the cross was made. Similarly, the colours of the
cotyledons of the plants of the second hybrid
generation can be seen in the pods borne by the
plants of the first hybrid generation ; that is to say,
in the case of the cross made in 1910, in the
autumn of 1911.

In Plate I. are shown the results of crossing a
green-seeded with a yellow-seeded pea, as far as the
second hybrid generation which we are now con-
sidering. Above, to the left, are seen a group of
ten peas, the cotyledons in which are green ; on the
right a group of six peas containing yellow cotyledons.
The skin of the vividly yellow pea to the top left
of this group has been removed in order to show
the colour of the cotyledons themselves. These two
groups represent the parents of the cross. The
group of three yellow peas in the middle, below,
contain the cotyledons of the first hybrid generation.

PLATE I.— MENDELIAN INHERITANCE OF THE COLOUR OF THE
SEED IN THE CULINARY PEA

(Top left) Green Parent. (Top right) Yellow Parent.

(Three Peas in middle line) First Hybrid Generation.

(In the pods) Second Hybrid Generation.

INHERITED SEED-GHARAGTERS 59

And in the two pods are seen the cotyledons of the
second hybrid generation ; the skins of the two peas
to the extreme left of the upper pod have been removed
in order to show the yellow and the green of the
cotyledons of the second hybrid generation in juxta-
position to one another. In the two pods shown there
are eleven yellow and four green — as close an
approximation to a ratio of three yellow to one green
as can obtain amongst fifteen seeds.

The colours of “ the seeds ” borne on a plant
of the first hybrid generation which has been pro-
duced by a cross between a yellow-seeded and green-
seeded pea are simply the colours of the first two
“ leaves ” or cotyledons of the plants of the second
hybrid generation.

I have endeavoured to illustrate this — and,
incidentally also, the general fact with which the
reader is now familiar, that the colours of “ the
seeds ” (whether yellow or green), borne by a plant,
are the colours of the first two leaves of its children —
by taking a photograph, reproduced in Fig. 18, of a
pod containing the cotyledons of the second hybrid
generation ; then sowing the seeds in a row in the
order in which they were in the pod, and photo-
graphing the seedlings thus raised (Fig. 19). This
illustration shows, I think, sufficiently clearly that
the characters, yellowness and greenness, of cotyledons
are characters which appear at a very early stage in
the development of the plant, so early in fact that
they can be seen in the seed (if the seed-coat be
transparent) before they are sown to produce the

6o

BREEDING

plant. Indeed, not only can the colour of the
cotyledons be seen before the seed is sown, it must
be seen then, or not at all, inasmuch as it cannot
be seen for long after the seed has begun to germinate,
because as the growing plant absorbs the food-
material in the cotyledons, their colour fades. But
it could still be clearly seen in the seedlings represented
in Fig. 19.

Stated in general terms, therefore, the second
hybrid generation consists of three plants with yellow
cotyledons, and one with green, in every four. In
practice this result is seen in the existence of yellow
and green peas in the ratio of three to one in the
pods of a plant of the first hybrid generation. In our
imaginary cross made in the summer of 1910 the
numbers of these yellow and green seeds could be
counted in the autumn of 1911. These seeds would
be sown in the spring of 1912. It would be found
that the green seeds produced plants on which all
the seeds were green ; that one of every three yellow
seeds would produce a plant on which the seeds were all
yellow, and that the remaining two yellows would
produce plants which would bear yellow and green
seeds in the ratio of about three to one. Stated in
more general terms, this means that of every four
plants (on the average) of the second hybrid genera-
tion : one (recessive) plant with green first two

“ leaves ” gives rise solely to plants like itself when
it is allowed to self-fertilise ; two (hybrid) plants
with yellow first two leaves will produce plants with
yellow, and plants with green first two leaves, in the

G Y G Y Y G Y

Fig. 18.— the seven SEEDS WHICH GAVE RISE TO THE
SEVEN SEEDLINGS IN FIG. 19, IN THEIR NATURAL
POSITIONS IN THE POD

Colours of the seeds. G Green ; Y Yellow.

Fig. 19.— the seven seedlings RAISED FROM THE SEVEN SEEDS SHOWN IN THE

POD IN FIG. 18

Colours of the cotyledons of the seedlings. G Green; Y Yellow.

INHERITED SEED-CHARACTERS 6i

ratio of three to one ; and one (dominant) plant with
yellow first two leaves will give rise solely to plants
like itself. The mode of inheritance of the colour
(whether yellow or green) of the first two ‘‘ leaves,”
or cotyledons, of the plants is therefore precisely
analogous to that of the stature (whether tall or
dwarf) of the plant. And if the reader wishes to see
the inheritance of the colour of the cotyledons set
forth in diagrammatic form, all that is necessary
is to substitute yellow for tall and green for dwarf
in the frontispiece.

But in order that the way in which this result
appears on the plants themselves may be under-
stood, I give, on the next page, an actual instance
of the total numbers of yellow and green “ seeds ”
constituting the second hybrid generation ; and also
of the total numbers of yellow and green “ seeds ”
constituting the third hybrid generation.

The crosses were made by me in 1905 ; the
second hybrid generation (which I give here)
was recorded in the autumn of 1906 ; and
the third hybrid generation was recorded in the
autumn of 1907. By “ plant number ” is simply
meant the number by which a particular plant was
named in the records of my experiments. In the
first column to the left of the Table are given the
number of yellow and green seeds constituting the
second hybrid generation. It will be seen that every
one of the twenty-six plants bears both yellow and
green seeds ; in other words, every plant, of the
first hybrid generation, produces, after self-fertilisa-

62

BREEDING

spns IL99J^
/o -oj^-

7

13

14

15
3

15

sp»3s atoilix
So 'OH

OC W wow Tj< CO 00 W Q

W cow CO -H OS -HO cot-® CO

MqvinH

turnJ

8-6

10-6

11-6

146

166

166

18-6

19-6

22-6

23- 6

24- 6

26-6

SpdiS U39iQ

, fo ’OH

WOOS ^ o w w o w CO i-i w eo t-

W CO CO -H w ^ W -H ^ ww

sp39S aioipx
So -OH

t-OO t- CO -H W W 05 W 00 t- W W OS t- OS OS o «o

eo O 05 w O O O O 05 CO ^ W O W W CO W O W -h C<l

i

jaqvinH

iuvij

WWW WWWWWWWWWWWWWW w w w w w

fH w cb cb (» ds 6 fH w eb w CD cb OS S iC iS S

THfHtHrHrHiHfHiHiHiH WWWWW

1

SpddS U3dJ^

So ’OH

-H Tft 00 W W 00 OS o 00 W -H 00 -H 00

W-HCO W N W-H^-HW -H -H

1

spd9S (nojtsA
So ’OH

COCO®®-H®t-OSOSOOSt-OSW®WQOSt-OOWWt-0-^t-
W Ht W CD W OS CO W Tt< 00 ® W ® ® W CO CO O t- CO CO OS

I— 1 eo —H 1-H f-H

g

jaqynnH

iurrij

»HWcbHchwcbb-cbds6tHwebT»<wcbt^cbds6fHweb^wcb

rHtHrHr-liHr-IrHiHrHr-IWWWWWWW

5

Sp9»S U99Xf)

So ’OH

® CO o 00 ® W ^ CO 00 W OOOSWCOC-®« t-

cq —1 ^ W CO eo W ^ WWW ^ <

§

sp99S (H02J9A
So ’OH

coo-^o— <'H*^t-t-osw-HWco-H'^eo®-Ht-'^-Hcowt-®o
O5eowwwwwoso5t-w®co'<**'^eo®— <t-®oeo®OHi<Ti<

rH ^ 1— * p— 4

1*

1

xoqvmH

tuvij

cpepeococoeocoeoepcpcoeoeoeoeocoeoeoeocoepeoeoepcpeo

T^web^wcbt^cbdsoTHWeb^wcbt^dbdsoiHwm^wcb

iHiHiHiHiHiHtHiHiHr-lWWWWWWW

1

Sp99S U99XQ

So ’OH

sp99S aioipA
So ’OH

W ® 00 W 00 00 ^ t- ® 00 ® W O ® -H o CO OS F-

CO 1— l«-H fHfH^f—C^f— i-H C^I— <f-4fH I— (

05 0ooocoeowTj(Qoo-HQeO'^oow®'^wt-t-®oo®'^t-t-
■hJi ^ o 00 CO eo ® CO w w ® t- CO eo w w f- 00 w -h ® eo w

fH 04 I-H i-H f-H

x9qvinH

i'md

WWWWWWWWWWWWWWWWWWWWWWWWNW

$p99S "U99XO
So ‘OH

w ® eowww-HQo— <-^w-^ ® F.-COWW ©os

CO W -H F-< <M (M (M ^ c<l -H -H Tt f-H W CO

SP99S 0M\PA
So ’OH

OQl-Q©W©^F-.eOOW®eOHi<<35t^W©®©©W©OOCO

eo®t-woocow®wwo5rt4eocoeo®^-HO©coeo©F-— <eo
i-H w 1— H i-H ^ 1-^

x9qvinH

tuvjj

THe«eb^wcbi>cbd>6fHweb^wcbi^cods6fHweb^wcb

rHrHr-lrHrHrHiHrHi-4T-)WWWWWWW

fill

ep»9g U99JL0
So -OH

Wrf<FH04Tj<OOHj<<»OSW®WCOOF-F-HWeO©t-W'^eO©©Tl<
f-Hi— <f-Hf-H f-H(Mf-H f-<i-HfH<Mf-If-H t—i WCO^W

(S5 P

^I’^i

.3-2‘S'^
a, ^«o

sp99g ouri2»A
So ’OH
j.9qvtnH

COt-®©-Ht-QOWOO-HfW®®'Tt<WW©'H*<CO®CO-HW®t-

wweowweo®®wW'^Hteo®weocowweo-H^ooooeO'^

r-IWeOHl'W®t-«©O^NeOH^WCO^®©OFHgQ^^^g

INHERITED SEED-CHARACTERS 63

tion, a family consisting in every case of plants with
yellow, and plants with green first two leaves.

To return to common parlance. Either five or
six of the yellow seeds constituting the second hybrid
generation were sown, and the colours of the seeds
(containing the third hybrid generation) borne on
the plants raised are given on the same line in the
Table, to the right of the record of the seeds of the
second hybrid generation. Thus, five of the yellow
seeds borne on plant No. 1 were sown and gave rise
to five plants numbered 1*1, 1*2, 1*3, 1*4 and 1*5,
all of which happened to produce both yellow and
green seeds. In other words, all of the yellow-seed
leaved plantlets contained within the seeds sown,
were hybrids. But of the five yellow seeds of
Plant No. 2 that were sown, two — ^namely, those which
produced plants numbered 2*1 and 2*3 contained
pure dominant plantlets because the plants raised
from them produced only yellow seeds. The import-
ant point to notice is that every grown plant of the
first hybrid generation produces both yellow and
green seeds containing the second hybrid generation
in embryo {see the first column in the Table) ; i.e.
segregation or reappearance of the parental charac-
teristics occurs in every family in the second hybrid
generation. On the other hand, some of the grown
plants of the second hybrid generation — namely, the
hybrid ones — produce both yellow and green seeds
(containing the third hybrid generation in embryo),
whilst others, the pure dominants, produce yellow
seeds only.

64

BREEDING

If the green seeds containing the cotyledons of
the second hybrid generation had been sown they
would produce greens only ; but as I did not test
the true breeding of these so-called “ extracted ”
greens in this particular experiment, the fact that
they breed true is not illustrated in the Table on

p. 62.

Fig. 20 shows the distribution of yellow and
green seeds borne by Plant No. 12, which was
selected at random. Yellow seeds are indicated
by O and green ones by o. It will be seen that
no order can be detected in the distribution of the
seeds of the two colours.

This diagram brings home to one’s mind the fact
that the colours of the seeds of a plant, when we
mean the colours of the contained cotyledons, are
the colours of the offspring (or, rather, of the first
two leaves of the offspring) of that plant ; whilst
the colour of the seed-coats of the seeds borne by a
plant is the colour of a part of the plant itself, just
as the colour of the pods is, or the colour of the
stem. That is why the colours of the seed-coats
(whether grey or white) of the seeds borne by a
plant are always the same (see Fig. 11 or 12), and why
the colours of the cotyledons of the seeds- borne by
a plant need not be always the same (see Plate I.,
facing p. 58). And the apparent paradox that “ the
seed ” presents a pair of characters (grey and white)
which are the latest to appear in the life history of
the plant, and another pair (yellow and green), which
are the earliest to appear, is explained by the fact that

Fig. 20.— diagram TO ILLUSTRATE THE DISTRIBUTION
OF YELLOW AND GREEN SEEDS IN PLANT No. 12 IN
TABLE ON PAGE 62

(Yellow Peas are indicated by full black; i(reen ones by outlines.)

INHERITED SEED-CHARACTERS 65

the seed-coats (which may be grey or white) are not
developed till the plant is mature, whilst the cotyle-
dons (which may be yellow or green) are among the
first parts of the plant to be developed.

I have been at pains to make this point clear,
partly because I am concerned to bring the actual
facts discovered by Mendel home to the reader, and
partly because the cotyledon characters of Pisum are
the most valuable that I know for illustrating the
Mendelian phenomenon and for testing the truth of
general statements made with regard to it. For it
must be plain already that one of the features by
which the Mendelian phenomenon may be recognised,
when it occurs in a breeding experiment, is the definite-
ness of the ratios in which the parental characters
reappear or segregate in the second hybrid generation.
It is therefore desirable, in critical experiments of
this kind, to use material which will furnish the
maximum number of individuals at a minimum
expense, and in a minimum space. That end is
attained by selecting characters, to experiment with,
which appear as early as possible in the life history
of the animal or plant which bears them, and this
condition is better fulfilled by the characters of
the cotyledons of Pisum than by any other that I
know. A plot of land fifteen yards square will
bear 100,000 seeds.

The actual numbers which Mendel himself obtained
in his experiments with the colour of the cotyledons
are as follows : He made fifty-eight crosses on ten
plants ; and found the yellow colour of the cotyledon

F

66

BREEDING

to be dominant over the green in every case. The
second hybrid generation was distributed over 258
plants, and consisted of 6,022 yellows and 2,001
greens ; a very close approximation to the three-to-
one ratio. Mendel gives the distribution of yellow
and green seeds on ten plants in his paper ; it was
as follows : —

Y.

G.

Y.

G.

1.

25

11

6.

20

6

2.

32

7

7.

32

13

3.

14

5

8.

44

9

4.

70

27

9.

50

14

5.

24

13

10.

44

18

Of the total number of individuals (i.e. seeds) which
composed the second hybrid generation (namely,
6022 + 2001 = 8023) only 519 yellow- cotyledoned
ones were tested : 166 produced yellows only, i.e.
were pure dominant yellows ; 353 produced yellows
and greens in the ratio of three to one, i.e. were
hybrid yellows. In other words, 519 of 8023 seeds
(which contained the second hybrid generation) were
sown, 166 gave rise to plants which produced yellow
seeds only, 353 gave rise to plants which produced
yellow seeds and green seeds in the ratio of tliree to one.

We will now proceed to a consideration of the
mode of inheritance of the shape of the cotyledons
in Fisum, The pair of characters in this case is
round (which is dominant) and wrinkled (which is

INHERITED SEED-CHARACTERS 67

recessive). The inheritance of these characters is
precisely analogous to that of the colour of the
cotyledons ; and a diagrammatic representation of it
may be made by substituting round for tall and
wrinkled for dwarf in the Frontispiece.

The shape of the seed (whether round or wrinkled)
is determined by the cotyledons, and not by the
seed-coats ; we may therefore say that the pair
of characters is roundness of the first two leaves of
the plants and wrinkledness of these structures.
The shape of the cotyledons, therefore, is just as
valuable a character, for the reasons stated above,
as their colour, to experiment with for the purpose of
testing general statements with regard to the Men-
delian phenomenon. But over and above this it
possesses an interest which the colour of the cotyledons
does not. In the first place, the distinction between
round and wrinkled peas is one of great economic
importance, the nature of which will be fully discussed
later ; and in the second place, it is possible in the
case of this character to see below the surface, as it
were, of the Mendelian phenomenon, and thus to
obtain a truer view of the essential nature of this
process. But at present we are concerned with the
phenomena as they present themselves to the senses
unaided by the microscope and the scales. Fig. 21
represents the result of crossing of round-seeded with
wrinkled-seeded pea. The round cotyledons of the
first hybrid generation do not differ, so far as the
unaided eye can see, from those of the round
parent.

68

BREEDING

Mendel himself made sixty crosses between round-
seeded and wrinkled-seeded varieties on fifteen plants.
The second hybrid generation consisted of 7,324
seeds, of which 5,474 were “ round or roundish,”
and 1,850 were wrinkled ; 565 of the rounds were
tested 193 gave rounds only, 372 rounds and
wrinkled in the ratio of three to one.

The seven pairs of characters of the culinary pea
experimented with by Mendel have now been dealt
with. In all of them the mode of inheritance is
essentially the same. Two organisms differing in
respect of a single pair of characters produce,
when mated, a hybrid, which manifests the so-
called “ dominant ” member of that pair to the
more or less complete exclusion of the recessive one.
Dominance, however, is an unessential feature of
Mendelian inheritance. Mendel himself stated that
one member of each of his seven pairs was completely
dominant over the other of that pair. But it is now
known that, in the case of the texture of the pod, the
hybrid between the hard and the soft bears pods of
intermediate texture. And we shall see later that,
in the case of the very first pair of characters in
Mendel’s list, the round cotyledons of the hybrid
produced by crossing a round with a wrinkled-seeded
pea are only superficially indistinguishable from
those of the pure round ; and that a fundamental
difference between them can easily be demon-
strated.

The essential feature of the Mendelian phenomenon
is that which the mode of inheritance of the seven pairs

Fig. 21.-MENDELIAN INHERITANCE OF THE SHAPE OF THE
SEED IN THE CULINARY PEA

(Top left) Wrinkled Parent. (Top right) Round Parent.

(Four Peas in middle line) First Hybrid Generation.

(In the pods) Second Hybrid Generation.

INHERITED SEED-CHARACTERS 69

of characters dealt with by Mendel, and that of the
colour of the Andalusian fowl and the human eye,
possess in common, i.e. that in which the genealogical
table of the tall and dwarf peas and that of the
Andalusian fowl agree, namely the orderly reappear-
ance of the characters of the parents of the
hybrid, and of the hybrid itself, in the second
hybrid generation in definite numerical propor-
tions.

There is another characteristic of the Mendelian
phenomenon which can only be said to be an essential
one in the sense that it follows from Mendel’s theory*
to account for the composition of the second hybrid-
generation. This characteristic is the repetition,
in each successive hybrid generation, of the ratios
obtaining in the second. The second, and all
subsequent hybrid generations, should all have
exactly the same composition as one another, if
this characteristic is found to be an invariable feature
of Mendelian crosses. Let it be clearly understood
what this means. It means that, for however many
generations hybrid yellow peas are bred from, they
will produce pure yellows and hybrid yellows and
pure greens in the ratio of 25 : 50 : 25 per cent. What
is most remarkable in this generalisation, because
most at variance with current notions of heredity,
is the idea that although the parent of each successive
hybrid generation is as yellow as a yellow of a pure
strain, there will be produced an average number
of green, namely, 25 per cent., which will not

* See Chapter XL

70

BREEDING

diminish as the generation in question becomes more
remote from the green parent of the cross. In other
words, it is a remarkable thing that a yellow oi the
tenth hybrid generation, with nine generations consist-
ing solely of yellows behind it, should contain as
many greens as the second hybrid generation, which
has only one such generation behind it. There is no
possibility, here, of eradicating the green by breeding
only from the yellow ; unless pure yellows are bred
from. The evidence, as yet available, as to the ratios
of recessives, in the case both of cotyledon colour
and shape, has been collected by Mr. Lock in his
useful paper, “ The Present State of Knowledge of
Heredity in Pisum,^^* and is as follows. The name
of the investigator is given in the case of each separ-
ate result : —

Cotyledon Shape

Hybrid

Oeneration

Observer

Round

Wrinkled

Percentage

of

Wrinkled

Second •

Mendel

5,474

1,850

25-2

Tschermak

884

288

24-6

Bateson

10,793

3,542

24-8

Hurst

1,335

420

23-9

Lock

620

197

241

Third •

Tschermak

2,087

661

24-0

Lock

769

259

25-2

Fourth .

Lock

2,328

812

25-8

* Annala of the Royal Botanic Gardens, Peradeniya, Vol. IV., Pt. iii.

INHERITED SEED-GHARAGTERS 71

Cotyledon Colour

Hybrid

Oeneraiion

Observer

Yellow

Green

Percentage
of Green

Second .

Mendel

6,022

2,001

24-9

Correns

1,394

453

24-5

Tschermak

3,580

1,190

24*9

Bateson

11,903

3,903

24-7

Hurst

1,310

445

25-4

Lock

1,438

514

26-2

Third .

Correns

1,012

344

25-5

Tschermak

3,000

959

24-2

Lock

3,082

1,008

24-6

Fourth .

Correns

225

70

23-7

Lock

2,400

850

26-1

These figures show how closely the ratios of
recessives observed by different investigators
approximate to 25 per cent. ; but they also show
that beyond the fourth hybrid generation we have
no evidence for saying that this ratio is repeated
in the case of these two characters of the cotyledons
in Pisum.

CHAPTER VI

THROWING BACK, OR REVERSION

In the crosses which have hitherto been dealt with,
the result of mating two individuals which differed
in respect of a particular characteristic has been a
hybrid which has resembled one of the parents so
closely as to be indistinguishable from it ; or it has
been intermediate between them, as in the texture of
the pod in peas, and the colour of the feathers in the
fowl. We shall now proceed to cases in which the
hybrid differs from either parent, but is not inter-
mediate between them. Into this class fall those
cases in which the result of mating two varieties is
the production of the features of the wild ancestral
form from which the two varieties are supposed to
have descended. A very characteristic instance of
such a result is afforded by the result of crossing
the ordinary albino mouse with the so-called Japanese
waltzing mouse. These two varieties differ from
one another in respect of their colour, and in respect
of their customary movements ; the one walking and
running normally, the other exhibiting the so-called
waltzing movements which are indicated by its
name. We are concerned at present solely with the
colour of the two forms ; the question of the mode
of progression must be banished from the reader’s

72

THROWING BACK

73

mind until it comes to be considered separately.
The colour of the two varieties crossed will be first
briefly considered.

The common albino mouse (Plate II., No. 1) has
a pure white coat ; there is no pigment in it at all.
The tail, hands, feet and ears are very pale pink,
and the eyes are a deep pink, the colour of these
organs being due to the blood in them.

The Japanese waltzing mouse, with which I have
experimented, is coloured exactly like the albino
except that it possesses patches of fawn-yellow fur
on its shoulders and haunches. That is to say, it is
a fawn-and- white piebald with pink eyes (Plate II.,
No. 2). The extent of these patches of fawn-yellow
fur varies very little from individual to individual
in this variety.

A hybrid produced by mating these two varieties
is shown at No. 3. The coat is a dark grizzly grey,
hardly distinguishable from that of the house mouse ;
and the eyes are jet black. The tail may be either
deeply pigmented over its whole extent or, as in the
specimen shown, the dark pigment may not extend
over the whole of it. The hybrid shown at No. 3
is the usual result of mating mice Nos. 1 and 2. I
have, however, observed rare exceptions to it, but
these are usually attributable to the impurity of the
albino stock used, and need not detain us now. It
makes no difference to the result whether the albino
is the male and the fawn-and- white mouse the female
parent, or the albino is the female and the fawn-and-
white the male.

74

BREEDING

The result, then, of crossing a pure albino mouse
with a fawn-and-white, pink-eyed (waltzing) mouse is a
mouse which differs very little from the common house-
mouse. As a matter of fact, it is usually paler beneath
than, the house mouse is ; and the wild race which
it most resembles exists in St. Kilda, an island far
out in the Atlantic, to the west of Scotland. Be
this as it may, the result is a typical instance of
reversion on crossing. These reversionary mice con-
stitute the first hybrid generation.

When these hybrids are mated together they
produce a generation (the second hybrid generation)
which consists of a variety of forms which can be
classified in three groups (A, B, and C), which corre-
spond roughly to the albino, the fawn-and-white, pink-
eyed (waltzing) mouse, and the hybrid. Let us
examine them more closely in conjunction with the
figures of them on Plate II. For the present, mice
Nos. 4, 6 and 10 will be left out of account, and
attention will only be paid to those individuals in
each group which are below the label indicating the
group.

Mice included in Group A correspond exactly to
one of the parental forms, namely the albino ; there
is, therefore, only one mouse (below the label) in
Group A. These mice are indistinguishable from the
pure albino mouse.

Mice included in Group B correspond to the
reversionary forms which constitute the first hybrid
generation ; but the correspondence between analo-
gous types in the two generations is much less

THROWING BACK

75

exact than in the case of the albinos. In the first
place the colour may be grey or black ; in the second
place there may be patches, of greater or less extent,
of white fur, as well as the grey or black in the coat ;
various degrees of piebaldness being thus produced.
There may therefore be, roughly, four kinds of mice
in Group B : mice which have a completely grey coat ;
(No. 7) — self-coloured greys as they are called ; self-
coloured blacks (No. 8) ; piebald greys (No. 9) ; and
piebald blacks which are not figured. Amongst the
piebalds there may be every gradation between a
mouse which, but for the possession of a few white
hairs (on the forehead usually), would be a self-
coloured mouse, to one which has no more than
a small patch of grey or black, which is usually
near the ears. The extreme form of piebald in
this direction is a white mouse with black eyes ; but
this type has never occurred in my experiments.
There are also differences in the intensities of the
grey and black, so that the number of possible
different types of coloration in the second hybrid
generation is very great indeed. All the members of
Group B, however, possess the following characters
in common : (a) black eyes ; and (6) a greater or
less amount of a dark colour, either black or grey,
in the coat.

Mice included in Group C correspond to the
fa wn-and- white (waltzing) pink-eyed mice. But here,
again, the correspondence is not exact. For just as
in Group B, which corresponds roughly to the first
hybrid generation, there were blacks as well as greys.

BREEDING

76

so in Group C, which corresponds roughly to the fawn-
and- white mice (No. 2), there are “ lilacs ” as well as
fawns. The parallel, indeed, between the two cases
is very close, for the fawn colour may be regarded as
a dilute grey ; and the lilacs as a dilute black.
“ Lilac was the name I first gave to these mice,
and I have adhered to it, but “ pale lavender ” more
accurately conveys the exact colour. Then, again,
as in Group B, there may be “ seifs ” as well as
piebalds. But whilst in Group B the piebald was the
new appearance, in Group C the “ self ” is. In
other words, in the case of the dark colours (grey and
black) the “self” appears first in the first hybrid
generation, and the piebald does not appear till the
second ; but in the case of the paler colours (fawn
and lilac) the fawn exists first in a piebald form in
one of the parents of the cross, and does not appear
as a “ self ” until the second hybrid generation. •

There may, therefore, be — as in Group B — four
kinds of mice in Group C : self-coloured fawns (No. 11),
self-coloured lilacs (not figured), piebald fawns (No. 12),
and piebald lilacs (No. 13, which has a white face).

The second hybrid generation is made up of the
representatives of the three Groups, A, B and C, in
the ratio of about 25 per cent. A, 50 per cent. B,
and 25 per cent. C, or, roughly, one of A, two of B,
and one of C in every four, on the average. These
ratios obviously do not obtain in every litter, even
if only because the number of young in a litter is
not necessarily a multiple of four ; in fact, in none
of the twenty litters, selected at random and shown

THROWING BACK

77

on the following page, did the exact ratio occur.
The total numbers obtained in my experiment were :
Group A, 137 ; Group B, 287 ; Group C, 134.

Now let us consider the breeding properties of
the mice representative of the three groups, A, B, and
C. The albinos (Group A) breed true without excep-
tion. With rare exceptions, the mice in Group B
again produce representatives of Groups A (No. 15),
B (Nos. 16 and 17), and C (Nos. 18 and 19), when
they are mated together ; the exceptions referred to
do not produce albinos or fawns, but only mice with
black eyes and dark coats — ^that is to say, they breed
true to the characters of Group B. The mice in
Group C, with very rare exceptions indeed, breed
true to the characters of their group.

The breeding properties of the various colour
types within the three Groups, A, B, C, are as follows,
so far as my published records permit me to judge :
Black by black do not give grey ; but grey by grey
may give black, as also may grey by black. Lilac
by lilac do not give fawn ; but fawn by fawn have
not yet given lilac, nor have lilac by fawn. Self-
colour by self-colour occasionally give piebalds, but
piebald by piebald do not give “ seifs,” and piebald
by “ seifs ” give both piebalds and ‘‘ seifs.” It is
desirable now to pause and to consider the pheno-
menon represented on Plate II. as a whole.

The parallel between this phenomenon and that
exhibited by, for instance, tallness and dwarfness
in the pea, or by the colour of the Andalusian fowl,
is at once obvious. Albinism may be called the

78

BREEDING

Table showing the Colour-Characters of the
Mice in Twenty Litters of the Second
Hybrid Generation

Litter

No.

Oroup A
{albinos)

Group B

{Dark-coloured mice
with black eyes)

Group C

{Pale-coloured mice
with pink eyes)

Total

1

none.

grey, grey, grey, black.

piebald fawn, piebald
fawn.

6

2

albino, albino.

piebald grey, piebald

fawn.

6

3

albino, albino.

grey.

piebald grey, piebald
grey, grey, black.

none.

6

4

albino.

grey, grey, grey, pie-
bald grey.

piebald fawn.

6

6

none.

grey, grey, black.

fawn, piebald fawn.

6

6

none.

grey, grey, grey, grey,
piebald black.

fawn, lilac.

7

7

albino, albino,
albino, albino.

grey, piebald grey.

none.

6

8

albino, albino.

grey, grey.

none.

4

9

albino.

grey, grey, piebald
grey, black.

none.

5

10

none.

grey, grey.

fawn, fawn, fawn.

6

11

none.

grey, piebald grey,
piebald grey.

piebald fawn, piebald
fawn.

6

12

albino, albino

grey, piebald grey.

piebald lilac.

6

13

albino.

grey, grey, CTey, grey.

fawn.

6

14

albino.

grey, piebald grey.

fawn, fawn, fawn.

6

16

albino.

grey, grey, grey, grey.

piebald fawn.

6

16

albino.

grey, grey, grey, pie-
bald grey, piebald
grey, piebald black.

piebald fawn.

8

17

none.

grey, grey, piebald grey.

lilac.

4

18

albino, albino.

piebald grey, piebald

fawn, lilac, piebald lilac.

7

19

albino, albino.

gr^,^grey, grey,
grey, black.

fawn.

6

20

albino.

fawn, fawn, piebald
fawn.

6

The mode of progression of the mice, whether waltzing or normal, is left out
of account in this Table.

THROWING BACK

79

recessive character, the “ fawn-and- white ” type of
coloration may be called the dominant ; although
it may be argued that, in a sense, it is just as recessive
as albinism inasmuch as it disappears from the first
hybrid generation as much as albinism does ; but it
is convenient to speak of it as the dominant character,
and it is justifiable if we state the case by saying
that in the cross shown on Plate II. “ some colour ”
is dominant over “ no colour ” ; further, if the white
Andalusian is regarded as the recessive it is legitimate
to regard the white mouse as recessive, and, if this
is done, the fawn-and-white falls into its place as
the dominant.

In all three cases — the pea, the fowl, and the
mouse — ^the two characters (of the two parents
respectively) reappear in the second hybrid generation
of which they each make up 25 per cent. ; the remain-
ing 50 per cent, being composed of individuals
resembling the hybrids of the first hybrid generation
in all three cases. The parental characters, when they
have reappeared in the second hybrid generation,
breed true without exception in the case of the
so-called ‘‘ extracted ” recessive ; and, with the
extremely rare exceptions afiorded by the mice in
Group C, they also breed true in the case of the
“extracted” dominant. And in all three cases the
hybrids of the second hybrid generation again produce
the dominants, hybrids and recessives with the
exceptions afforded by the true-breeding mice in
Group B.

We are thus again brought round to the conclusion

8o

BREEDING

that the character borne by the hybrid is not essential
to the Mendelian phenomenon ; in the three cases
which we are considering the hybrid has been
either (a) indistinguishable from one parent, or (h)
intermediate between the two, or (c) it has borne
the character of the remote ancestral form of both.
What is common to all three cases is the reappearance
in the second hybrid generation of the “ dominant ”
parental character, the hybrid character (when
different from this) and the “ recessive ” parental
character in the ratios of 25 per cent., 50 per cent,
and 25 per cent, respectively — a phenomenon to
which the term segregation has been applied.

The fact that the hybrid frequently bears
characters which are peculiar to itself, and the
occurrence of segregation, are the two features of
the results of crossing which lead to the production of
novelties ; and they correspond, in the main, to two
distinct practical methods. In the first of these
methods the novelty is obtained in the first hybrid
generation by virtue of the fact that the hybrid bears
characters peculiar to itself. The hybrid in these cases
is usually not reversionary, but intermediate between
the two parents, as in the case of the Andalusian
fowl, and that of the roan colour in cattle, which is
produced by crossing red with white, and which, when
m.ated with roan, gives 25 per cent, red, 50 per cent,
roan, and 25 per cent, white. The novelty which
arises in this way is unstable.

In the second of these practical methods the
novelty is not obtained till the second hybrid genera-

THROWING BACK

8i

tion, the first hybrid generation consisting usually
of reversionary forms. In this class of cross nothing
must be expected from the immediate result of the
cross. Nothing, for instance, could be more dis-
appointing to the mouse fancier than the result of
crossing the albino with the Japanese waltzing mouse
— namely, an animal scarcely distinguishable from one
that could be caught in a trap, any day, in the pantry.
Yet I know no more beautiful colour in an animal
than that of the “ lilacs ” which are produced in
small numbers by mating these hybrids of the first
generation together. I am given to understand that
this colour is new to the “ fancy,” but I do not
know if there is a class for it at the shows.

It is very important to keep these two ways in
which novelties are obtained by crossing, distinct in
the mind. In the “ first generation method,” as it
may be called, an unstable new form is the immediate
result of the cross ; and to obtain it again it is better
to repeat the cross than to breed these hybrids
together, partly because by this means 100 per cent,
of the offspring are of the desired kind (as opposed
to the 50 per cent, produced by mating the first
crosses together) ; and partly because of the greater
vigour of first crosses. In the “ second generation
method,” as it may be called, a stable new form is
obtained in the second hybrid generation ; the hybrid,
which is usually reversionary, is useless in itself, and
merely serves as a sort of mill which will turn out
any desired quantity of the new forms.

The two methods are like the two chief methods

82

BREEDING

of photography — the method of Daguerre and the
method of Fox-Talbot, who invented the negative.
By Daguerre’s method a positive image was produced
on a silvered surface, and the picture could not be
copied except by an elaborate process of electrotypy.
A new daguerreotype could only be reproduced by
exposing a fresh plate. The negative, like the
reversionary hybrid, is useless in itself ; but it revolu-
tionised photography. Daguerre’s might be called
the one -generation method, and Fox- Talbot’s the
two -generation method of photography. A breeder
who threw away his first crosses because they did
not possess the character he was working for, would
be as foolish as a photographer who threw away
his negatives because the light parts in the objects
appeared dark, and the dark light, in them.

First crosses are, however, not always bred because
they possess new characteristics, but because they
very often possess greater vigour than either of the
parents crossed. This is another reason why it
is better to raise those hybrids which possess
characteristics peculiar to themselves, such as roan
cattle, by repeating the cross than by breeding roan
cattle inter se, I do not know, and it does not much
matter, whether the excessive vigour of the first cross
is to be considered as a phenomenon of reversion,
but it is a fact that, in the case of my mice, reversion
to the ancestral condition in regard to disposition
is just as invariable a result of the cross as the rever-
sion in regard to colour. The difference between
the disposition of the hybrid and that of the

THROWING BACK 83

albino is as great as that between the colour of
the two.

Though I did not make pets of my albinos, they
were perfectly tame, and would very seldom attempt
to escape from the cage when it was opened. But
the greatest caution had to be exercised when a
cage containing first crosses was opened, for if care is
not taken the mouse will spring from the nest directly
the lid is lifted, and be lost. I lost one or two mice
in this way before I learnt this characteristic of the
first crosses. But even if the mouse does not escape,
the wildness of its disposition is manifested in the
frenzied way in which it darts about the cage when
it is subjected to treatment (such as mating up, or
the separation of the males from the females in a
family when they become adult) under which the
albino behaves with perfect tameness. The excessive
vigour of the first cross is manifested also in the
glossy condition of its coat, which I have always
noticed to be much sleeker than that of any of the
other mice — its parents or its offspring — in my
experiments.

The phenomenon of the inheritance of colour in
these mice has now been described ; and the appli-
cation of the knowledge of this and the other results
which we have discussed, has been briefly referred
to. The luxuriant “ condition ” of the hybrid has
been illustrated, and this reference to a physiological
character leads naturally on to the consideration of
another character in respect of which the two parent

84 BREEDING

forms differ — namely, the nature of their customary
movements.

It is not necessary, I think, to describe the
walking and running movements of the common
white mouse. Moreover, it would be difficult to give
a description of its demeanour which would be gener-
ally true, because the habits of the individual mouse
is determined to a very great extent by the treat-
ment to which it has been subjected. For instance,
the habits of one of my mice, more than a thousand
of which were kept in a room whilst the breeding
experiment was in progress, would be very different
from that of the ewe lamb of the schoolboy. The
characters which ‘‘ tame ” mice possess in common
are sufficiently familiar to the reader. They can
run straight away if free to do so. But the so-called
waltzing mouse cannot.

The actual waltzing itself, which gives the breed
its name, does not, of course, resemble waltzing, and
is not executed on the hind legs only. A better
name, which is sometimes given it, is “ spinning.” The
animal runs round and round in a small circle, the
diameter of which is about half the length of the
animal’s body excluding the tail, at so great a pace
that the mouse becomes a blur. No. 2 on Plate II.
represents a waltzing mouse in the act of waltzing,
with great faithfulness. This excellent representation
of the waltzing habit is due to the skill of Mr. Gerrard,
of Camden Town, who kept some of the waltzing
mice under observation for a considerable time before
stuffing them. An individual mouse does not

THROWING BACK 85

always spin in the same direction, but sometimes
with the hands of the clock, and at other times in
the contrary direction. A “ waltzing ” mouse is not,
however, always spinning. It spins at night, as a
rule, and sleeps during the day when it is not
being fed. But it can always be recognised as a
“ waltzer,” even if it is not spinning, so long as it
is awake, by the following characteristics : first, an
apparently complete lack of control over the move-
ments of its head, which is thrust up and down and
moved rapidly from side to side in a peculiar manner
which cannot be forgotten, but is not easily described ;
secondly, by a curious habit which it has, when
placed in a large open space, such as the floor, of
backing very vigorously as if it were on a slow-
moving avalanche, and moving its head, which is
kept near the floor, rapidly from side to side. Not
only can the adult “ waltzer ” be recognised as such
when it is not waltzing, but the young can also be
recognised before they can waltz, because they can
hardly keep on their four legs. The waltzer is
characterised by great delicacy and stupidity. The
death-rate is higher amongst them than amongst
ordinary tame mice ; they often fall out of the cage
when the door is opened, and are greatly inferior
to the albino, for instance, in the care of the brood.
They seem to be deafer than ordinary mice ; at any
rate, they take fright much less readily at a small
noise than an ordinary mouse does. For instance,
if a noise is made with the tongue and teeth whilst
a waltzer is at food, he will not, as the albino probably,

86

BREEDING

and the hybrid certainly, will, scuttle back into the
hind part of the cage, if this is divided off by a parti-
tion. The waltzer differs from the ordinary tame
mouse in the matter of disposition as much in one
direction as the wild mouse or the hybrid does in the
other. , An albino, if set free, can run away, but does
not ; a hybrid or wild mouse, if set free, can run away
and does ; but a “ waltzer ” cannot escape ; it cannot
keep up a run in direct line for long, and soon lapses
into spinning.

The cause of the spinning is supposed to be an
abnormality in certain organs in close connection
with the internal part of the ear. There are three
of these semicircular canals, as they are called, in
connection with each ear in the normal vertebrate
animal, one horizontal and two vertical for each ear,
and they are said to be concerned in the normal
animal in maintaining the balance of the body. It
was formerly stated that the waltzing mouse lacked
the horizontal canal in connection with each ear ;
but this statement has been proved to be incorrect
by the application of a new method of preparing the
semicircular canals and of preserving them in their
natural state. All six canals are present. But a
deficiency in the nerve supply of the semicircular
canals is said to have been discovered; and the
waltzing habit may be due to this. At any rate, it
seems to be generally agreed that this phenomenon
of waltzing is determined by some abnormality in
the organs which are responsible for preserving the
balance of the animal. A physiologist once expressed

THROWING BACK 87

to me his conception of the causation of the waltzing
habit in the statement that it is due to the absence
of the power of becoming giddy. The unpleasant
sensations which would soon stop a normal mouse
from spinning round are not felt by the waltzing
mouse, and what ultimately brings the bout of
spinning to an end is not nausea, but exhaustion.

Let us, now, turn to the inheritance of the waltzing
character. When two waltzing mice are mated
together the offspring produced are all waltzers.
The character breeds true.

When a waltzing mouse is mated with a mouse
with normal movement, as in my own cross repre-
sented on Plate II., the hybrid produced never
exhibits the waltzing movements. This, at any
rate, has invariably been the case in the hundreds
of hybrids raised by Dr. von Guaita and by myself.
The waltzing character is, therefore, recessive and
normality of progression is dominant. Being a reces-
sive character, waltzing should reappear in the
second hybrid generation in one individual in every
four. It does reappear in the second hybrid genera-
tion, but not in a quarter of the individuals ; there
were 97 out of 555, which is less than a fifth.
But I do not think that this proportion should
lead us to the conclusion that waltzing does not
behave as a Mendelian character in inheritance,
because I believe that the reason that the number of
waltzers falls short of the expected ratio is simply
that waltzers are more delicate constitutionally than
normal mice ; and that relatively more waltzers

88

BREEDING

die before tlie age at which the characters of the
litters are noted, than normal ones do. But to
return to the actual reappearance of the waltzing
character in the second hybrid generation. The
waltzing of the “ extracted ” waltzers, as the
individuals of this generation which manifest this
character are called, does not differ from that of
individuals of the pure race, except that in occasional
individuals it seems to be accentuated.

The waltzing character, when it reappears in the
second hybrid generation, is not confined to mice
coloured like the pure waltzer. No. 2 ; that is to
say, it does not only occur amongst the mice of
Group C, but amongst those of Group A and B
also. Waltzing representatives of each of the three
groups are shown at No. 4, No. 6 and No. 10. We
are now in view of a new and very important
subject. Hitherto we have been considering the
relation to one another of two characters which
stand to one another in the relation of dominant
and recessive — i.e. characters which constitute a
single pair, and affect the same organ or feature of
the plant or animal. We now come in view of the
question : What is the relation between members of
distinct pairs of characters — i.e. characters of distinct
features of the animal or plant — for instance, the
colour of a mouse and its mode of progression ? The
answer in this case is very simple. There is no
relation. The colour of a mouse and the nature of
its movements, i.e. whether normal or “ waltzing,” are
inherited entirely independently of one another. The

THROWING BACK 89

pure race of waltzers breed true to their colour and
“ waltzing ” ; but in the second hybrid generation
produced from a cross between a waltzing mouse
and an albino, there is no association between colour
and the waltzing character. The waltzing character
is distributed at random over the three colour-groups,
A, B and C — i.e. without any preference to any
particular colour-group, as can be seen on page 91
and Plate II. The numerical proportions in
which the various combinations of colour and mode
of progression occur will not be considered in detail
now ; suffice it to say that rather less than a quarter
of the mice in each colour-group. A, B and C, are
waltzers. Our attention may now be turned to
points of theoretical and practical interest in the
results displayed on Plate II.

To consider the facts from their theoretical side
first ; it is seen that the fact of the independent
inheritance of the colour and mode of progression,
which could not be detected by observing the results of
breeding from the pure waltzing race for an unlimited
number of generations, is at once revealed (in the
second hybrid generation, to be precise) by the simple
experiment of crossing two mice which differ in their
colour and the mode of their progression. The
practical interest of such facts is too obvious to need
insisting upon. The number of new varieties in the
second hybrid generation is considerable ; and
some of them may be counted upon to breed true.
The re-combination of characters possessed by the
two parents of the cross may be considered first.

90

BREEDING

The albino waltzer (No. 4) possesses the recessive
character, waltzing, of one parent, and the recessive
character, albinism, of the other parent, and as both
these characters breed true when they reappear in
the second hybrid generation, the new albino waltzer
will be a constant variety from the first. The colour-
ing of tlie pure waltzer and the normal progression
of the albino are combined in mouse No. 12. Mice
in which the waltzing character is associated with
one of the colours or arrangements of colour not
existing in either of the two parents of the cross
constitute a further set of new forms. The most
beautiful of these is undoubtedly the lilac waltzer
(a specimen of which I have not figured), and the
most curious of them is the form shown at No. 6, a
mouse in which the coloration of the first hybrid
generation is associated with the waltzing character
— a house-mouse that cannot run away. Whether it
was a constant variety or not would depend upon
whether, when mated with its like, it bred true to
its colour or not; for it would be certain to breed
true to the waltzing character if mated with a
mouse like it. But as the majority of the mice in
Group B produce, again, representatives of Groups A,
B and C, it is impossible that a constant race of mice
like No. 6 could be raised, except after long waiting
for two mice, of opposite sexes, both of which breed
true to grey coat and black eye.

In this chapter I have set forth an example of
reversion on crossing, and the result of breeding from
these reversionary hybrids. The inheritance of colour

Table showing the Distribution of the Waltz-
ing Character over the Twenty Litters
Exhibited on Page 78

Litter

No.

Group A
{albinos)

Group B

{Dark-coloured mice
with black eyes)

Group C

{Pale-coloured mice
with pink eyes)

Total of
normal
mice

Total of
waltziny
^mice

1

none.

grey, grey, grey,
black.

piebald fawn, pie-
bald fawn.

6

none.

2

albino,

albino.

piebald grey, pie-
bald grey :W.

fawn.

4

1

3

albino :W,
albino ;W.

piebald grey, pie-
bald grey, grey,
black :W.

none.

3

3

4

albino :W

grey, grey, grey,
piebald grey.

piebald fawn.

5

1

6

none.

grey, grey, black.

fawn :W, piebald
fawn.

4

1

6

none.

grey, grey, grey,
grey, piebald
black.

fawn, lilac.

7

none.

7

albino ;W.
albino, albino,
albino.

grey, piebald grey

none.

5

1

8

albino,

albino.

grey, grey.

none.

4

none.

9

albino.

grey, grey, piebald
grey, black.

none.

5

none.

10

none.

grey, grey :W.

fawn, fawn, fawn.

4

1

11

none.

grey :W, piebald
grey :W, pie-
bald grey :W.

piebald fawn :W,
piebald fawn ; W.

none.

5

12

albino,

albino.

grey, piebald grey.

piebald lilac :W.

5

1

13

albino.

grey, grey, grey.

fawn.

6

none.

14

albino.

grey.

grey :W, piebald
grey.

fawn :W, fawn,

fawn

4

2

15

albino.

grey, grey, grey.

piebald fawn.

6

none.

16

albino.

grey, grey, grey,
piebald grey,

piebald grey,

piebald black.

piebald fawn.

8

none.

17

none.

grey, grey :W, pie-
bald grey ;W.

lilac.

2

2

18

albino :W,
albino :W.

piebald grey, pie-
bald grey.

fawn, lilac, pie-
bald lilac.

5

2

19

albino,

albino.

grey, grey, grey.

fawn.

6

none.

20

albino.

grey, black ;W.

fawn, fawn, pie-
bald fawn4

6

1

W printed after the name of the colour of a mouse in the above Table indicates
that the mouse was a waltzer. All the other mice were normal in their mode
of prograssion*

92

BREEDING

in the mice represented on Plate II. runs on the
same lines as that of, for instance, tallness and dwarf-
ness in peas, and colour in the Andalusian fowls.
All these three instances agree in the production
of “ dominants,” “ hybrids,” and “ recessives ” in
the second hybrid generation in the proportion
1 : 2 : 1 ' in every four, and differ in the character of
the hybrid which, in the case of stature in the peas
was indistinguishable from one parent, in the case
of colour in the fowls was intermediate between that
of the two parents, and in that of colour in the mice,
constituted a reversion to the ancestral type of
coloration. The mice, further, afforded an instance
of a cross in which two pairs of independently inherited
characters were involved. A closer consideration of
such cases will form the subject of the next chapter,
and the line of argument to be followed now will
lead us back again to the phenomenon of reversion,
and show how, at any rate, some cases of it may be
explained in the light of Mendelian phenomena.

CHAPTER VII

THE RELATION BETWEEN CHARACTERS BELONGING TO
DISTINCT PAIRS

The reader is already familiar with the relation
between two characters of the same pair, such as
yellowness and greenness of the cotyledons ; that is
to say, with the results which follow when an indi-
vidual bearing one character of a pair (yellow) is
crossed with one bearing the other member of the
same pair (green). The question to be dealt with now
is the relation between members of distinct pairs
of characters, such as yellowness of cotyledons, be-
longing to one pair of characters, and roundness
or wrinkledness belonging to another pair. A proper
understanding of this relation is of great practical
value, because it enables the breeder to effect the
combination of desirable characteristics existing in
distinct strains with great swiftness and precision.

The nature of this relation will at once become
apparent when the result of a cross between two
individuals which differ from one another in respect
of characters which belong to two pairs of characters
has been described. For this purpose I shall describe
the result of a cross between a pea with wrinkled
yellow cotyledons, and a pea with round green ones.
Let the exact nature of this cross be clearly under-

93

94

BREEDING

stood before the result of it is described. Plate I.
represents the result of a cross between two races
of peas which differ in the colour of their cotyledons.
The two races crossed only differed in respect of a
single characteristic — colour. Both the parent forms
happened to be round. Fig. 21 represents the result
of a cross between two races of peas which differ
only in the form of their cotyledons. Both the
parent forms happened to be yellow — but this is
not shown in the figure. In the cross about to be
described the parent forms differ in respect of both
the colour and the form of their cotyledons ; that is to
say, the results shown separately in Plate I. and Fig. 21
are shown together. There is nothing new in the
case we are about to deal with : it is merely two
cases which have already been considered separately,
in separate crosses, considered together in one cross.

/ Indeed, every detail of the result of the cross between
the yellow wrinkled and the green round can be pre-
dicted from the knowledge which we now possess of
the result of crossing a yellow with a green, and a
round with a wrinkled. Yellow is dominant to green
and round dominant to wrinkled ; the result, there-
fore, of crossing a yellow wrinkled with a green round
is a yellow round. This is shown in Plate III. ; the
yellow wrinkled race is represented by a group of
six peas to the (top) left of the picture, the green
round one by a group of eight peas to the (top) right.
The first hybrid generation is represented by five
yellow round peas between and a little below the
parent forms.

Plate III.— THE RESULT OF CROSSING A YELLOW WRINKLED
WITH A GREEN ROUND PEA

(Top left) Yellow Wrinkled Parent. (Top right) Green Round Parent.

(Five Peas in middle line) First Hybrid Generation.

(In the pods) Second Hybrid Generation.

CHARACTERS OF DISTINCT PAIRS 95

A cross between a yellow round and a green
wrinkled would have involved the same two pairs
of characters. The hybrid would have been, as in
the case I have shown, yellow round ; that is to say,
exactly like one of its parents. The second hybrid
generation would have been the same in the two cases.
The difference between the two cases, of course, lies
in the distribution of the characters over the parent
forms ; in other words, in the case I have figured,
each parent has a dominant member of one pair
and a recessive member of the other pair. But
when a yellow round is crossed with a green wrinkled
the former has both the dominant members and the
latter both the recessive ones. That is why in this
case the result of the cross is identical with one of the
parents. I have, however, described a cross between
two individuals, each of which bears a dominant
and each a recessive character, for a definite reason,
to illustrate the fact that the dominance or recessive-
ness of a character is not a result of constitutional
vigour or weakness of the animal or plant that bears
it, but is a property peculiar to that particular
character. A cross between a round and a wrinkled
does not show this ; nor does a cross between a
yellow and a green ; nor even does a cross between a
yellow round and a green wrinkled. In all these
cases the hybrid is like one of its parents, and we
have no means of knowing that its character is not
determined by the superior vigour of that parent.
But directly we have seen the result of a cross between
two forms, each of which possesses a dominant

BREEDING

96

character, we perceive that dominance attaches not
to the individual but to the character; Some
individuals in certain animals, such as horses, are
said to possess the power of impressing their charac-
teristics on their offspring whenever they are mated,
which, in the case of stallions, is pretty often. What-
ever be the nature of this power, which is called
prepotency, it is clear that it has nothing to do with
dominance. Prepotency is an attribute of individuals,
and capricious in its appearance. Dominance is an
invariable attribute of particular characteristics. In
the case before us there is no question of a certain
individual impressing its characters on its offspring.
The hybrid takes one of its characters, yellow, wholly
from one parent, and another, round, from the other.

We are thus brought in view of one of the con-
clusions to which experimental breeding has led,
namely, the conception that living things are made
up of a number of characters which arise separately
and are transmitted separately. Whatever may be
the value of this suggestion as a help to the under-
standing of evolution, there is no doubt as to its
value as a guide in the practice of breeding.

The second hybrid generation produced by the
self-pollination of the yellow round hybrids referred
to above must now be considered. Briefly, this
generation consists of individuals presenting all the
four possible combinations of yellow, round, green
and wrinkled ; namely (9) yellow round, (3) yellow
wrinkled, (3) green round, and (1) green wrinkled, in
the proportions indicated by the figures in brackets

CHARACTERS OF DISTINCT PAIRS 97

prefixed to them. This proportion is a very impor-
tant one ; it is, therefore, desirable to make this
point perfectly clear. In the second hybrid genera-
tion resulting from a cross between a yellow wrinkled
and a green round pea there will be :

9 Yellow round, 3 Yellow wrinkled, 3 Green round,
and 1 Green wrinkled

in every sixteen peas, on the average.

It will at once be seen that the peas bearing two
dominant characters are in the majority : there are
nine out of sixteen. Next come peas with a dominant
and a recessive character, which may be either yellow
and wrinkled or green and round. Of each of these
combinations, there are three in the sixteen. Last
of all, in point of number, come those peas both of
the characters of which are recessive, namely the
green wrinkled ones ; of these there is only one in
every sixteen.

This 9 : 3 : 3 : 1 proportion follows from the
co-existence of two 3 : 1 ratios in the second
hybrid generation produced by a single cross, in the
following way. The two pairs of characters, yellow
and green, and round and wrinkled, are distributed in
their proportions of three to one, at random, over the
individuals composing the second hybrid generation.
What is meant by “at random ” is that the shape
of the cotyledons is not affected, one way or the
other, by their colour, or their colour by their shape.
It would be an even chance that a round pea were a
yellow or green, or that a yellow one were a round

H

BREEDING

98

or wrinkled, if yellows and greens, and rounds and
wrinkleds occurred in equal numbers. But there
are three yellows to one green in every four, and three
rounds to one wrinkled in every four. And the
proportions in which each of the four combinations
occur is arrived at by multiplying together the ratios
in which each of the characters in a combination
occur separately : —

Yellow and round .
Yellow and wrinkled
Green and round
Green and wrinkled .

. . 3x3=9

. . 3x1=3

. 1x3=3

. 1x1=1

A glance at Plate III. will show the kind of
approximation to the 9:3:3: 1 ratio which is obtained
in a small number ; there are thirty-eight seeds
altogether, and the various combinations of characters
occur in the following proportions : —

18 Yellow round, 7 Yellow wrinkled, 12 Green round, and
1 Green wrinkled

That is to say, there are too many green rounds and
too few green wrinkleds ; but deviations of this magni-
tude in so small a sample are of no significance.

So far the visible characters of the plants — i.e. of
the cotyledons of the young plants seen through the
seed-coats — have alone been considered. The dis-
tinction between pure and hybrid individuals bearing
the dominant character has been left out of account.
It must now be considered. The simplest case is
that of the green wrinkled ; every one of these,
appearing in the second hybrid generation, will

CHARACTERS OF DISTINCT PAIRS 99

breed true, inasmuch as both of its characters are
recessive.

The green rounds breed true to their recessive
character (greenness), but may or may not breed
true to their dominant one, roundness. There are,
therefore, two kinds of green rounds — a kind that
breeds true to both of its characters and produces
green rounds only, and a kind that breeds true to
colour but not to shape, and produces green rounds
and wrinkleds. Amongst every three green rounds
there will be one of the former and two of the latter
kind, on the average.

The yellow wrinkled are in like case. They breed
true to their recessive character (wrinkledness), but
may or may not breed true to their dominant one
(yellowness). There are, therefore, two kinds of
yellow wrinkled, a kind that breeds true to both of
its characters and produces yellow wrinkleds only,
and a kind that breeds true to shape but not to
colour, and produces yellow and green wrinkleds.
Amongst every three yellow wrinkleds there will be
one of the former kind and two of the latter, on the
average.

There are four kinds of yellow rounds : (1) A kind
that breeds true to both of its characters and produces
yellow rounds only ; of which kind there is one in
every nine on the average. (2) A kind that breeds true
to colour and not to shape and produces yellow rounds
and wrinkleds ; of which kind there are two in every
nine. (3) A kind that breeds true to shape but not to
colour, and produces yellow and green rounds ; of

100

BREEDING

which kind there are also two in every nine. And,
lastly, (4) a kind that does not breed true to, but is
hybrid in respect of, both its characters, and produces
all four kinds — yellow round, yellow wrinkled, green
round, and green wrinkled. Of this kind there are
four in every nine on the average. On the opposite
page are given the actual characters . of the seeds
of ten plants which were all raised from the yellow
round seeds of a hybrid generation, such as that
which we have considered ; it will be seen that there
are two instances of the first kind (Plants Nos. 5 and
8), one of the second (Plant No. 7), two of the third
(Plants Nos. 1 and 4), and five of the fourth kind
(Plants Nos. 2, 3, 6, 9 and 10). Of course it does not
always happen that, in a sowing of ten seeds, repre-
sentatives of each of the four kinds occur. Nor,
when they do occur, amongst ten plants is there
likely to be a close approximation to the ratio given
above. The composition of the second hybrid genera-
tion we are considering may, therefore, be summarised
as follows : —

1. Yellow round —

(i) Pure to colour and shape . . 1

(ii) Pure to colour but not to shape . 2

(iii) Pure to shape but not to colour . 2

(iv) Hybrid in both respects ... 4

9

2. Yellow wrinkled —

(i) Pure to colour and shape . . 1

(ii) Pure to shape but not to colour . 2

8

Table to show the distribution of yellow round, yellow wrinkled,
green round, and green wrinkled seeds, containing cotyledons of the
third hybrid generation, on ten plants (1 to 10) of the second hybrid
generation. From actual specimens.

The numbers in each horizontal row, under each plant, relate to the
seeds in individual pods. Thus, there were sixteen pods on the first
plant, eleven in the second, and so on. The first pod on the first plant
contained eight seeds, the second six, the third five, and so on.

YR

YW

GR

GW

YR

YW

GR |GW

1

YR

YW

GR

CW

4

4

2

4

1

3

"l

1

5

1

6

1

3

1

2

5

3

3

3

2

1

4

1

5

1

1

1

1

3

6

2

4

2

1

1

1

3

2

1

6

1

2

3

1

2

1

1

4

5

5

1

1

1

1

3

2

3

1

4

2

1

2

2

1

1

4

2

1

3

1

1

1

1

3

1

2

3

1

1

1

2

2

3

1

1

3

2

1

3

1

1

1

2

1

2

49

21

37

111

10

1

39

17

16

4

6

7

8

15.

YW

GR|GW

E

YWjGRICW

YR

YW

GR

CW

1

1

5

2

6

4

1

1

5

1

5

5

1

1

3

2

7

2

1

2

6

1

8

2

3

2

3

1

6

4

2

2

2

1

7

3

1

1

3

6

3

1

4

6

4

1

4

5

4

1

1

4

3

5

3

1

2

2

6

1

2

4

4

1

5

3

2

1

2

2

4

1

1

2

5

1

3

1

4

5

2

1

1

4

2

1

3

2

1

6

3

1

1

2

3

3

2

43

18

16

4

78

25

86

YR

YW

GR

CW

6

1

1

5

1

4

2

4

2

3

3

6

1

5

1

4

2

3

2

4

1

2

1

3

2

2

1

2

2

1

4

1

3

3

2

1

1

3

2

2

2

1

1

73!

30

YR YW OR CW
6

5
7

6
6
6
6
6

5

6
4
3
2
3
3

74

10

YR

YW

GR

YR

YW

CRiGW

3

2

1

3

1

3

1

3

1

1

1

2

1

1

3

2

3

1

1

1

3

1

4

1

1

3

1

1

2

3

2

2

1

1

1

4

1

3

3

1

2

1

1

1

1

1

1

26

3

8

2

23

7

7

6

102

BREEDING

3. Green round —

(i) Pure to colour and shape , , 1

(ii) Pure to colour but not to shape , 2

3

4. Green wrinkled —

Pure to colour and shape . , 1

It will be observed that there is only one pure
breeding representative of each of the four categories
yellow round, yellow wrinkled, green round, and
green wrinkled.

The ratios in the above may be arrived at by
multiplying together the ratios in which the parti-
cular characters which make up each combination
occur separately, having regard to the question
whether the character is in a pure or hybrid state.
Thus, when a yellow is crossed with a green the
second hybrid generation is composed as follows :
1 pure yellow, 2 hybrid yellow, 1 green, in every
four ; and the second hybrid generation from a cross
between a round and a wrinkled has the following
composition : 1 pure round, 2 hybrid round, 1

wrinkled, in every four, on the average. The com-
bination of these six classes will, therefore, occur in
the following ratios : —

[1. Yellow round] —

Pure yellow and pure round , 1x1 = 1

Pure yellow and hybrid round . 1x2 = 2

Hybrid yellow and pure round 2x1 = 2

Hybrid yellow and hybrid round 2x2 = 4

—9

CHARACTERS OF DISTINCT PAIRS 103

[2. Yellow wrinkled] —

Pure yellow and wrinkled, . 1x1 = 1

Hybrid yellow and wrinkled . 2x1 = 2

[3. Green round] —

Green and pure round
Green and hybrid round

—3

1X1 = 1

1X2 = 2

[4, Green wrinkled] —

Green and wrinkled

—3

1x1 = 1

— 1

The epithet “ pure ” is omitted before the
recessive character in the above table to lay stress
on the fact that the recessive character is always
pure.

To consider for a moment the practical application
of the facts which have now been set forth, it will
be seen that there are four combinations of characters
in this second hybrid generation, two of which are
different from the parent forms started with, namely,
the green wrinkled and the yellow round. Of these
four combinations only one type — that which possesses
two recessive characters, the green wrinkled — can be
counted on to breed true, straight away. We have
to wait no longer in this instance than the second
hybrid generation for a constant new variety. In
the case of the other three combinations, it is necessary
to wait till the next generation before we can be
certain of getting a constant type. For directly we
find a plant (bearing a fairly large number of seeds)

104

BREEDING

which produces only yellow rounds, or only yellow
wrinkleds, or only green rounds, we can be sure that
it is pure in respect of both colour and shape, and will
continue so. And it may be laid down as a general
rule that the permanent combination of two recessive
characters may be effected in the second hybrid
generation, but that the permanent combination of
a recessive and a dominant character (belonging, of
course, to distinct pairs), or of two dominant
characters, cannot be effected till the third hybrid
generation. It is curious that the great majority of
tne best culinary peas are green wrinkled. But this
fact is probably not merely due to the facility with
which these two characters can be combined ; for
though there is no advantage, from the culinary
point of view, in the green over the yellow, the
advantage of the wrinkled over the round will be
seen to be very great.

An instance of the value a familiarity with this
9 : 3 : 3 : 1 ratio, as a help to understanding otherwise
unintelligible results of cross-breeding, will now be
given. One of the characteristics which distinguish
the various breeds of poultry is the form of the comb.
The commonest type of comb, and that possessed by
the wild ancestor of our domestic poultry, is known
as the single comb, which is represented in Fig. 22.
The comb is flat from side to side, and its free edge
is marked by deep indentations. Another type of
comb, which characterises the Wyandottes, is shown
in Fig. 23, and is known as the “ rose ” comb. It is

Fig. 22.— single comb

(Black Leghorn.)

Fig. 23.— rose comb

(Partridge Wyandotte.)

Photographs by the kindness of the Rev. T. W. Sturges.

CHARACTERS OF DISTINCT PAIRS 105

squat, and its surface is marked, all over, by many
convolutions. A third type (Fig. 24), known as the
“ pea ” comb, occurs in the Sumatra game, for instance,
and may be said to consist of three low ridges, a
median and two lateral ones. A fourth type of
comb, known as the walnut comb, and only occurring
in the Malays, is shown in Fig. 25. It consists of a
globular excrescence not unlike a walnut.

If a fowl with a “ pea ” comb is mated with one
bearing a “ rose ” comb the resulting hybrids have
‘‘ walnut ” combs. When these hybrids are mated
together, the generation produced has the following
remarkable composition : 9 “ walnut,” 3 “ rose,”
3 “ pea,” : 1 “ single.” No “ single,” it will be remem-
bered, was put into the cross. What is the origin of
the single comb in this second hybrid generation,
and what are the two pairs of characters responsible
for this 9 : 3 : 3 : 1 ratio are questions which natur-
ally present themselves.

The result of crossing the yellow wrinkled with
the green round pea gives the clue. It will be remem-
bered that in that instance both the dominant
characters occurred together in the individuals which
appeared in the proportion of nine in the sixteen ;
and that the dominant characters occurred separately
in the two lots of individuals which appeared in the
ratio of three in the sixteen, i.e. the yellow in the
yellow wrinkled, and the round in the green round.
In this case, therefore, we should suppose that the
two dominant characters are “ rose ” and “ pea.”
What the recessive characters corresponding to them

\

106 BREEDING

are will be indicated shortly. For the present, let us
observe the result of their union. It will be at once
noted that we cannot, as in the case of the peas,
observe the co-existence of the two characters
“ rose ” and “ pea ” in the hybrid. But this should
give no cause for surprise. In the case of the peas,
one pair of characters relates to colour and the other
to shape ; so, not only is it natural to expect to
see them separately in the hybrid, it is impossible
to imagine how they could become merged. But in
the case of the fowls’ combs, both the characters
relate to shape, and it is therefore hard to conceive
how, when they both exist in the same comb, they
can escape blending, or at any rate producing a
shape which is different from either, and in which
neither can be recognised separately.

The “ walnut ” character is, therefore, due to the
co-existence in the same comb of the “ pea ” and
“ rose ” types. The “ single ” comb, which makes
up one-sixteenth of the second hybrid generation
we are considering, is the result of the absence of
the two characters “ pea ” and “ rose.” The case is
entirely analogous to that of the peas : the yellow
rounds have two dominant characters ; so has the
“ walnut,” namely, “ pea ” and “ rose ” ; the yellow
wrinkled has one dominant character and one reces-
sive ; so has the “ pea,” namely presence of the
“ pea ” character, and absence of the “ rose ” — if
the “ rose ” were present the comb would be
“ walnut.” The green round also has one dominant
character and one recessive ; so has the “ rose,”

Fig. 24.— pea comb

(Sumatra Game.)

Fig. 25.— walnut COMB

(Malay.)

Photographs by the kindness of the Rev. T. IF. Sturges.

CHARACTERS OF DISTINCT PAIRS 107

namely presence of “ rose ” and absence of “ pea ” ;
the green wrinkled has two recessive characters, so has
the “ single ” — absence of both “ pea ” and “ rose.”
And the proportions in which these analogous things
occur in the second hybrid generation are the same,
thus : —

9 “ Walnut ’’ 3 “ Pea.” 3 “ Rose.” 1 “ Single.”

[=“Pea”and [=absence of both

“Rose”] “Pea” and “Rose”]

9 Yellow round, 3 yellow wrinkled, 3 green round, 1 green wrinkled.

It will be remembered that it was said earlier in
this chapter that the same results would have followed
if, instead of crossing a yellow wrinkled with a green
round pea, a yellow round were crossed with a green
wrinkled. The two crosses only differ in the fact
that in the former each parent has a dominant
character, whilst in the latter both dominant
characters exist in one parent. The cross we have
described in the case of the fowl’s comb, namely
“ pea ” by “ rose,” is analogous to the one displayed
in Plate III., namely, yellow wrinkled by green
round. In both cases each parent has a dominant
character. The cross in the case of the combs,
analogous to that between yellow round and green
wrinkled, is a cross between “ walnut ” (possessing, or
consisting of, two dominant characters, “ pea ” and
“ rose ”) and “ single.” Here, as in the case of the
peas, precisely the same results follow in the second
hybrid generation, namely 9 “ walnut,” 3 “ rose,”
3 “ pea,” and 1 “ single.” The appearance, in the

io8

BREEDING

second hybrid generation, of two entirely new
characters — the “ rose ” and “ pea ” combs — which
were not present in either of the parents or in the
first hybrid would have been unintelligible to the
experimenter who was not familiar with the 9 : 3 : 3 : 1
ratio, and with the proof, which this ratio affords,
that we are dealing with a cross between two
forms which differ in respect of two pairs of
characters.

The reader will have noticed that a new con-
ception of the nature of the Mendelian pair of
characters has been invoked to explain the pheno-
mena of inheritance presented by the characters
of the fowl’s comb. Mendel’s own results led
to the conclusion that characters which were
inherited in Mendelian fashion were associated
together in pairs, such that one member of a pair
was dominant and the other recessive. No indica-
tion was given of any features which were peculiar
to dominant characters — that is to say, no clue
was given by which a dominant character could
be known or suspected to be dominant before
the result of mating it with a recessive one was
known. The two members of a pair possessed this
feature in common, that they were not merely drawn
at haphazard from the characters of the plant, but
both pertained to the same part of the organisa-
tion of the plant. Thus, yellow and wrinkled do
not constitute a pair, but yellow and green. The two
members of a pair only differ from one another by

CHARACTERS OF DISTINCT PAIRS 109

the fact that when two individuals, each of which
possesses one of them, are mated, a hybrid is produced
in which one of them, which is called the dominant,
exists to the more or less complete exclusion of the
other.

This is the general conclusion, as to the relation
between the two characters of a pair, which was
derived from a consideration of Mendel’s own experi-
ments with peas. Now, it will be seen that from the
breeder’s point of view it is very desirable to possess
some sign by which a dominant character may be
known before the crosses are made ; also one is
naturally curious to know what it is that makes one
character of a pair dominant and the other recessive.
A conception of the nature of dominant, in contra-
distinction to recessive, characters has already been
hinted at in the explanation which was given of the
result of crossing a fowl with a “ pea ” comb with
one with a “ rose ” comb. The two pairs of characters
involved in that cross were supposed to be the
character of the comb known as ‘‘ pea ” (dominant)
and the absence of that character (recessive) ; and
the character of the comb known as ‘‘ rose ”
(dominant) and the absence of that character (reces-
sive). In both cases the presence of a particular
character constituted the dominant member of the
pair, and the absence of that character the recessive
member. The dominant character is due to the
presence of something — the recessive to the absence
of that something. It is not easy at present to see
how this conception may be applied to the pairs of

no

BREEDING

characters we have already considered, such as yellow
and green, round and wrinkled ; and before I attempt
to show how it can be applied to them I will describe
a cross involving two pairs of characters to which
its application is obvious.

This cross is between two kinds of peas {Pisum)
which differ in the colour and arrangement of the pig-
ment in their seed-coats. In one of them there is a
rich brown mottling of anastomosing brown lines on a
paler background ; this is characteristic of the
so-called “ maple,” or “ partridge,” peas which can
be had of any corn chandler. The brown mottling
is only visible on close inspection ; in the mass the
peas appear brown. A specimen of a pea with a
“ maple ” seed-coat is shown, magnified, in Fig. 26.
The seed-coat of the other variety, with which the
“ maple ” is crossed, is marked with a great number
of minute purple spots on a background, which is
pale greenish grey in the newly ripe pea, but becomes
dark brown with age. Certain garden peas, such
as the French sugar pea, exhibits this type of
coloration. A pea with this grejdsh seed-coat always
has purple flowers, whether there are purple spots
on the seed-coat or not. Further, the grey may
exist in the seed-coat without purple spots, but the
purple spots are not known to occur on seed- coats
which are not grey. This fact does not concern us
now, but is of great importance in the theory of
reversion, which will be set forth later. For the sake
of brevity, the type of seed-coat which has purple
spots on a grey background will be referred to as

" Maple." •• Purple-spotted maple.'

es g

1

•5^

^Iz

u- <« 1-^

U a H

S.'" o

«*-o U

■®“ S (A

e « W

W

o « rrj
■« S' CD

Q
»
H'
H

. o

J Oh
o CD

a •

« W

Qu

Q

Z

<

: M

d J

*0.

5 ^

CHARACTERS OF DISTINCT PAIRS iii

“ purple spot ” simply. A specimen is shown in
Fig. 26.

When peas bearing these two varieties of seed-coat
are crossed, there results a hybrid, the seed- coat of
which exhibits both the “ maple ” and the “ purple
spot ” character, which is seen below and between
the two parent forms in Fig. 26. This result is
rather surprising, because it might have been supposed
that “ maple ” and ‘‘ purple spot ” were the dominant
and recessive members of a single pair of characters
from the fact that they were stated to characterise
one and the same part of the plant, namely, the seed-
coat. It will be remembered that one feature of
characters which constituted a Mendelian pair of
characters was said to be the fact that they pertained
to the same part of the organisation of the animal
or plant which bore them ; and it might be supposed
that the converse was true, namely, that two
characters which pertained to a particular part of
the plant, if so facto, constituted a pair of Mendelian
characters. But the fact that both “ maple ” and
“ purple spot ” are simultaneously present in the
hybrid shows that these two characters do not con-
stitute a pair, but are members of distinct pairs.
The generalisation, however, in regard to the common
location of the two members of a pair of characters,
and the converse of this generalisation are saved by
the fact that there are two layers in the seed-coat
of the pea, an outer and an inner, and that the
pigment to which the “ mapling ” is due is
lodged in one of them and the pigment which

II2 BREEDING

gives rise to tlie purple spots is situate in the
other.

The second hybrid generation produced by the
self-fertilisation of the hybrids just described shows
at once what the pairs of characters involved in this
cross are. It consists of nine plants, the seed-coats
of which bear both “ maple ” and “ purple spot,”
three with “ maple ” seed-coats only, three with
“ purple spot ” only, and one with neither “ maple ”
nor “ purple spot,” but a pale homogeneous grey
coat, amongst every sixteen plants, on the average.
The cases of a 9 : 3 : 3 : 1 ratio in the second hybrid
generation, which have already been described, at
once show what the two pairs of characters are. In
these cases, it will be remembered, the individuals
which occur in the ratio of nine possess two dominant
characters ; the two lots, which occur in the ratio of
three, possess respectively the dominant member of
one pair and the recessive of the other; whilst the
individuals, which occur in the ratio of one, possess
the recessive members of both pairs. The two pairs
of characters in the instance under consideration are,
therefore, “ maple ” (dominant) and the absence of
“ maple ” (recessive) ; and “ purple spot ” (dominant)
and the absence of “ purple spot ” (recessive). Cases
such as this have given rise to the theory that one
(the dominant) of the two characters, which make
a pair, consists in the presence of something, and
that the other (the recessive) consists in the absence
of that something. The theory has been called the
Presence and Absence hypothesis. The reader who

CHARACTERS OF DISTINCT PAIRS 113

is reminded of logical exercises, and suspects that
this theory, like other products of the imagination,
may bear but a very distant relation to actuality,
may discover that his suspicion is happily — to a large
extent, at any rate — without foundation by making
the cross which I have described, if he has a few
square yards of ground at his disposal ; when he
will obtain about once in every sixteen plants, in
the second hybrid generation, a plant bearing peas
in the seed-coats of which the two absences of
“ maple ” and “ purple spot ” are combined — peas
which can be seen and handled.

We have dealt in this chapter with pairs of
characters which are inherited in complete independ-
ence of one another, i.e. with cases in which
characters belonging to one pair are unaffected by
characters belonging to another, and have come in
view of a simple theory of the nature of dominant
and recessive characters. The further consideration
of this theory will be dropped for the present, to be
resumed in the chapter after next. In the next
chapter we shall deal with cases in which characters,
belonging to one pair, are very profoundly affected
by characters belonging to another pair.

z

CHAPTER VIII

' THE INTERPRETATION OF REVERSION

For instances of reversion which may be explained
in the light of Mendelian facts of inheritance we can
again draw on the characters of Pisum. One of the
characters with which Mendel experimented was the
colour of the seed-coat, which was either grey
(dominant) or white (recessive). The grey seed-coat
might, or might not, be marked with the purple
spots which contributed to the subject matter of
the last chapter. We are now concerned with the
grey seed-coat which is stated to be destitute of the
purple spot. As a matter of fact, purple spots are
present, but they are very faint and scarcely dis-
cernible. So that the difference between the two
kinds of greys is not that one has, and the other has
not, purple spots, but that one has pronounced purple
spots (which have already been seen in one of the
parents of the cross described in the last chapter.
Fig. 26) and the other scarcely discernible ones (which
cannot be seen in the pea to the top left of Fig. 27).
The latter will be referred to simply as grey to avoid
circumlocution ; the former as grey with purple
spot. We are now concerned with the spotless
grey (Fig. 27, top left), and with the pure white,
which is shown at the top right of Fig. 27. When

“Grey with purple spots.” \Vhite.

INTERPRETATION OF REVERSION 115

peas characterised by these two types of seed-coat are
crossed, the resultant hybrid may either have a seed-
coat like the grey parent or it exhibits the full
purple spotting. This latter result is shown in Fig. 27,
the hybrid grey seed-coat with purple spots being
seen below and between the two parent forms. The
hybrid here possesses a character, the purple spot-
ting, which is absent from both parent forms, but
exists in the wild pea found in Palestine and elsewhere,
which probably corresponds more closely than any
other to the wild ancestral form from which our
cultivated peas have descended. The reappearance,
therefore, of this purple spotting as the result of a
cross between two forms, neither of which possesses
it, constitutes a typical instance of reversion or
throwing back.

The second hybrid generation, produced by the
self-fertilisation of these reversionary hybrids, consists
of peas with purple -spotted grey seed-coats, peas
with grey seed-coats, and peas with white seed-coats,
in the ratio of 9 purple-spotted grey, 3 grey,
and 4 white. This is an entirely new proportion,
the meaning of which is not at first sight obvious.
The theory which has been put forward to explain it
is as follows : It is supposed that two pairs of
characters are involved in this cross : they are

grey ” and “ absence of grey ” ; and “ purple spot ”
and “ absence of purple spot ” — exactly as in the
instance described at the end of the last chapter,
“ maple ” and its absence, and “ purple spot ” and
its absence. But in the instance now under con-

ii6

BREEDING

sideration the results are supposed to be complicated
by the fact that one of the characters of one pair is
affected by a character belonging to the other pair.
It is supposed that the character “ purple spot ”
can only exist in a pea which has a grey seed-
coat. In other words, the character “ purple spot ”
cannot be manifested unless the character “ grey ”
is also present. But the character “grey” can be
manifested in the absence of the “ purple spot.” If
there was not this connection between the characters
of two distinct pairs we should expect the composition
of the second hybrid generation to be analogous to
that which we have seen to follow from other instances
in which two pairs of characters are involved. We
should expect it to consist of 9 grey with purple
spot (two dominant characters), 3 grey, but with-
out purple spot (one dominant and one recessive
character), 3 purple-spotted, but without greyness
of seed-coat (one dominant and one recessive
character), and one white, i.e. without purple spot or
grey ness (two recessive characters) : an ordinary

9 : 3 : 3 : 1 proportion. This is what we should
expect if we did not suppose that purple spot and
greyness of seed-coat were connected in the manner
stated above, viz. the impossibility of the mani-
festation of purple spot in the absence of grey ; and the
possibility of the manifestation of grey in the absence
of purple spot. Now, let us see how this supposition
affects the 9 : 3 : 3 : 1 ratio. The second 3 represents
three plants the seed-coats of which would exhibit
purple spots on a white seed-coat on the supposition

INTERPRETATION OF REVERSION 117

that purple spot was not dependent on the simul-
taneous presence of grey for its manifestation.
But on the supposition that the purple spotting
cannot exist in the absence of grey, the seed-coats of
these three plants will be simply white, and, there-
fore, indistinguishable from the single white plant
occurring at the end of the series, in the ratio of one
in the whole sixteen. The first 3 in the series will
not be affected by this supposition, because grey can
exist in the absence of purple spot ; nor will the 9,
because both characters are present. The second
hybrid generation will, therefore, consist, according
to this theory, of nine plants the seed-coats of which
are grey with purple spot, three grey and four white.
And this, it will be remembered, is the result which
was actually obtained. Of course, the fact that the
theory harmonises with the facts does not prove that
the theory is true. Its truth can only be tested by
finding out whether all the consequences which follow
from it actually occur.

But assuming that it is true — and I, for one, believe
that it approximates fairly closely to the truth — it
brings us within reach of an explanation of one of
the phenomena with which the breeder is most
familiar, namely, that of reversion, or throwing back.
Reversion on crossing, in such cases as those which
we have been considering, is due, according to this
theory, to the reunion in one individual of two
characters, the simultaneous presence of both of
which is necessary for the existence of the ancestral
character. In the instance discussed above, the

ii8

BREEDING

purple spotting of the wild ancestor of our culinary
peas depends on the co-existence in one individual
of two characters, the purple spotting itself and a
grey seed-coat. At some period, probably after the
pea had been domesticated, these two characters
became separated; how, we cannot do more than
guess. And the reappearance of the ancestral
character on crossing is due to the union of two
individuals one of which has one, and the other has
the other, of the two characters necessary for the
production of the ancestral character.

Another instance of reversion is afforded by the
colour of the flower of the culinary pea. The three
colours known are shown on Plate IV. On the
right is shown the commonest type, the pure
white flower ; on the left is what may be called the
pink flower ; the flower is not however, as a matter
of fact, equally pink over its whole extent, the outer
single petal, the “ standard,” is nearly white, and
the paired inner ones, the “ wings,” are salmon pink,
the keel, enclosed by the “ wings,” is white with
green veins.

Between these two flowers, and above them, is the
so-called purple flower, which approximates most
closely in its colour to that of the wild ancestor of
garden peas. The standard is a pale purple in which
the blue is in excess of the red ; the wings are a dense
purple in which the red preponderates over the blue.

When a pink-flowered variety is crossed with a
white-flowered one, the result is a purple-flowered
plant. The pink and the white flowers on Plate IV.

Plate IV— MENDELIAN INHERITANCE OF THE COLOUR OF THE FLOWER IN THE CULINARY PEA

Two flowers of a plant Two flowers of a plant produced by Two flowers of a plant

of a pink-flowered race. crossing the pink with the white. of a white-flowered race.

INTERPRETATION OF REVERSION 119

are from plants belonging to pure pink-flowered and
pure white-flowered races respectively ; and tbe
purple flower is from a plant of the first hybrid
generation produced by crossing these two races.

These purple-flowered hybrids produce, when
self-fertilised, a second hybrid generation which
consists of purple-flowered, pink-flowered, and white-
flowered plants in the ratio of 9:3:4 respectively.
The explanation of this case is precisely similar to
that suggested to account for the result of crossing
a grey- with a white-seed-coated pea. But it was not
so obvious in the present case, for a reason which
will be given shortly. The two pairs of characters
in this case are supposed to be “ blue ” and “ absence
of blue,” and “ pink ” and “ absence of pink ” ; but
the blue, like the purple spot, cannot be manifested
in the absence of the pink ; so that the composition
of the second hybrid generation which, if the blue
could be manifested in the absence of pink, would

is, as a matter of fact —

9 Pink and Blue . . [= Purple] i.e. 9 Purple

3 Pink and not Blue . [= Pink] i.e. 3 Pink

3 Blue and not Pink . [= White] \ . wv.*f

1 not Blue and not Pink [= White] ^ ^

The type of second hybrid generation which has
been met with twice in this chapter deserves further

be —

9 Pink and Blue
3 Pink and not Blue
3 Blue and not Pink

1 not Blue and not Pink

. [= Purple]

. [= Pink]

. [= Blue]

. [= White]

120

BREEDING

consideration. It may be expressed in general terms
in the following scheme : —

[Parents] Pale form x Colourless form

[First Hybrid Generation] Dark form

[Second' Hybrid Generation] 3 Pale form

9 Dark form
4 Colourless form

The three categories are written in this order,
3:9:4, instead of 9:3:4 for a reason which will
shortly appear. The manner in which the two cases
which we have considered fit into this scheme is as
follows : —

Seed-coat Colour

[Parents] Grey x White

[First Hybrid Gen.] Grey with purple spots

[Second Hybrid Gen.] 3 Grey

9 Grey with purple spots
4 White

Flower Colour

[Parents] Pink x White

[First Hybrid Gen.] Purple

[Second Hybrid Gen.] 3 Pink

9 Purple
4 White

It will also be apparent, on reflection, that the
phenomena of the inheritance of colour observed in
the mice fits into the same scheme : —

INTERPRETATION OF REVERSION 121

[Parents] Fawn and white x White

with pink eyes | with pink eyes

[First Hybrid Gen.] Grey with black eyes

[Second Hybrid Gen.] 1 Fawn or lilac, with pink eyes

2 Grey or black, with black eyes
1 White with pink eyes

The attempt to fit this last case into the scheme
brings out another point — the similarity of the pro-
portion 1:2:1 to that of 3:9:4. This similarity
should, however, be merely accidental, because the
3:9:4 (or 9:3:4) is supposed to follow from
the 9 : 3 : 3 : 1 ratio, and its connection with the
1:2:1 ratio is thus very remote and devious.
For my own part, I am inclined to think that
the resemblance is not accidental. I find it
difficult, for instance, to regard the fact that in
both the 3:9:4 and the 1:2:1 the colourless forms
occur in the proportion of 25 per cent, of the genera-
tion as a mere fluke, and entirely without significance.
But I do not think that it is profitable to discuss
this matter until more cases of second hybrid genera-
tions, in which these ratios are said to occur, have
been bred on a scale large enough to satisfy statisticians
that they really do occur.

For instance, in the case of the mice, it is possible
that the ratio in that case is 3 : 9 : 4 and not 1:2:1.
And the question which of the two it is has still to
be decided.

Moreover, in this chapter I have been merely
concerned to point out that certain cases of reversion

122 BREEDING

can be easily described in terms of Mendelian here-
ditary processes.

The reader should be familiar by now with the
manner of inheritance of Mendelian pairs of characters
in the result of crossings in which only one such
pair is involved, and in crossings in which two such
pairs are involved ; and with those cases in which
the members of one pair of characters exert no effect
on members of another pair, i.e. in which cases there
is a 9 : 3 : 3 : 1 ratio in the second hybrid generation ;
and also with those cases in which a member of one
pair of characters depends for its manifestation on
the presence of a member of another pair, in
which cases there is a 9:3:4 ratio in the second
hybrid generation.

Our review of the main phenomena of Mendelian
inheritance is now concluded. Most of the illustrative
cases are such as can be bred by the reader if he has
the curiosity to do so. A full description of the
conduct of a breeding experiment, together with
the names of the varieties which exhibit the characters
referred to, will be found in Chapter X.

We now pass on to the subject alluded to at the
end of the last chapter, namely, the nature of the
Mendelian pair of characters. This question is of
the greatest interest to the practical breeder and to
the investigator. But we are as yet very far from a
satisfactory answer to it. The work of the practical
breeder would be greatly simplified by the possession
of some simple rule wliich would tell him before he

INTERPRETATION OF REVERSION 123

made his crosses which characters would behave as
dominants and which as recessives. To the biologist,
some conception of the nature of the Mendelian pair
of characters is a necessary preliminary to any
attempt to estimate the true significance of the
Mendelian phenomena.

1

CHAPTER IX ’O

THE NATURE OF THE MENDELIAN PAIR OF CHARACTERS

A QUESTION which naturally and frequently presents
itself to the mind is : “ What is it that makes one
character dominant over another when they meet
in a cross ? ” A common answer to this question
is that the dominant character is strong and the
recessive one is weak ; but this is too vague a notion
to be of the nature of an explanation ; moreover,
it is, in fact, little more than a restatement of the
difficulty involved in the question.

Before proceeding to the two theories which
appear to be worthy of consideration, I think it is
desirable to warn the reader against an idea, which
has found its way into print, that there is some
connection between selection and dominance ; that
a dominant character is the result of a long process
of selective breeding. There is no ground whatso-
ever for such an idea, and it is difficult to imagine
how it can have arisen.

Two theories of the nature of the Mendelian pair
of characters, and of the cause of dominance in parti-
cular, have obtained wide credence. One of them
has already been exploded, the other is still supported
by a considerable party of adherents.

The first of these is the idea that the oldest of

124

PAIRS OF CHARACTERS 125

the two members of a pair of characters is the domi-
nant. When, for instance, one of the members of
a pair of characters is that borne by the species
in which the character occurs in its wild state, and
the other only occurs in a state of domestication, it
would be supposed, according to this theory, that
the “ wild ” character, being the older, would be the
dominant, and the “ domesticated ” one, being the
more recent, would be the recessive. In mice, for
instance, the grey of the house mouse would be
expected to be — and is — dominant over the white
of the albino. Now, there are a great number of
cases which support this theory. For instance, the
dominant members of the pairs of characters pre-
sented by the cotyledons in Pisum — ^yellow and
round — both occur in the wild specimens of the
wild Pisum which have been examined. There are,
however, a number of cases in which this theory
breaks down entirely. For instance, the single comb
in the domestic fowl, which is known to be the form
of the comb in the wild ancestral stock of the jungle,
is recessive to the “ rose ” and to the pea ” comb.
The absence of a tail in rumpless breeds of fowl is
dominant to the ordinary tailed condition which
characterises the majority of domestic breeds, and,
of course, the wild representatives of the species.
So this theory falls to the ground.

We see, therefore, that though many characters
of wild species are dominant to their domesticated
counterparts, the fact that certain of them are not
dominant suggests that, in those cases where they

126

BREEDING

are, dominance is not simply due to the fact that
the character in question belongs to the wild species ;
and that some other cause must be sought for.

This is suggested by the second theory. The
reader will remember that, in order to explain the
result, of crossing a pea with a maple seed-coat with a
pea with a purple -spotted one, we invented the theory
that, in this case, the pair of characters did not
really consist of two characters, like yellow and green,
but of a particular character, maple, as the dominant
member, and the mere absence of this character as
the recessive member of the pair. And the impres-
sion may have been given that this theory was only
invoked to explain an outstanding case. So far
from this being the case, this theory, which is called
the Presence and Absence hypothesis, is now being
applied to all the other Mendelian characters. In
many cases the fitness of this application is at once
obvious. In the case of eye-colour in man, for instance,
the duplex character which is dominant, is due to the
presence of the brown pigment, and the recessive
character, simplex, is due to the absence of this pig-
ment. Stated in its most general terms, this theory is
that the dominant character is due to the presence of
something, and the corresponding recessive character
is due to the absence of that something. It will be
perceived that this conception of the constitution of
the Mendelian pair of characters differs profoundly
from that which we should naturally derive from a
contemplation of, for instance, the seven pairs of
characters studied by Mendel. In all these seven

PAIRS OF CHARACTERS 127

cases the recessive character appears to be as much
due to the presence of something as the dominant
one — greenness as much as yellowness, wrinkledness
as much as roundness, and so on ; and it appears
that in these seven cases we really are dealing with
seven genuine pairs of two equally real characters,
and that in no case is one of the characters merely
the absence of its counterpart. But we shall see
that not only can the presence and absence hypothesis
be made to apply to these cases, but that it probably
throws much light on the real nature of these
characters.

The application of this theory to the characters
of the cotyledons does not consist merely in the
statement that wrinkledness is simply the absence
of roundness. That would be a mere juggling with
words. The application of this theory consists in
the statement that wrinkledness is due to the absence
of that which determines roundness. And this raises
the question : What is it that determines roundness ?
To answer this question it is necessary to consider
the part played by the cotyledons in the economy
of the plant. They are, as the reader is probably
aware, storehouses of food material on which the
young plant depends almost exclusively from the
moment at which germination begins until the young
plant has attained a height of about six inches.

This food material is, of course, laid down in the
cotyledons whilst the seed is still in the pod. It
first appears in the form of a sugary solution, and
as the seed matures, this sugar is gradually converted

128

BREEDING

into starch. This starch is in the form of grains
which, in the case of the round pea, may be best
described as potato -shaped. They are shown in
Fig. 28. These grains can easily be seen, in the
following way : A dry round pea is soaked in water
for, twenty- four hours; a thin slice is then cut off
with a razor, the cut surface of the pea is then
wetted with a drop of water, and scraped. The
whitish fluid scraped off will be found to consist
almost entirely of these starch grains, when examined
through the microscope.

If the starch grains of a wrinkled pea (Fig. 29)
are examined in the same way they will be found to
present a very different appearance from those of the
round pea, for whilst the latter are entire, the former
are compound ; their substance is subdivided by
radiating strands of a refringent yellow matter. More-
over, whilst the grains of a round pea are distinctly
elongate, those of a wrinkled one are roughly
circular, if we neglect the indentations caused by
the radiating strands.

It is evident, then, that the starch is in a different
state in the round pea from that in which it occurs
in the wrinkled. The probable nature of this differ-
ence is indicated by the following experiment : If
a round pea is weighed, and then soaked in water
for twenty-four hours, and then weighed again, the
weight of water absorbed can be determined. If
this weight is then expressed as a fraction of the
weight of the dry pea, a certain figure, which I have
called the “ absorptive capacity,” is obtained. The

Fig. 28.— starch GRAINS OF ROUND PEA

Magni&ed about 300 times.

Fig. 29.— starch GRAINS OF WRINKLED PEA

Magnified about 300 times.

PAIRS OF CHARACTERS 129

amount of water absorbed by a round pea iu twenty-
four hours is about 80 per cent, of its weight when
dry. In other words its absorptive capacity is 80. It
does not absorb its own weight of water in twenty-
four hours. A wrinkled pea, on the other hand,
absorbs more than its own weight of water in twenty-
four hours. Its absorptive capacity is about 120.

The cause of this difference is probably that, in
the case of the wrinkled pea, the whole of the sugar
is not converted into starch. In fact, the wrinkling
itself, which takes place during the ripening of the
seed, is probably due to this incomplete conversion
of the sugar into starch. All peas, wrinkled as well
as round, are round before they are dry ; that is,
in the state in which they are eaten. Cooks think
that all peas are round. It is only gardeners, who
sow or harvest the dry seed, who know the distinction
between round and wrinkled. Wrinkled peas wrinkle
because more water escapes from them than from
round peas; and more water escapes from them
because water can escape from a sugary solution
more easily than it can from the grains of starch.
The water is, so to speak, locked up in the solid
starch grain much more securely than it is in a solu-
tion of sugar.

It is not possible — at any rate it would be very diffi-
cult— to measure the relative amounts of water given
off by round and by wrinkled peas during the process
of ripening. But what can be done is to measure
the converse of this process, namely, the relative
amounts of water taken up by round and wrinkled
J

130

BREEDING

peas when they germinate; for, inasmuch as round
and wrinkled peas are equally round before they lose
their water, and are equally round when they have
taken it up again, we have some ground for assuming
that each kind takes up as much when it germinates
as it loses when it ripens. The fact, therefore, that
a wrinkled pea takes up more water than a round
one does, when it germinates, may be taken to mean
that it loses more than a round one does when it
ripens.

The theory that less of the sugar in the wrinkled
pea is converted into starch is supported by the
well-known fact that wrinkled peas are sweeter than
round ones, and also by the fact that all the favourite
eating peas on the market are wrinkled ones.

We are now in a position to give tentative answers
to the questions : What is it that determines roimd-
ness ? and. What is it that determines wrinkledness ?
and then to indicate the way in which the presence
and absence hypothesis applies to this pair of
characters.

In a round pea, it would seem that all the sugar
is converted into starch ; in a wrinkled one, only part
of it is ; and the wrinkling is primarily due to the
escape of the water from the solution of sugar left
over after ripening ; and, consequently, in the last
resort, due to the absence of that which completes
the conversion of the sugar into starch or, at any
rate, to an insufficiency in the quantity of that
substance, whatever it is. The round pea has the
full dose — to use Mr. Bateson’s phrase — of this

PAIRS OF CHARACTERS 131

substance ; the wrinkled pea an insufficient dose.
Something is absent from the wrinkled pea which is
present to the full in the round.

It will be gathered from the above that the
object of our inquiries is not merely to analyse an
organism into its component characters and to leave
the matter there, but to push deeper and obtain
some insight into the fundamental nature of these
characters. The reader will do well to lay it to heart
that in the case of the pair of characters, roundness
and wrinkledness, which Mendel placed first on his
list of the seven with which he experimented, we are
at present only on the threshold of an investigation
of the true nature of the characters themselves.

The application of the presence and absence
h}q)othesis to the other pair of cotyledon characters
will now be considered. The two characters which
meet the eye are yellow and green. But the matter
is not so simple as this. Mr. Bunyard has shown
that there is a yellow and a green pigment both in
the yellow and in the green cotyledon. When both
are present at the same time, as in the ripe but still
moist pea, the green masks the yellow. All peas,
both yellow and green varieties, are green when
they are eaten. Just as cooks think that all peas
are round, so they think that all peas are green.
It is only gardeners who sow and harvest them
who know the distinction between yellow and green.

The ripe but still moist cotyledons of both yellow-
and green-seeded varieties are, therefore, green. The
yellow kind become yellow as they ripen ; the green

132

BREEDING

do not change colour during this process. The
yellowing of the former is brought about by the
gradual fading and disappearance of the green pig-
ment, which thus leaves the yellow pigment (which is
present in both kinds) exposed. The successive
stages in the fading of the green can be easily observed.
The simultaneous presence of both green and yellow
pigment in yellow and in green peas has also been
demonstrated.

To sum up, green-seeded varieties contain two
pigments in their cotyledons, a yellow and a green ;
neither of them fade during the process of ripening,
and, inasmuch as the green masks the yellow, the
ripe seed is green. Yellow-seeded varieties also
contain the same two pigments, but the green fades
during the process of ripening ; so that the ripe
seed is yellow. This fading of the green pigment
in the yellow pea is supposed to be brought about
by the presence of some substance which is absent
from the green pea.

It is not possible to indicate the manner in which
this presence and absence hypothesis applies to the
remaining of the seven characters dealt with by
Mendel, except, of course, in those cases in which
its application is obvious — for instance, to the colour
of the seed-coats, which may be grey or white. It
would be easy to say that normality of stem was
due to the presence, and fasciation to the absence,
of some factor controlling the normal succession of
nodes in the stem ; but this is doing no more than
re-stating the problem in other terms. And until

PAIRS OF CHARACTERS 133

an exhaustive inquiry into the causes determining
the nature of these two characters — normality and
fasciation — has been carried out it is absurd to attempt
to interpret them by means of any theory whatsoever.
With regard to tallness and dwarfness, a start has
been made by Mr. Keeble, who has shown that tallness
depends on two characters, length and (rather
curiously) thickness of internode ; dwarfness depends
on the absence of both of these qualities. Peas
which possess only one of them fall into the category
known as half-dwarf.

Before we leave Mendel’s seven characters it will
be well to pay some attention to a question which
can be most conveniently discussed here, though it
may seem that none of the questions so far dealt
with in this chapter are worth discussing until it
has been answered — and in the affirmative, too. The
question is. Is there such a thing as dominance ?
Does one character hold the stage in the first hybrid
generation to the complete exclusion of its partner ?
If an affirmative answer cannot be given to these
questions, what reason is there in discussing the
causes determining dominance, a thing whose reality
has not yet been established ? But in the present
case it is desirable to present a consistent scheme
of the whole set of phenomena, and then to consider
to what extent it is justified. Let us take the first
pair of characters, cotyledon shape. If we content
ourselves, as the impressionist does, with the mere
impression produced on the unaided eye, the answer
to the question in regard to the existence of dominance

134

BREEDING

is an unqualified affirmative. A round is crossed
with a wrinkled, and the hybrid is a round, indis-
tinguishable from the pure parent round. But if
we refuse to be content with this impressionist
aspect, and set about to find out as much as possible
about, the true nature of wrinkledness and roundness,
both in the pure and in the hybrid rounds, we shall
find that the matter is not by any means so simple.

The starch grains of pure round and wrinkled
peas have already been dealt with. Those of the
round are elongate and entire, those of the wrinkled
are circular and compound, the number of com-
ponent pieces varying between 2 and 8, the commonest
number being 6.

The starch grains of the cotyledons of the first
hybrid generation are shown in Fig. 30. They
constitute about as complete a blend between the
two parental types as can be imagined. In the first
place the grains are intermediate in shape ; this has
been demonstrated by measurements. They are also
intermediate in compoundness, in two ways ; in what
may be called the distribution of compoundness, some
of the hybrid grains being compound and some of them
single ; and, secondly, in what may be called the
degree of compoundness of those grains which are
compound ; the number of component pieces in the
hybrid is intermediate between the number in the
starch grains of the wrinkled pea and entirety, as we
may name the feature presented by the grains of
round peas. The number of component pieces in
those hybrid grains which are compoimd varies

Fig. 30.— starch grains OF A GROSS
BETWEEN A ROUND AND A WRINKLED
PEA

Magni&ed about 300 times.

PAIRS OF CHARACTERS 135

between 2 and 4, and the commonest number, based
on a large number of counts, is 3.

It is only to the superficial observer that domi-
nance is complete in this instance, the first character
dealt with by Mendel. The microscope at once
reveals the fact that the pure and hybrid rounds,
though externally indistinguishable, are in fact of
radically different constitutions. This fact is also
revealed by a determination of the absorptive
capacities of the pure parental and hybrid rounds.
It will be remembered that the absorptive capacity
of the wrinkled pea was stated to be about 120, and
that of the pure round 80. The absorptive capacity
of the hybrid round is almost exactly intermediate
between the two.

The hybrid round, therefore, is seen to resemble
the pure only in the fact of being round, and to differ
in what must be regarded as more fundamental
properties ; and it is probable that with further
inquiry the number of cases in which simple domi-
nance occurs, i.e. where the hybrid is identical with
the dominants, will be greatly reduced.

The following considerations are suggested by
the facts set forth in this chapter. In the experi-
mental study of heredity, it is not sufficient to find
that round and wrinkled, or any other characters,
constitute a Mendelian pair, and are inherited in
Mendelian fashion. Some attempt must be made
to get behind the appearance of roundness, for
instance, and to find out what roundness really
consists in. If no such attempt is made the fact

BREEDING

136

that the result of a cross between a round and a
wrinkled is ap'parently a round identical with the
round parent, but is in reality a blend between the
round and wrinkled, is entirely missed. The vulgar
reluctance to be clear in the mind as to the things
which are represented by words used has not been
without its effect on the complexion of the Mendelian
doctrine as expounded by the less considerable of
its adherents. When a character has been found
to behave in Mendelian fashion in inheritance, it
is considered that the nature of the character and
its mode of inheritance have been determined ; and
that any further investigation of the matter is
superfluous. In this way a great many instances of
characters which apparently behave in a Mendelian
way in inheritance have been accumulated. But
as to the essential nature of the characters
themselves we are still ignorant. It is probable
that, when the seven pairs of characters investigated
by Mendel have been thoroughly worked out, the
terms now used to designate these characters will
have no more than the most remote application to
them, and that, for the purpose of accurately repre-
senting the essential nature of these characters,
they will be wholly valueless.

It is here that the value of a theory like the
presence and absence hypothesis comes in. This
hypothesis leads to the investigation of the nature
of the characters themselves. This is its chief
value. If it does not do this, if it is used merely
to restate already known facts in new terms, it is

PAIRS OF CHARACTERS 137

useless. But if it stimulates an investigation of
the nature of characters described as Mendelian, it
is justified.

How far the truth of this hypothesis is borne
out by the available evidence is an entirely different
question, which will now be discussed. There is a
large number of pairs of characters to which this
theory obviously applies. There are also a number of
cases, like cotyledon shape and colour, which this
theory correctly describes, though the manner of
application is not at first sight easy to see. But
there is also a considerable number of cases which
constitute obstacles in the way of its general appli-
cation. For instance, there are certain cases where
apparently the same character is dominant in one
animal, or plant, and recessive in another. For
instance, the possession of horns is dominant in
sheep, but recessive in cattle, to the hornless con-
dition ; white in pigs appears to be dominant to
black, but recessive to it in sheep ; and, lastly, the
dissection of the margin of the leaf is dominant
to the entire condition of the edge in the nettle,
and recessive to it in Chelidonium, It is doubtful
in this last case whether the condition of entirety
is of the same nature in the two cases ; but the
possession of horns in cattle and sheep certainly
would seem to be.

Then there is a further set of cases, in which the
dominant character seems to consist in the absence
of something and the recessive in the fresence of
that something. For instance, the possession of

BREEDING

138

five heavily pigmented, brown bands in the snail
of the hedgerows (Helix hortensis) is recessive to the
pale yellow, entirely bandless condition. Again, it
is not, at present, easy to see how the fact, discovered
by Mr. Bifien, that susceptibility to the attacks of
“ rust ” in wheat is dominant to immunity from such
attacks can be brought in line with this hypothesis.
In the case of the snails, it must be supposed that the
possession of bands is due to the absence of some-
thing which is present in the pale yellow snail
and prevents them appearing. In the case of the
wheat it must be supposed that resistance is due to
the presence of some anti-toxin, and susceptibility
to the presence of some substance which prevents
the formation of the anti-toxin. With further
inquiry the scope of this hypothesis can be better
estimated ; at present its chief service is that it
stimulates this inquiry.

Another point raised by this Presence and
Absence hypothesis can be brought out by comparing
it with Prof, de Vries’s conception of the Mendelian
pair. This theory, as stated in the first chapter,
is that specific characters blend on crossing ; but
that varietal ones exhibit Mendelian phenomena.
The reason for this is that the specific characters are
new creations, isolated and without partners, whilst
varietal characters are the result of the transference
of an already existing character from one state to
another, as, for instance, from the patent to the
latent, or vice versa. Thus when one organism is
crossed with another from which it differs in the

PAIRS OF CHARACTERS 139

possession of a specific character, this specific
character meets nothing in the other organism
corresponding to it, and the result is a permanent
blend. Such crosses are called unisexual. But if
one organism is crossed with another from which
it differs in the possession of a varietal character, the
particular character in a patent state in one organism
meets the same character in a latent state, in the
other organism, and the result may or may not be
a blend, but is invariably segregation. These crosses
are called bisexual, or balanced crosses, the latter of
which is a much better term. Prof, de Vries’s theory
of the nature of the Mendelian pair differs from the
presence and absence hypothesis in that, according
to him, the pair consists of a character in an active
or patent state (dominant) and the same character
in a latent state (recessive) ; whilst according to
the other theory the dominant member of a pair of
characters consists in the presence of something
and the recessive in the complete absence, not the
latency, of that something. It is premature to hold
the scales between these two before a much greater
body of evidence bearing on the question is available.
The investigation designed to furnish this evidence
must start from a perception of an essential difference
between the two theories ; the most important of
these, in my opinion, is the difference, according to
the two theories, between the constitution of the
recessives. According to the one theory, the
recessive is the dominant character in a latent
state; according to the other, it is the absence of

140

BREEDING

the dominant character. Now, in the latter theory
the recessive character is nothing ; there is no
difference, according to it, between the absence of
fasciation and the absence of greyness of seed-coat —
a point which deserves careful consideration. And
one way of attacking the question of the relative
truth of the two theories is to find out if the recessive
members of, say, Mendel’s seven characters are
identical, or if they do not each manifest something
corresponding to their dominant counterpart.

CHAPTER X

THE CARRYING OUT OF A BREEDING EXPERIMENT WITH

A PLANT

I HAVE two objects in view in dealing with this
matter : first, to render the reader familiar with

the means by which the facts which have already
been related have been determined, and in particular
to give directions for the repetition of the actual
crosses made by Mendel ; and, secondly, to lead up
to an exposition of Mendel’s theory, through a
familiarity with the germ cells themselves, acquired
by actually handling them.

I think it may be said that in the case of no other
branch of science can results of such interest be
obtained at so small a cost. The experiments by
means of which the high-water mark of modern
chemistry has been reached, such as — to take an
instance at random — the determination of the pro-
rperties of radium, are so expensive that there are
few who can command the means to perform them,
and they are practically outside the reach of the great
mass of teachers of chemistry altogether. But in
the case of that branch of biology which, rightly or
wrongly, has attracted a share of interest which is
comparable with that attracted by radium in
chemistry, the case is very different. The obser-

142

BREEDING

vations made by Mendel can be repeated by any-
one who has, or can obtain the use of, a few square
yards of ground ; the expenses in addition to this
are trifling, and amount to no more than the few
shillings which will buy the seeds and the pea -sticks.
And anyone who has a kitchen garden, and room to
spare in it, can contribute new results of great value
with very little extra outlay.

The following directions are based on six years’
experience in my own garden.

It is desirable that the ground in which peas are
to be grown should be deeply dug in the autumn
before the spring in which the seeds are to be sown.
My own ground was dug two spits deep, and an
ample quantity of farmyard dung incorporated with
the bottom and with the top spit. If the land is at
all clayey, it should be well hmed ; but if it is coarse,
open and sandy, chalk should be used instead of lime.
During the digging a sharp eye should be kept open
for wireworms (the larvae of the skipjack beetle);
for if many of these are left in the ground they will
play havoc amongst the seedlings in the spring. It
is well to remind those who are not familiar with
wireworms that they are very tough, and that it
is therefore necessary to pull them in two, to make
certain that they are destroyed.

The first week in March is the earliest date at
which it is advisable to sow the seed ; for even when
seed is sown at this time the seedlings do not begin
to break through the ground until the beginning of
April. The chief enemies of the swollen seeds, before

EXPERIMENT WITH A PLANT 143

the seedlings have shown above ground, are mice ;
and these vermin must be destroyed once and for
all, with a stern hand. When the seedlings have
begun to unfold their leaves they are by no means
free from enemies, either above or below ground.
But the subterranean enemies are now no longer
mice, but wireworms. The signs of the attack of
one of these on a plant are unmistakable. The
leaves of a seedling which has been making vigorous
growth will one day be seen to flag and droop ; if
the soil round the underground part of the stem is
examined, a wireworm will in nine cases out of ten
be found there. The enemies above ground are
birds, and these are best guarded agai/ast by nets
hung loosely over the rows on low sticks. Having now
dealt with the enemies of the peas during their earlier
stages, we may revert to the actual sowing of the
seed.

The seed should be sown in drills, which should be
drawn with an ordinary hoe. Adjacent drills in
the case of tall peas should never be closer than
six feet asunder. What may be gained in an increase
in the number of plants by planting the rows closer
will be lost in a decrease in the quality and quantity
of the seed produced by the plants, and in a lowering
of the health of the plants themselves. I can most
heartily endorse the opinions on this matter expressed
in that most excellent of all books on gardening,
William Cobbett’s “ English Gardener,” published
in 1838.

The seed should be sown between two and three

144

BREEDING

inches deep in the ground. The distance of the seeds
apart depends on whether it is desired to produce
as large a number of plants as possible in a given
space, or as large a number of seeds as possible per
plant. In the former case, it is not advisable to
place the seeds nearer than three inches apart in the
drill, but two rows may be sown in one drill, three
inches apart from each other, in such a way that a
seed of one row is opposite an interval in the other
row. If seeds are sown thicker than this, I find that
there is a danger of weaker plants being fatally over-
shadowed by their better-favoured fellows, or, at any
rate, of their being prevented from flowering or
setting seed. In the case in which it is desired to
obtain as many seeds as possible per plant, the seeds
may be sown at least nine inches apart in a single
row in the drill. Whether, if the seeds were sown
one foot apart, the plants would be able to avail
themselves of the additional soil-space and light,
and bear more seeds than plants grown nine inches
apart, I do not know. But that plants grown nine
inches apart set a vastly greater quantity of seed
than plants grown three inches apart I know as the
result of experiment.

It is desirable to cover the patch of ground, on
which the peas are growing, with netting before the
time when the pods have swollen to the size at which
they are usually gathered for the table. The unripe
peas at this stage constitute an irresistible temptation
to birds, especially to jays and hawfinches, which can
do a great deal of damage in a very short time. And

EXPERIMENT WITH A PLANT 145

I have sometimes lost nearly half my harvest through
not netting the peas in time. The netting is effected
by hanging ordinary strawberry nets over a frame-
work consisting of stout ashen poles standing eight
feet out of the ground, and long, slender horizontal
ones of the same material, nailed to the tops of
them. Slender poles are better than laths (which I
have tried) for this purpose, as the nets can be slipped
over them easily, whilst they catch on the laths.
The small birds which can get through the mesh
of a strawberry net do little, if any, harm.

If the poles are a good eight feet out of the ground
and the net is not permitted to sag between the
cross-pieces, the presence of a net does not interfere
with the giving of such attention as the plants require
during the summer. This attention is of two kinds ;
one which is always necessary, hoeing and weeding ;
another which may or may not be required, cross-
fertilisation.

It is desirable that the ground immediately
round the plants should be kept free of weeds by the
hand, say, once a fortnight, and that the six feet
between adjacent rows be kept free of weeds, and its
surface broken by the hoe, especially two or three
days after heavy rain so as to prevent the surface
from caking and the water in the soil from evaporating.
The Dutch hoe is best for this purpose, because with
this tool it is not necessary to walk over the part
which has been hoed, during the process itself.

The cross-fertilising can easily be done by any
one with nimble fingers.

146

BREEDING

The pea carries its reproductive organs in its
flowers ; and the brightness of the flowers is supposed
to serve as an attraction to insects whose services are
required to transfer the fertilising element from one
flower to another. Whether the pea- flower ever
required these services, I cannot say. But it is certain
that it does not do so now ; for the pea is normally self-
fertilised. It may be necessary to remind the reader
that the reproductive organs of both sexes are present
in one and the same flower in the pea, as in the
majority of flowering plants. The male cells are
contained in the nine yellow bodies, called anthers,
borne on long, delicate stalks, inside the flower ; the
anther together with its stalk is called the stamen. The
female cells are contained in the rod-like structure
called the ovary, which has a curved and tufted tip
called the fistil. If a newly but fully opened flower of a
pea be plucked and examined, it will be found to con-
sist of the following parts : Five green “ petals ”
encircling the flower and appearing to be continuous
with the flower stalk, called sepals : a single, erect
petal, known as the standard ; a pair of petals, facing
one another, more or less, the wings ; and inside the
wings a single, boat-shaped, and much stifler and
greener structure with a projecting keel, which gives
this part its name, the carina. Inside the carina
are the anthers, and, surrounded by the anthers, the
pistil. If the flower is fully open, the anthers
will be ripe ; that is, they will have burst and
discharged their pollen grains. Each pollen grain
contains a male reproductive cell. The pollen

EXPERIMENT WITH A PLANT 147

appears to the naked eye as a fine yellow dust,
but the individual grains composing it can easily
be seen by the aid of a small hand lens.

If the ovary be slit by a vertical incision, it
will appear to be — what, indeed, it actually is — the
future pod in miniature. It is almost exactly like
the ripe pod full of peas ; everything is there, but
in miniature ; the whole pod is not longer than
one-third of an inch. Inside each of the miniature
peas is a minute cell, the female reproductive cell,
or egg, or ovule.

Before such an egg can develop it must unite
with or, in other words, be fertilised by a male
reproductive cell. In the case of a great many
plants there are special arrangements by which the
eggs are always fertilised by pollen from another
flower, the pollen being transferred by the agency of
the wind, or by that of insects. This is what is
known as cross-fertilisation. But in the pea, and
in many other plants as well, the ovules in a parti-
cular flower are fertilised by the pollen from the same
flower. This is known as self-fertilisation. But the
essential process in the two cases is the same : it is
the union of a male with a female reproductive cell.
In all cases of biparental reproduction, which is all
we are concerned with in this book, both in plants
and in animals, fertilisation consists essentially in
this union of a male with a female reproductive cell,
and until this union has taken place the develop-
ment of the egg into the future plant or animal
cannot proceed.

BREEDING

148

The pea, as has been already stated, is self-
fertilised. Fertilisation occurs before the flower
has opened. The fact that a flower is open is a
certain sign that fertilisation has taken place.
Fertilisation occurs in the following manner : At a
stage in the development of the flower, shown in
Fig. 31, i.e. long before the flower has opened, in
fact whilst its petals, the only one
of which to be seen at this period
is the folded standard, are still of
a greenish colour, the anthers burst
and the pollen grains escape into the
space enclosed between the two
halves of the carina. In the centre
of this space lies the pistil, so that
it at once becomes surrounded with
Fig. 31.— Stage in ^n immense number of pollen grains,
^he development male reproductive cells in the
when fertilisa- pollen grains reach the ovules in the
following way. Some pollen grains
stick to the free end of the pistil, and each one
of them sends out a structure, known as the pollen
tube, which grows like a root into the tissue of the
pistil and travels down it until it reaches one of
the ovules. This pollen tube contains the male
reproductive cell, and when this has fused with an
ovule, the ovule proceeds to develop.

When, therefore, it is desired to make a cross
between two kinds of peas, it is necessary to select a
flower, on the plant which is to function as the female
parent, of such an age that the anthers have not

EXPERIMENT WITH A PLANT 149

yet burst in it (i.e. a little younger than the one
shown in Fig. 31), that is to say, of such an age that
the ovules in it have not yet been fertilised. The
reader is recommended to discover the external ap-
pearance of the latest stage in the growth of the flower
at which the anthers usually burst, by picking, say,
ten stages in the development of the flower, ranging
from a bud in which the petals do not extend beyond
the sepals, to the nearly open flower. The stage,
in the development- of the flower, at which the
pollen is set free will be found to be much earlier
than was expected. A flower, at a stage of develop-
ment preceding this, must be selected to function as
what we may, for convenience sake, call the female
parent, or female flower.

Having described the appearance of the flower
suitable to function as female parent, we can proceed
to a description of the method by means of which one
variety may be crossed with another. A very simple
set of implements is required for this — (1) a pair
of fine forceps ; (2) a pair of fine scissors ; (3) a
glass tube, or wide-mouthed bottle with cork
stopper, to contain methylated spirit. A hand
lens is also a desirable addition but is not a
necessity.

Having selected the flower of the right age to
function as female plant, the first thing is to make
certain that it has not already fertilised itself, by
seeing that the anthers have not yet burst (with a
hand lens if available) ; the next thing to do is to make
it impossible for the flower to be self-fertilised, and

BREEDING

150

this is done by cutting out the anthers. The reader
will doubtless evolve the details of his own method
of procedure in carrying out cross-pollination. For my-
self, I have found that a simple way of getting at the
parts to be manipulated is to hold back the standard
and wings, and to make a transverse tear right across
one half the carina near its base, and then bend the
whole of the carina in front of the tear to one side,
so that the anthers and stigma project free from the
part of the carina, in front of the tear. The stamens
are then easily removed by nipping them ofi with
the forceps. The forceps should then be dipped
in the spirit to sterilise them, in case the anthers
may have been accidentally burst during these mani-
pulations.

The flower is now ready to receive the pollen of
the flower which is to function as male parent. This
should have been selected before the female flower
was operated on in the manner just described, so that
it may be at hand as soon as the female flower is
ready.

The age of the flower to function as male parent
should be older, but not much older than that which
acted as female. The reason for this is not that an
older flower would not contain enough pollen, but
that it might contain too much ; I mean pollen
which does not belong to it, brought to it by bees.
For though the bees who visit the pea flowers, once
they are fully open, do not effect cross-pollination,
because self-pollination has already taken place, they
may bring to the flower some pollen from another

EXPERIMENT WITH A PLANT 151

flower. It is, therefore, not wise to use as pollen-
bearer a flower which has opened.

Having selected a bud of a suitable age, the plan
I adopt is to pluck it from the plant on which it is
growing, to pull off the sepals, the standard and the
wings, and to snip a piece about the size of a large
pin’s head from the extreme tip of the carina with the
scissors. This flower may, for convenience, be called
the “ male flower.” After this operation the scissors
should be dipped in the methylated spirit, so as to
kill any pollen grains that may be sticking to them,
and so avoid mixing them with the pollen of the next
flower similarly operated on.

The female flower should now be held between
the thumb and finger of the left hand, and its pistil
thrust through the entrance in the tip of the carina
of the “male flower” made by snipping ofl its tip,
into the mass of pollen which is just inside, if the
anthers have burst. There is no difficulty in telling
whether they have or not. If they have, the pistil
when withdrawn will be seen to be yellow with pollen ;
if they have not, it will bear no trace of pollen ; which
means that too young a bud was taken; in which
case it must be thrown away, and an older one
procured.

When the pistil of the flower which is to function
as female parent has been thoroughly covered with
pollen, the part of the carina in front of the tear
made across it, which was slipped aside to set the
pistil and anthers free, may be put back in its natural
position so that it encloses the pistil ; the wings

152

BREEDING

which have been folded back during the manipula-
tion are allowed to revert to their natural
position, one on either side of the carina, and
the standard also allowed to return to the posi-
tion natural to it in a bud of this age, shown in
Fig. 31.

A bud treated in this way grows, opens, and
blossoms, then withers and dies, in exactly the same
way as a bud which has not been interfered with.
But the course of its subsequent history, as traced in
the characters of its progeny, has been profoundly
altered by what has been done.

We are brought very close up against the mystery
of heredity when the extreme narrowness of the
channel through which the characters of organisms
have to find their way between each generation is
forced upon our attention, as it is by actually
making a cross between, say, a tall and a dwarf.
From the flower of a dwarf pea will grow out a pod
containing seeds all of which will be dwarf. But if
such a flower is deprived of its stamens and its pistil
covered with the pollen from a tall pea, all of its
offspring will be tall peas. We are thus firmly held,
without any possibility of escape, and forced to
recognise the fact that the character of the tall pea
is contained in each one of the pollen grains, them-
selves scarcely visible, as individual grains, to the
naked eye. This is how the problem of heredity
presents itself to the mind to-day : How are the
characters of the organism represented in the germ
cells which produce it ? The solution of this problem

Magaified about 50 times.

Magnified about 160 times.

Fig. 32.— pollen GRAINS OF THE
CULINARY PEA

EXPERIMENT WITH A PLANT 153

put forward by Mendel will be dealt with in the next
chapter.

When the petals of the pollinated flower have been
folded into their place a label should be tied round
the stalk of the flower by means of an 8-shaped loop,
one half of which encircles the flower stalk, and the
other the main stem. A number, corresponding to
a number in a record of all the necessary details with
regard to the cross, should then be written on the
label. A label bearing the same number, preceded
by “ f. of ” (meaning “ father of ”), should be tied
in the same manner round the stalk of the flower
which furnished the pollen with which the cross
was made. And nothing further need be done to
it until the autumn when the pod is hard and dry
(see Fig. 33), when the seeds can be taken out of
it and put into a packet with the number on the
label tied round the flower in the summer, written
on it. In order to render the possibility of error as
remote as possible, it is well to cut off the label and
put it into the packet with the seeds.

It is convenient to describe here a system of
keeping the records of my experiments which I have
found to work very well. I have two classes of
numbers from 1 onwards in each year — “ mating num-
bers ” and ‘‘ sowing numbers.” Mating numbers,”
which are written on the labels tied round the crossed
flowers, and on those tied round the plant which
furnished the pollen, are distinguished by the year
in which the cross is made ; for instance, the first
cross I shall make this year, 1911, will be 11.1, and

154

BREEDING

the hundredth 11.100. But it is not necessary to
write the prefix indicating the year on the labels
themselves, because the only crosses harvested in a
given year are crosses made in that year ; this prefix
must, however, of course be wiitten on the packet in
which the seeds are kept, because the packets may
be kept for many years. ‘‘ Sowing numbers,” which
are written on the pegs marking where peas are
sown, are distinguished by capital letters, A standing
for the first year of the experiment, B for the second,
and so on. But here, as in the case of the “ mating
number,” it is not necessary to write the prefix indicat-
ing the year on the pegs, because all the pegs in the
ground in any given year relate to seeds sown in that
year ; this prefix must, however, as in the case of
the mating prefix, of course, be written on the packets
in which the harvested seeds are kept.

I strongly recommend anyone undertaking breed-
ing experiments not to use note books, but loose
sheets of squared paper. The year, in full, should
be written or, still better, stamped on one of the top
corners of every sheet.

With regard, now, to the harvesting of the peas.
A plant should be pulled up, as a general rule, when
all its pods, save the uppermost one or two, are hard
and dry. The stem should be pulled off as far below
the level of the ground as the unaided fingers can
get ; it is necessary to do this because stems some-
times branch very near the ground, and the two
divisions of one plant might easily be harvested as
two plants if the precaution which I have advised

Fig. 33.— to illustrate METHOD OF LABELLING MALE
AND FEMALE PARENTS OF CROSS

EXPERIMENT WITH A PLANT 155

is not taken. It is not advisable to dig up the whole
root of the plant, partly because it is a superfluous
and inconvenient cumbrance in the subsequent deal-
ings with the plant, and partly because the roots do
good to the soil if left in it, by contributing to its
humus and to its supply of nitrogen-fixing bacteria.

All the plants related to by one peg should be
pulled up at once and tied together near the roots
with a coarse open twine which will not cut or break
the stems. The piece of twine should first be tied
round the peg (notched on either side so that the
twine does not slip) relating to the plants tied up.
If these bundles consist of about ten plants, ten such
bundles should be tied together to form a “ secondary
bundle.” The terms “ peg ” to signify a bundle of
peas tied up with a single peg, and ‘‘ bundle ” to
signify a secondary bundle as just defined, have come
into everyday use between myself and the assistant
who helps me with harvesting. The bundles should
be hung up so that the tops of the plants hang down-
wards in any place where they are likely to dry, and
are secure from the depredations of birds, rats or
mice. The likelihood of their drying properly
depends partly on the dryness of the air of the shed
or room they are hung in, and partly on whether they
were harvested durmg a drought or at any rate in
sunshine ; or in rain or damp. It is hardly necessary
to say that, if it can possibly be avoided, plants should
not be harvested during wet or damp weather. I
did this one year, and the recollection of the
unpleasantness of handling mouldy pods, and of the

BREEDING

156

destruction of a great part of the crop by the ger-
mination of the seeds in the pods and their subsequent
drying, is not likely to allow me to make the same
mistake again.

To keep the plants secure from the attack of
birds they must be hung up indoors. To keep them
secure from rats and mice is a more difficult matter.
The plan I have found to answer this purpose is to
suspend the bundles on stout wire stretched from
one wall of a disused stable to the other, taking care
that the bundles do not touch the walls or any other
object from which rats or mice could get at
them.

For hanging the bundles on to the wire I have
found a form of picture -hanger so useful that a better
thing for the purpose could not have been invented.
This is the brass picture-hanger which has a broad
hook at one end from which the picture is sus-
pended, and is itself nailed to the wall by three nails
which pass through three holes, one in each of the
three subdivisions of its trifoliate base. In the use
to which I put it, the hanger is hung the other way
up ; it hangs from the wire by the hook, and a loop
from the twine suspending the bundle is passed over
the two lateral subdivisions of its base.

If the weather is hot, the peas will probably be
dry enough in ten days’ time for the business of
recording to be started. But it is well to give them
time to become thoroughly dry; say a fortnight or
three weeks.

k word as to the patch of ground on which the

EXPERIMENT WITH A PLANT 157

peas have been grown. When the last bundle has
been taken in, it is my practice to dig and well manure
the ground as soon as possible, and to plant it at
once with vetches, to be dug in, in January, as
a green manure. For sowing the vetch, the ground
is left fairly rough, and shallow drills, one foot
apart, are drawn across it ; the seed is sown
broadcast, and the ground is then nicely raked
over with a fine rake. If the weather has been
favourable it should be possible to finish the har-
vesting and get the vetches planted by the end of
August ; and, of course, the earlier they can be
got in the better, for the more growth will they have
made when the time comes to dig them in. Besides
acting as a valuable green manure, the vetches serve
other useful ends. To a certain extent they keep
weeds down by smothering them; and, by keeping
the frost out of the ground they render it possible
to begin tilling the grormd at an earlier date in the
year. In January of this year, 1911, a frost, which
had left a piece of ground, adjoining the patch
in which the vetches were, so hard that it was
impossible to work it with a spade or a fork, had not
touched the ground in which the vetches grew. This
was perfectly soft and open, and the vetches were
dug in with the greatest ease.

My practice is not to use this patch again for
peas before two years have elapsed.

The reader who wishes to repeat Mendel’s obser-
vations will require to know the names of those
varieties which bear the various characters dealt

158 BREEDING

with by Mendel. I give this information in the
following Table : —

Character

concerned

Dominant
Member
of Pair

Name

of

Variety

Recessive
Member
of Pair

Name

of

Variety

Shape of
Cotyledon

Bound

Express

Wrinkled

Laxton’s Alpha

Colour of
Cotyledon

YeUow

Sangster’s
No. 1

Green

Express

Colour of
Seed-coat

Grey

Grey Field
Pea

White

Songster’s

No. 1

Texture
of Pod

Hard

Grey Field
Pea

Soft

French Sugar
Pea (Messrs.
Suttons)

Colour of
of Pod

Green

Sanqster’s

No. 1

Yellow

Gold von

Blocksherg
(Haage und
Schmidt,
Erfurt, Ger-
many)

Distribution
of flowers
on stem

Normal

Sanqster’s
No. 1

Fasciated

Mummy Pea
(Henry Eck-
ford, Wem,
Shropshire)

Stature

TaU

Carter’s

Quite

Content

Dwarf

William Hurst

The varieties for which no source is specified in
this Table can be supplied by, or obtained through,
any established seedsman. Some care must be taken
with the Grey Field Pea. The term “ grey ” is

EXPERIMENT WITH A PLANT 159

sometimes used among seedsmen in so wide a sense
as to signify peas with a maple seed-coat. The
strain which I have used were given me by Mr.
R. H. Lock. In the case of all the pairs of characters
excepting that relating to the distribution of flowers
on the stem, the two varieties named difier from one
another only in respect of the two characters which
they are stated in the Table to bear.

The simplest way of obtaining a result to illustrate
the ratio obtained when two pairs of characters are
involved {see Plate III.) is to cross a yellow wrinkled
variety with a green round one. A good yellow
wrinkled one to use is British Queen^ and a good
green round Express,

The best way of exhibiting the results of crosses
involving characters either of the seed-coat or of
the cotyledon is to mount them in cotton wool in
glass-topped cardboard boxes about three-quarters
of an inch deep. Figs. 11, 12, 21, 26, and 27, and
Plates I. and III. are direct photographs from speci-
mens mounted in this way. The method is as
follows : Plenty of cotton wool is laid in the box ;
half of each of the pods is cut ofl at the base, and
the peas attached to it are gummed into the other
half. The specimen is now laid on the cotton wool
and the lid put in its place and kept there with
seccotine. The whole process requires patient and
careful manipulation.

CHAPTER XI

Mendel’s explanation of his results

In this chapter we pass from the description of facts
as to which men with unimpaired sense of sight
are all agreed, to the consideration of hypotheses
about which men whose judgment estimated by the
usual standards is sound are widely at variance.
At this point, therefore, we bid farewell to the broad
daylight of observation, and enter the dark and
treacherous alleys of inference. Before we take
this step, it is well to reflect on the nature of the
errand on which we are engaged.

A theory commonly held as to the nature of the
explanation of a sequence of events is that explana-
tion consists in the discovery of the principles under-
lying and determining this sequence. The essence
of this theory is that explanation consists in seeing
below what is observed, and flnding out how it is
worked ; and this probing downwards is regarded
as one step of a grand downward journey of explora-
tion, the goal of which is the discovery of the causes
of all phenomena. Those who hold this theory of
the nature of explanation believe that each downward
step is one stage nearer to the truth below — a day’s
march nearer home.

Now, if this view of the matter were the true

160

MENDEL’S EXPLANATION i6i

one it would follow that the deeper we probed,
that is, the nearer we got to the underlying truth,
the more closely we should agree. The very reverse
of this is, however, the case. The further we leave
the crust of phenomena behind, and the more we
approach fundamentals, the more are we at variance
with one another. The extent to which we agree
with one another is not directly, but inversely, pro-
portional to the depth to which we probe. And
when we reach the most fundamental questions of all,
such as the theory of the relation between spirit
and matter, so far from finding unanimity, we
find the profoundest and widest gulf of disagree-
ment.

True it is, that when we try to explain the things
which reach us through our senses, we leave the crust
of indisputable phenomena behind. But we do not
leave it behind, above us, by burrowing beneath it ;
we leave it behind in the other direction by retreating
into the recesses of our own minds. And the further
we retreat into these recesses, the more does our
explanation become an expression of our own mind,
and the less relation does it bear to anything in the
world outside. That is why the ‘‘ deeper ” we probe
the more violently do we disagree with one another.
It is only so long as we stick to the crust that we
agree.

We may regard the facts described by Mendel as
established. The theory which he put forward to
explain them is probably a very close approximation

L

i62

BREEDING

to the true one. And it is given here partly for
its general biological interest, and partly because it
is true up to such a point that it is possible, by means
of it, to predict the results of all the simpler types
of matings which the practical breeder is likely to
want to make.

In the previous chapter we were confronted with the
fact that there must be something in the male repro-
ductive cell, or germ cell, as it is called, contained
in the pollen grain taken from a flower on a tall
plant, which made the offspring of a dwarf plant
fertilised by this pollen, develop into a tall plant.
Mendel’s theory is that every germ cell, whether
male or female, of a tall plant contains something
representing the character tallness. It is not known
what this something is, so it does not matter what
it is called. The term that has come into general
use is “ factor.” The germ cells of tall peas are,
therefore, said to contain the factor for tallness.
Similarly the germ cells of a dwarf pea contain
the factor for dwarfness. So that when a tall pea
is self-fertilised, a male germ cell containing the factor
for tallness unites with a female germ cell also con-
taining the factor for tallness, and the result of this
double dose of tallness in the fertilised ovule is that
it develops into a tall pea. Similarly, when a dwarf
pea is self-fertilised a male germ cell containing
the factor for dwarfness unites with a female germ cell
also containing this factor, with the result that a dwarf
pea is produced. In both cases, it will be noted, two
doses of the same factor are necessary for the pro-

MENDEL’S EXPLANATION 163

duction of the character represented by that factor
in the germ cell.

Now, what happens when a germ cell bearing the
factor for tallness unites with a germ cell bearing
the factor for dwarf ness ? We know the answer to
this question. A tall pea is produced. The fertilised
egg contains in this case, as in the others, two factors ;
but one is the factor for dwarfness, and the other
that for tallness. But the latter seems to dominate
completely over the former, and allows it no share in
determining the stature of the plant ; for the hybrid
tall pea which is produced is indistinguishable from
its pure tall parent, except that it possesses the
exceptional vigour which is characteristic of hybrids,
and is often a little taller than its pure tall parent. It
should be noted that the result of crossing “ female
dwarf ” with “ male tall ” is the same as that of cross-
ing “ male dwarf ” with “ female tall.” In technical
parlance, the reciprocal crosses are identical. We have
now to consider the condition in which these factors
exist in the germ cells produced by the hybrid tall. And
in so doing we get to the kernel of Mendel’s theory.
It might be expected that one kind of factor occurred
in all the germ cells, and that this factor represented
a mixture between tallness and dwarfness ; or it
might be concluded from the fact that the hybrid
itself was tall that all its germ cells contained the
factor for tallness, and that the dwarf factor had
been eliminated altogether. Neither of these views
were held by Mendel. He supposed that half of the
germ cells, produced by the hybrid, contained the

BREEDING

164

tall factor, and the other half the dwarf factor ; the
tall factor has not been adulterated with dwarfness
by its union with a dwarf factor, nor the dwarf by
the tall. The two kinds of factors occurring in equal
numbers amongst the germ cells of the hybrid are
as pure, in respect of the character they represent, as
the same two factors existing in the germ cells of
the parental tails and dwarfs, according to the theory.

Assuming this to be true, what result would be
expected to follow from the self-fertilisation of such
a hybrid ? Let us be perfectly clear, before we
attempt to answer this question, as to what we are
considering. On the one hand there are all the ovules
in all the flowers on a plant, amounting, if the plant
has been given plenty of room to grow, to anything
between two and flve hundred ovules ; half of these
ovules are supposed to contain tall factors, and the
remaining half, dwarf factors. On the other hand, there
are all the pollen grains on the plant. The numbers of
these must be immense, and I know of no estimate
of their number in a single flower ; nor is the matter
of any importance for our purpose. Suffice it that
the number of them is vast. Half of these pollen
grains are supposed to contain tall factors and the
remaining half dwarf factors. I shall speak of pollen
grains as synonymous with the male germ cell, to
avoid circumlocution. What is the result of the
fertilisation of these few hundred ovules by the few
hundred pollen grains necessary for the purpose ?
Let us consider first the ovules with the tall factor.
They have an equal chance of being fertilised by a

MENDEL’S EXPLANATION 165

pollen grain with a tall factor as by a pollen grain
with a dwarf factor ; that is to say, the following
unions will be, approximately, equally frequent : —

Tall factor (borne x Tall factor (borne
by ovule) by pollen grain)

Tall factor (borne x Dwarf factor (borne
by ovule) by pollen grain)

Now let us consider the ovules with the dwarf
factors. These again have an equal chance of being
fertilised by a pollen grain with a tall factor as by
one with a dwarf factor ; that is, the following
unions will be, approximately, equally frequent : —

Dwarf factor (borne x Tall factor (borne
by ovule) by pollen grain)

Dwarf factor (borne x Dwarf factor (borne
by ovule) by pollen grain)

All the ovules have now been accounted for,
and it will be seen that the result is four kinds of
unions of germ cells, when these unions are classi-
fied according to the factors contained in the two
germ cells imiting in each case. Moreover, these
four kinds of union are, on the average, equally
frequent. They may now be written in one list,
with their frequency written after them in per-
centage : —

Tall-Tall
Tall-Dwarf
Dwarf-Tall
Dwarf-Dwarf .

25 per cent.

25

25

25

»

if

99

These are the kinds of unions of germ cells classified

i66

BREEDING

according to the factors in the two cells uniting
in each case. We have now to consider what will be
the nature of the plants which develop from these
fertilised ovules. Let us deal, first, with the first and
last on the list. It may be supposed that the Toll-
Tall type will develop into a pure tall, and that
the Dwarf-Dwarf will produce a dwarf, inasmuch as
this was the theory of the reproduction of tails and
dwarfs with which we started. Anyhow, we shall
see how this supposition fits the facts.

Now, with regard to the second and third types
of union on the list. It should be remembered that
the character written first in each pair is that borne
by the ovule, and the character written second
that borne by the pollen grain. The Tall-Dwarf
couple, therefore, was produced by pollinating a pistil
on a tall plant with pollen from a dwarf; and the
Dwarf -Toll couple by pollinating a pistil on a dwarf
plant by pollen from a tall. Now, we know that in
the case of this character reciprocal crosses are iden-
tical ; that is to say, that the result of both of these
crosses is a hybrid tall plant. We are, therefore, now
in a position to say what will be the average result of
the random union of a large number of ovules, 50 per
cent, of which bear the factor for tallness and the
other 50 per cent, the factor for dwarfness, with a
similar number of pollen grains containing similar
factors in the same proportions. By “ random ” I
mean that germ cells bearing a particular factor do
not show a preference for germ cells bearing a par-
ticular factor, whether similar or dissimilar.

MENDEL’S EXPLANATION 167

Table to illustkate Mendel’s explanation of
Segregation in the 1:2:1 Eatio.

Types oj fertilised
ovules

Percentage

of

Nature of Plants
into which they

Percenlage

of

frequency.

will develop

frequency.

Tall-Tall .

25

Pure Tall .

25

Tall-Dwarf .
Dwarf-Tall .

25 1

25 j

Hybrid Tall .

50

Dwarf-Dwarf

25

Pure Dwarf

25

It will at once be seen that the proportions in
which the three kinds of plant — pure tall, hybrid
tall and pure dwarf — would be expected to occur on
the basis of this theory is exactly that in which they
do occur.

The fact that the theory is in perfect accord with
the facts which it attempts to explain is not, of
course, proof that the theory is true ; no one would
invent a theory which did not fit the facts. This
is, however, not the place to discuss the question
how nearly true it is, for we are concerned now with
Mendel’s theory as stated by him, and not with the
question whether it is true or not, which will be
dealt with in a later chapter.

It was stated, categorically, above that the four
kinds of unions of germ cells occurred with equal
frequency. The truth of this statement can be
tested in the following way : Two people each take
a coin, and they both toss it at the same time. This
act is repeated, say, a thousand times, and the result

i68

BREEDING

written down each time. The result of any given
pair of throws may, of course, be one of three things :
head-head, head-tail, and tail-tail. And the total
result of a thousand throws will not be very far
from 250 head-heads, 500 head-tails, and 250 tail-
tails.

In this illustration the matter of sex has been
left out, so that the equal frequency of the four types
of unions of germ cells seen in the first column of
the Table on p. 167 has not been illustrated. This
can easily be done. One of the persons who tosses the
coins may be designated the ‘‘ male,” and the other
the “ female.” If the sex of the tosser be taken into
account and recorded, it will be found that the four
following events will occur equally frequently. Male
will be denoted by the symbol c?, and female by
the symbol $ : —

Head tossed by and Head tossed by $

Head tossed by and Tail tossed by 5

Tail tossed by (J and Head tossed by 5

Tail tossed by (J and Tail tossed by $

Before we leave this illustration of the manner
in which the Mendelian ratio 1 : 2 : 1 is brought about,
a matter of great theoretical and practical importance
must be dealt with.

This ratio 1 : 2 : 1 in the second hybrid generation,
which is the essential and diagnostic feature of
Mendelian inheritance, is brought about by the
random union of male germ cells (50 per cent, of
which contain the factor for the dominant, and 50

MENDEL’S EXPLANATION 169

per cent, the factor for the recessive character) with
female germ cells containing the same factors in the
same proportions. The union, be it noted, is ran-
dom; it is what is called a chance result. And
inasmuch as it may be — nay, must be — important
to know, for practical or theoretical purposes, whether
a particular ratio under investigation is a Mendelian
one or not, a familiarity with certain attributes of
chance is the barest necessity to the student of
heredity.

Two things may be meant by the statement that
a thing happens by chance, according as to whether
reference is being made to the result of a single event,
or to the total result of a large number of such events ;
according, for instance, as to whether we are refer-
ring to the result of tossing a penny or to the result
of, say, ten thousand such tosses. Nothing, of
course, is more uncertain than the result of a single
throw, yet certain features of the result of a very
large number of throws can be predicted with great
accuracy. We are not concerned here with the first
signification of the statement that a thing happens
by chance, with the result of tossing a single penny
once ; and we can, therefore, proceed directly to
a consideration of those features of chance which
bear on the causes which determine the Mendelian
ratios.

As a familiar instance of an aggregate chance result
the result of tossing a penny a number of times
may be considered. It is stated that the result of
tossing a penny an infinite number of times is an

170

BREEDING

equal number of heads and tails, but as the truth of
this statement can never be tested, it does not concern
us. It is also stated that the nearer the number of
tosses approaches infinity the closer will be the
approximation to equality of the number of heads
and tails thrown. This can be tested, and concerns
us here very deeply; but before we deal with its
significance a minor feature of it, which is sometimes
forgotten, may be pointed out. As the number of
times a penny is tossed becomes larger and larger,
the absolute difference between the number of heads
and of tails becomes larger and larger ; but, on the
other hand, the ratio of the two approximates more
closely to the ratio 50% : 50%. Thus at 10 throws
there might be 6 heads and 4 tails, an absolute
difference of 2 ; but at 100 there might be 54 heads
and 46 tails, an absolute difference of 8 ; yet the
latter is a closer approximation to 50 per cent, heads
50 per cent, tails than the former. To sum up, the
essential feature of the aggregate result is that the
approximation to the expected result becomes closer
as the number of trials becomes greater.

Now, since the Mendelian ratio is the result of
the union of germ cells which unite without regard
to the factors which they contain, that is to say,
by chance, it is evident that this essential feature
will also characterise the Mendelian ratio. That is
to say, a very close approximation to the 25 per
cent, of recessives (which is the ratio which can be
tested on the largest scale) is not to be expected in
a small number such as that of the peas in a pod, or

MENDEL’S EXPLANATION 171

the children of a human family, any more than
three heads and three tails are to be expected when-
ever a penny is tossed six times. The statement
that the recessives appear in the ratio of 25 per
cent, to the whole second hybrid generation stands
the experimental test well, when it is tested on
a large scale. In the case of the colour of the
cotyledons in the pea, I raised a generation of
139,837 individuals ; the ratio in which the greens
occurred was 24.88 per cent. The point I wish to
insist on here is that when small numbers of indivi-
duals are being dealt with small deviations from the
expected number do not indicate that the Men-
delian theory does not apply to that case, any more
than a close approximation is proof that it does.
For instance, if a brown-eyed couple marry, the
fact that their first four children are brown-eyed
does not prove that both parents are pure-bred
brown. The next child might be blue-eyed, which
would show that both the parents were hybrid
brown. Nor would the production by a brown-
eyed woman of three brown-eyed and one blue-eyed
child prove that the father was, like the mother, a
hybrid duplex ; he might be a simplex, and the
family would, in that case, be exhibiting such an
approximation to 50 per cent, duplex : 50 per cent,
simplex as might be expected in so small a number.

Having dealt with this question, we may now
return to the Mendelian theory itself.

There is a way of illustrating the result of the
random union of the germ cells, which is in some

172

BREEDING

respects better than the tossing of coins. It is done
with counters. Equal numbers of red and white
ones are required. Only one pair of characters is
being taken into consideration in the case to be
illustrated. The female germ cells, half of which
bear the dominant member of. a pair of factors
whilst the other half bear the recessive member, are
represented by a collection of 100 red counters (repre-
senting the factor for the dominant character) and
100 white (representing the factor for the recessive
character). The male germ cells, containing similar
factors in the same proportions, are represented
by a precisely similar collection of red and white
counters.

Three columns are now ruled on a large sheet of
paper, to receive the three kinds of couples of counters
which can be drawn from these two collections.
These three kinds are, of course, red-red, red-
white, and white-white. A counter is now drawn
at random from one collection, and another counter
from the other, and the pair is placed in the column
prescribed for it on the paper, according as to whether
it is red-red (RR), red-white (RW), or white-
white (WW). Another pair is drawn in the same
way until, say, a hundred pairs have been placed
on the paper. It will be found that the numbers
of the three kinds of pairs approximate to the ratio
25 per cent. RR, 50 per cent. RW, 25 per centl WW.

The illustrative value of this device may be
greatly increased by making it a rule, when a red and
a white are drawn together, to put the red on the

MENDEL’S EXPLANATION 173

top of the white when the pair is placed in its
column on the paper. This is supposed to represent
the dominance of red over white ; and if the pairs
of counters are placed on the paper so carefully that
the lower counter is in each case concealed, the
conclusion derived from a superficial glance at the
paper with the counters on it, will be that reds and
whites occur in the proportion of 75 per cent, and
25 per cent, respectively. Closer inspection of these
apparent reds will reveal the fact that, of these
seventy-five, fifty consist of a red and of a white
counter, and correspond to the hybrid individuals
bearing the dominant character ; and the remaining
twenty-five consist of two reds, and correspond to
the individuals which both bear, and breed true to,
the dominant character.

This device brings home vividly to the spectator
the reason, according to the Mendelian theory, why
the recessive which appears in the second hybrid
generation, the extracted recessive as it is called,
breeds as true to the particular character under
consideration as the recessive with which the cross
was made. The extracted recessive breeds true for
precisely the same reason as the pure one does ; it
is the result of the union of like germ cells, or, to
speak more fully and accurately, of germ cells con-
taining factors for like characters. The factor for
dwarfness, for instance, contained in the ovule of a
hybrid tall pea is identical with such a factor borne
by a pure dwarf pea ; and when two such factors
borne by the hybrid tall unite and result in the

174

BREEDING

production of a dwarf, the dwarf character is no more
tainted with tallness than a white counter becomes
pink by lying in a heap with red ones.

The cause of the pure breeding of the 25 per
cent, tails appearing in the second hybrid genera-
tion, namely, the union of two germ cells containing
similar character-factors, is also well demonstrated
by means of these counters. So, also, is the fact
that the remaining 50 per cent, of the tails are
hybrids, and, moreover, that these hybrids of the
second hybrid generation will produce a third hybrid
generation exactly similar to that produced by the
hybrids of the first hybrid generation, for the reason
that the contents of the germ cells of both these
hybrids (of the second and first hybrid generation)
are identical.

It is furthermore plain why, according to Mendel’s
theory, this segregation of characters in these pro-
portions should continue as long as the seeds of
hybrid plants are sown. A hybrid, whether it occurs
in the first or the fiftieth hybrid generation, will pro-
duce germ cells half of which contain the dominant,
and the remaining half the recessive character ; and
the generations produced by them will, consequently,
have the same composition. But we must guard
against the error of assuming that Mendel’s theory is
true, and of then dogmatically stating, from the
standpoint of this assumed basis, that the segregation
in these ratios will be continued for all time. If
the theory is true they will. Let us, then, turn the
tables ; and test its truth by finding out if they do.

MENDEL’S EXPLANATION 175

The most convenient way of representing the
supposed causes of Mendelian segregation on paper
is to use the four-square table. Along the top of
the table are written the two kinds of factors that
occur in equal numbers amongst the male germ cells
of the hybrid, namely, the tall, for which the letter
T is written, and the dwarf, for which D is written.
The same is written along the left of the table, but
the T and D here refer to the factors in the female
germ cells.

c?

T D

TT

TD

DT

DD

Two squares alongside one another are called
a row. Two squares superimposed on one another
are called a column. The T above the whole square
relates to the column below it, and the D, to the right,
to the column below it. Similarly the T and D at
the left of the table relate to the rows to the right
of them. In the square formed by the intersection
of a row with a column is written the result of the
union of the factor at the head of the column with
the factor at the left of the row. This Table is
merely another way of stating what has already been

BREEDING

176

set forth on p. 165. But it is introduced here to
pave the way for the sixteen-square Table which is
indispensable for explaining the 9 : 3 : 3 : 1 ratio.

There are certain general terms and formulae
introduced by Mr. Bateson which have come into
general use. They may now be briefly indicated.

The letter D is used to signify the dominant
character, and R the recessive, and consequently
the hybrid is spoken of as DR, the pure dominant
as DD, and the recessive as RR. It will be noted
that the single letters D and R are stated to signify
characters, whilst the organisms bearing these
characters are represented by two flgures, DR, DD,
RR, etc. This latter fact is an expression of a part
of the Mendelian theory according to which the
organism is a double structure. Indeed, it may
almost be said that this is not a theory but a fact.
Every organism with which we are concerned in this
book is formed by the union of two germ cells. And
the simplest evidence of this duplicity is that the
number of chromosomes in the cells of an animal
or plant is twice that of the number in each of the
two germ cells which united to produce it.

A single word for a germ cell is wanted. This is
supplied in gamete,^ A more urgent need is a single
word for the result of the union of two gametes, the
animal or plant into which it develops. For this,
the word zygote\ has been coined.

In the illustration above, the single counter

♦ From the Greek ya/j.€w, marry,
f From the Greek (vy6y, a yoke.

MENDEL’S EXPLANATION 177

represented the gamete, and the pair the zygote.
The single letters at the top of the four-square Table
represented gametes, and the pairs in the squares
zygotes.

This Table may now be written in a generalised
form in which D and R represent dominant and
recessive characters in general. Thus ; —

(?

6 " K

DD

DR

RD

RR

Another form of such a Table, in which the capital
and small sizes of the same letter are used instead of
D and R to represent the dominant and recessive
characters of a pair, should be referred to here, because
it is necessary to use this form of notation in the
general form of the sixteen-square Table, which is
used to represent the union of gametes of hybrids
between forms differing in respect of two pairs of
characters. This form of the Table is given on the
next page.

An inspection of this Table, or a casting back of
the memory to the counters, will reveal the fact that
there are two kinds of zygotes. One is the result

M

BREEDING

178

of the union of gametes bearing similar character-
factors, and is called a Tiomozygote, the adjectival
form of which is homozygous : of this there are two
kinds, DD and RE. The other kind of zygote is

A

ot-

the result of the union of gametes bearing dissimilar
character-factors, and is called a heterozygote, the
adjectival form of which is heterozygous : of this
there is only one kind, DR.

The heterozygote may either bear the character
of its dominant parent, as in stature in peas; or it
may possess a character peculiar to itself and
inseparable from the heterozygous condition, as in
the case of the Andalusian fowl. In such a case
it is useless to attempt to fix the hybrid character.
A similar case amongst cattle is red-roan, which is
a heterozygote between red and white.

It now remains to consider the results of the
different types of matings which may be made
between the three types of zygotes, DD, DR, and
RR. It is hardly necessary to point out how impor-

(J gametes
A

AA

Aa

aA

a

MENDEL’S EXPLANATION 179

tant a finger-end familiarity with such results is,
both to the student of heredity and to the prac-
tical breeder. It will be convenient, first, to display
them in tabular form, so as to obtain a general view
of them, and to see which have already been dealt
with, and which remain to be considered.

Zygotes mated

Nature and r dative numbers, in
'percentage, of zygotes resulting from
preceding union

DD

DR

RR

1

DD

DD

100

2

RR

RR

100

3

DD

RR

100

4

DR

DR

25

50

25

5

DR

DD

50

50

6

DR

RR

50

50

The first four of these types of mating have
already been considered. No. 1 represents the true-
breeding of pure dominants, according to the fol-
lowing scheme : —

(J gametes
D D

-2

o>

a

o3

W)

0

D

DD

DD

DD

DD

D

i8o

BREEDING

No. 2 represents the true-breeding of pure recessives,
according to a precisely analogous scheme in which
R is substituted for D. No. 3 represents the cross-
ing of DD with RR, according to this scheme : —

R

<v

a

c3

fcC

Of

R

(J gametes
D D

(J gametes
R R

DR

DR

o

gametes

O

DR

DR

DR

DR

DR

DR

and No. 4 represents the mating together of these
hybrids, which has already been represented in this
scheme on p. 177. Nos. 5 and 6 have not yet been
dealt with, and are very important for more reasons
than one.

One feature which these two types of mating
possess in common is so simple that even when it
is noticed at all it is usually passed over with a
few words. Yet this feature has, in my opinion, a
very important bearing. Although these two types
of mating are usually considered last, as a not very
important corollary to the Mendelian theory, they
possess this feature in common, that of all the six
matings, in the above Table, in which the forms mated
differ from one another (i.e. in all except 1 and 2)
they are the only two in which the two parental

MENDEL’S EXPLANATION i8i

forms are reproduced in the next generation in equal
numbers. In No. 3, one parental form, the recessive,
is not reproduced at all. In No. 4, where the two
parents, the hybrids, are similar, they only account
for a half of the next generation, the remaining half
being made up of the grand-parental dominants and
recessives in equal numbers.

This extremely simple feature of matings. Nos. 5
and 6, is also characteristic of the inheritance of
sexual characters ; the result of the union of a male
with a female is the production of equal numbers
of males and females on the average. And this fact,
which the founders of the Mendelian theory have,
until recently, neglected, has become the corner-stone
of the Mendelian theory of the inheritance of sex,
and will also, in my opinion, play a very important
part in elucidating the origin of the Mendelian mode
of inheritance.

These two types of mating may now be dealt
with separately. No. 5 is represented on the four-
square Table as follows : —

D

09

a

bc

Of

D

gametes
D K

DD

DR

DD

DR

i82

BREEDING

This type of mating is not one of any practical
value, but its existence should be borne in mind,
because superficially the results of it are indistin-
guishable from the result of mating two dominants
(DD and DD) in those cases in which the character
of the hybrid is the same as that of the dominant ;
and if this fact is not borne in mind the false con-
clusion might be drawn from the results of breeding,
that a dominant character had become fixed when
it really had not. Take, for instance, the duplex
eye in man ; the result of mating a hybrid duplex
with a pure one is, as far as we can see from the
immediate result of the mating, the same as that of
mating two pures, namely, the production of nothing
but duplexes ; the fact that half of them are hetero-
zygous, or hybrid, and half homozygous, or pure,
can only be determined by subsequent breeding. And
so long as DR is mated with DD, and for however
many generations, the apparent result will be the
same as mating DD with DD. What, then, it may
be asked, does it matter whether DD is mated with
DR or with DD, if the result is the same ? The
answer is that it does not matter provided that the
union of DR x DD can be ensured, but as this is a
laborious matter (involving as it does the testing
of both forms before they are mated) the union
DR X DR is sure to occur sooner or later ; and then
the fat is in the fire, and a quarter of the generation
produced consists of recessives.

Mating of type No. 6 is represented in the four-
square Table on the next page.

MENDEL’S EXPLANATION 183

R

BQ

I

I

bo

Of

R

(J gametes
D R

DR

RR

DR

RR

This type of mating is of great practical value
because it is the one by means of which the question
whether an animal or plant bearing the dominant
character is heterozygous or homozygous can be
most expeditiously decided. This question could,
it is true, be answered by mating the individual
in question with another hybrid, if a hybrid of the
first generation were at hand (a hybrid of the second
generation can only be known to be such by testing
it, and that is what we are trying to do) ; and even
if this were available an inconveniently large number
of offspring would have to be produced before there
were sufficient grounds (derived from this test) for
believing that the individual tested was not a hybrid.
Directly a recessive is produced by mating a par-
ticular individual, which may be called A for brevity,
with a hybrid, it is certain that A is a hybrid ; but
the production of a family of six containing no
recessives does not by any means prove that A is
a dominant ; the next child might be a recessive.
But the probability of A being a dominant becomes

i84 breeding

less and less as the size of a family containing no
recessive increases.

By far the most convenient way, as stated at
the beginning of the last paragraph, of testing the
zygotic constitution (that is, of determining whether
it is heterozygous or homozygous) of an individual
bearing the dominant character is to mate it with
a recessive. There are two advantages of this over
any other way ; the first and most practically impor-
tant is that an individual bearing the recessive
character can only be one thing, RE, in contra-
distinction to one bearing the dominant character,
which can, of course, be DD or DR ; so that the
recessive can be used at once, without any pre-
liminary time-wasting testing which is necessary to
determine the constitution of an individual bear-
ing the dominant character. The recessive wears
its gametic constitution on its sleeve, and no further
credentials are required. The second advantage of
testing by mating with a recessive is that the result
of mating a hybrid with a recessive is the production
of twice as many recessives as are produced by the
union of two hybrids, 50 per cent, as against 25 per
cent. ; and consequently twice as great a likelihood
of the form tested throwing a recessive, if it is a
hybrid, as there would be if it were mated with a
hybrid.

My hybrid mice of the first hybrid generation
were mated with albinos, not to test their hybridity,
for that was obviously not in question, but to test
the truth of the statement that hybrids mated with

MENDEL’S EXPLANATION 185

recessives produced hybrids and recessives in equal
numbers. The actual numbers produced by mating
hybrids and albinos were 378 hybrids and 368 albinos.
The following list of families, which were the first
ten raised, will give some idea of the distribution of
recessives over the families. The largest family in
the whole experiment in which no recessives occurred
was one which consisted of six hybrids.

Pakents

Offspring

Female

Male

Hybrids

Albinos

Albino

piebald grey

4 grey

None

Albino

piebald grey

1 grey, 2 piebald
greys, 1 pie-
bald black

5

Albino

piebald grey

6 greys

4

Albino

grey

2 greys, 1 black

2

Albino

grey

4 greys

1

Albino

grey

2 greys

2

Albino

piebald grey

1 black

5

Piebald grey

albino

1 piebald grey,
2 blacks

2

Grey

albino

1 grey, 1 black

1

Grey

albino

2 piebald blacks

4

CHAPTER XII

THE ANTECEDENTS AND FATE OF MENDEL’s
DISCOVERY

It may not be without interest to consider here how
it came about that Mendel’s discovery, made in
1865, remained unknown until 1900, and to empha-
sise the fact that what constitutes a great discovery
is not the mere unfolding of a sequence of events
before the eyes of an observer, but the appreciation
of the significance, or perhaps the invention of a sig-
nificance, or the reading of a meaning into, this
sequence by the observer.

To deal first with the first question. The prob-
lem of heredity, i.e. the question how the resemblance
between one generation and the next is brought
about, as it presented itself to biologists interested
in the matter at the time when Mendel worked, was
to determine how the characters of the parent got
into the germ-cells which produced the next genera-
tion. It was natural that the problem should pre-
sent itself in this way ; that the inquirer should start
with the grown parent, and wonder how its charac-
ters could be compressed into the minute germ, sub-
sequently to emerge and expand and develop into
the next generation. For, compared with the grown

MENDEL’S DISCOVERY 187

body of the parent, the germ was a recent dis-
covery. It was natural, therefore, that the fixed
starting-point in the inquirer’s mind should be
the body, and that the problem of heredity should
present itself to him as the task of finding out how
its characters could get into the germ-cells which (so
he had recently discovered) it produced, and which
give rise to the next generation.

There can be no doubt that that was the way
in which the problem presented itself to Darwin’s
mind. His theory of Pangenesis is an answer to
the question. How do the characters of the parent
get into the germ-cells which it produces ? He
imagined that this was brought about by every cell
of the body giving off a particle which somehow
reached the germ-cells, so that each germ-cell con-
tained a representative particle from every cell of
the body which housed it. This theory would account
not only for the inheritance of so-called inborn
characters, but also for so-called acquired ones ;
because if a dog had its tail cut off, the tail would not
be represented in the germ-cells of the dog unless
the particles had left the tail before it had been
amputated. Pangenesis died, not because Galton’s
experiments designed to test it had a negative
result, but because Weismann’s writings effected a
swinging round of biological opinion, through 180
degrees, to a statement of the problem of heredity
which was the diametric opposite to that which
had prevailed hitherto. The doctrine preached by
W eismann was that to start with the body and inquire

i88

BREEDING

how its characters got into the germ was to view the
sequence from the wrong end ; the proper starting-
point was the germ, and the real question was not
“ How do the characters of the organism get into
the germ-cells which it ^produces ? ” but “ How are the
characters of an organism represented in the germ
cell which produces it ? ” Or, as Samuel Butler has
it, the proper statement of the relation between suc-
cessive generations is not to say that a hen pro-
duces another hen through the medium of an egg;
but to say that a hen is merely an egg’s way of pro-
ducing another egg. According to Weismann, the
problem how the characters of the body get into
the germ- cells which it encloses is as unreal as the
problem which puzzled one of our kings — how does
the apple get into the dumpling ? The answer in both
cases is the same. The apple does not get into the
dumpling. The characters of the parent do not get
into the germ-cells which they enclose. The apple
and the germ- cells were there first. The germ- cells
are merely part of an unbroken line of germ plasm
which under certain circumstances, usually the fusion
of two of its constituent germ- cells, froths up and
produces a great excrescence, the body of the next
generation, and continues its existence in this body.
The germ plasm, according to this view, is immortal ;
the excrescence, the body, is mortal. The germ plasm
only comes to an end along a particular line when
the body containing it dies without leaving offspring.
The animal does not, as it appears to do, give rise to
germ-cells when it reaches maturity ; the germ-cells

MENDEL’S DISCOVERY 189

are not formed afresh by each generation. When
the fertilised egg divides, some cells are at once set
apart to become the germ-cells of the organism into
which this fertilised egg develops. It is the body
which is formed afresh at each generation : the germ
plasm has existed since the dawn of life. The
reader will, of course, understand that this is not
a statement of fact, but a presentation of the
Weismannian doctrine.

The relation between Mendelian theory and this
doctrine may be considered from two points of view.
In the first place, a general acceptance of Weismann’s
way of stating the problem of heredity was neces-
sary before the significance of Mendel’s observation
could be understood. In the second place, Mendel’s
discovery was a confirmation and amplification of
Weismann’s thesis.

These two points will be dealt with separately.
The publication of Mendel’s paper in 1865 was the
throwing of pearls before swine. The problem of
heredity, as it presented itself to those who were
interested in the matter at that time, was, as I have
already stated, to explain how the characters of an
organism got into the germ- cells which it produced.
To Mendel, on the other hand, the problem was so
obviously the converse of this — namely, the question
how the characters of an organism are represented
in the germ-cells which produce it, that he did not
stop to discuss the matter. Mendel was so sure that
this was the correct statement of the problem that
he made no defence of it, and proceeded at once to

BREEDING

190

account for his results on the assumption of its cor-
rectness. He knew it so well that he did not know
he knew it. But those who read Mendel’s papers
were still labouring to find out how the characters of
an organism got into the germ-cells which it pro-
duced. The result was that no point was seen in
Mendel’s theory, and it quickly lapsed into the oblivion
from which it was not rescued till 1900.

It is not as if Mendel were out of touch with the
recognised representatives of biological orthodoxy.
He was in correspondence with Carl Nageli, to whose
criticisms Charles Darwin paid more attention than
to those of any other of his critics. This correspondence
has been published by Prof. Correns. It includes a
patient attempt by Mendel to make clear to Nageli
those points which the latter had not understood in
Mendel’s paper. Not only did he do this ; he also
sent packets of peas resulting from his experiments
to Nageli, in the assurance that if he (Nageli) grew
them he could not fail to perceive the significance
of the results which he (Mendel) had obtained. But
so deeply was Nageli imbued with his view of heredity
that Mendel’s explanations, and his seeds as well, were
as water on a duck’s back. “ These ” says Mr.
Bateson,* referring to Mendel’s letters and his illus-
trative specimens, “ must have utterly failed to arouse
his (Nageli’s) interest, for when in 1884, the year
of Mendel’s death, he published his great treatise on
heredity, no reference was made to Mendel or his
work. That this neglect was due to want of com-

* “ Mendel’s Principles of Heredity,” p. 55.

MENDEL’S DISCOVERY

191

prehension is evident from a passage where he
describes an experiment or observation on cats
which, as it happens, gave a simple Mendelian result.
The Angora character (recessive) disappeared in a cross
with a certain common cat whose hair character is,
as we now know, dominant. The • crossbreds were
mated together, and the Angora character reappeared
in one individual among a litter of common cats.
This typically Mendelian fact was thus actually
under Nageli’s own observation, but from the dis-
cussion which he devotes to the occurrence it is clear
that Mendel’s work must have wholly passed from
his memory, having probably been dismissed as some-
thing too fanciful for serious consideration.”

Mendel’s theory could not be grafted on to con-
temporary biological opinion : the cross was too

wide, to borrow Butler’s simile. But, grafted on to the
biological opinion prepared by Weismann, it flourished
like a green bay tree ; and its immense value was at
once recognised.

To regard, now, the relation between Mendel’s
and Weismann’s work from the second-mentioned
point of view.* In the first place, Mendel’s results
support Weismann’s doctrine, because the only theory
which can as yet account for these results leaves the
characters of the soma or body entirely out of account
and relates solely to the contents of the germ cells
which produce the generation whose character has
to be explained. On no other theory but Weis-
mann’s would an extracted green pea of the fifth

• Second paragraph on p. 189.

BREEDING

192

hybrid generation be identical in appearance and in
its breeding properties with a green of a pure strain.
Yet these two are identical, as will be shown on p.
236. Both Mendel’s and Weismann’s are “ germinal ”
theories of inheritance. Mendel’s a particular, and
Weismann’s a general one, on the basis of which
Mendel’s is erected. And Weismann’s general
doctrine receives strong support from the fact
that the predictions based on Mendel’s theory are
fulfilled.

Mendelian inquiry does more than afford evi-
dence of the truth of Weismann’s general theory ; it
fills in the details of the general theory of heredity
outlined in charcoal by Weismann. As has already
been said, Weismann was the first to point out that
the problem of heredity was to find out how the
characters of an organism were represented in the
germ-cells which produced it. Weismann showed what
the question was. Mendelian inquiry is slowly build-
ing up an answer to it. And already, in the ten
years which have elapsed since Mendel’s papers were
discovered, a rich store of information relating to
the manner in which characters are represented in
the germ-cells has been accumulated. I have en-
deavoured thus briefly to indicate what I conceive
to be the real nature of the respective parts played by
Mendel and Weismann in building up our modern
conception of heredity, because it is often com-
plained that credit which is due to Mendel is given
to Weismann, and that ideas which are described as
Mendelian were really introduced by Weismann.

MENDEL’S DISCOVERY

193

I will proceed now to a short description of a
hybridisation experiment identical with one of Men-
del’s which was begun in 1820, two years before Men-
del’s birth. The work was done by John Goss in the
garden shown in Fig. 34, in the village of Hather-
leigh, in Devonshire, which is eight miles from the
northern border of Dartmoor. The results of the
experiment were published in the “ Horticultural
Transactions ” in 1822, in a paper which is repro-
duced on pp. 199-202.

As will be seen from his third paragraph, he pro-
ceeded as directed in Chapter X., except that he
applied the pollen (of the dwarf pea) on the following
day. He obtained three pods of hybrid seed. In the
sentence beginning “ In the following spring . . .
the phenomenon of dominance is described, and in
the next sentence the phenomenon of segregation.
But Goss went farther than this. Not only did he
witness the phenomenon of dominance and of segrega-
tion ; he also observed the true-breeding of re-
cessives. “ Last spring I separated all the blue peas
from the white, and sowed each colour in separate
rows, and I now find that the blue 'produce ord'y
blue, . . .”t

But, as the words which follow show, Goss failed
to discover that some of the plants bore only yellow
seeds, and some both yellow and green. The reason
for this, I have little doubt, is that Goss did not

♦ Bottom of p. 199.

t In the colours of peas “blue” is synonymous with green, and
“white” with yellow.

N

194

BREEDING

record his plants separately, plant by plant, after
they had been pulled up and dried ; but looked at
a pod here and there as they stood in the ground, or
shelled the whole lot from the dried bundles without
regard to the individuality of the plants from which
the seeds came.

That Goss did not examine the plants separately
after they had been pulled up is evident from the
fact that he never speaks of plants, but only of pods ;
that when he does deal with things separately, as in
the case of the green and yellow seeds of the second
hybrid generation, he says so. This conclusion is
also supported by the fact that Goss was not seeking
for an interpretation of the phenomenon he was
dealing with, but was trying to raise a new vegetable.
‘‘ The edible qualities of this fea I have not tried,
having hut fewl^

How it came about that Goss, with the pheno-
mena ready to his hand, failed to make the discovery
which is now associated with Mendel’s name is a
question which naturally presents itself. In the first
place, it is, of course, more true in Goss’s case than it
it is in Mendel’s that biological opinion was not ready
to receive a theory such as Mendel’s, which lies on the
surface of properly recorded results of such an experi-
ment as Goss made. If Mendel failed to graft his
theory on to current biological theory, such an
attempt made by Goss would have been doomed to
failure from the outset. Nevertheless Mendel did
what Goss did not do ; he extracted something from
or projected something into his observations which

Fig. 34.— JOHN GOSS’S GARDEN; WHERE HE MADE HIS CROSSES IN 1820

MENDEL’S DISCOVERY 195

though at the time it fell on deaf ears, was at once
recognised as of immense value when it was again
brought forth and exhibited before a generation of
biologists which had been prepared by Weismann.
Why did Goss fail to do this ? The answer is mani-
fold. In the first place, he was not seeking for an
interpretation of his results. And, in the case of
interpretations, he who does not seek will not find ;
he may not find what he is seeking ; but if he does
not seek at all he will find nothing at all. Another
reason why Goss failed to make Mendel’s discovery
is that he did not record the proportion in which his
types recurred in the second hybrid generation. But
this is scarcely remarkable in view of the fact that
not only did he not know that it was a second hybrid
generation : he did not know that it was a genera-
tion. He says (last paragraph, p. 200) : “ Should
this new variety of pea neither possess superior merit
nor be deemed singular in its bicoloured produce
...” From this it is evident that he perfectly
naturally regarded the coloured parts of the seeds
borne by a pea plant as its fruit, and did not know,
what he could not be expected to know, that they
were the cotyledons of the next generation. The
recording of the numbers was an essential preliminary
to the suggestion of the theory put forward by Men-
del ; and in my opinion the great future advances in
this line of work will be made by paying close atten-
tion to numerical ratios, and testing them with statis-
tical formulae. Another reason why Goss failed where
Mendel succeeded is that he did not record the ofi-

BREEDING

196

spring of each plant separately. And the reason for
this was, as before, that he did not know that “ the
seeds ” were the offspring of a plant. Mendel, who
of course knew this, laid the foundations of the
modern analytical method in the investigation of
heredity, because he perceived the importance of
tracing the offspring of each individual separately,
and of keeping the records of an experiment in such
a way that the ancestry and progeny of every indivi-
dual concerned may be looked up from them. Much
has been talked of the essential antagonism between
those methods which deal with individuals in the
mass and the Mendelian or analytical method. But
I have been unable to find that any such antagonism
exists. The analytical method is the only one by
which hereditary processes can be unravelled ; whilst
the significance of the numerical results of these
analytical experiments can only be estimated by
statistical formulae. Both methods appear to me to
be indispensable.

Goss failed where Mendel succeeded. But the
lessons of failure are not less illuminating than those
of success, and the lives of those who fail not less
valuable or interesting than the lives of those who
succeed. Be this as it may, I was unable to resist
the temptation to satisfy my curiosity as to the life
of a man who, at any rate, spent part of his time in
work which has occupied much of my own. I
owe the following facts and the photograph of Goss’s
garden to the kindness of the Rev. J. W. Banks,
vicar of Hatherleigh, Devonshire.

MENDEL’S DISCOVERY

197

John Goss was born in Hatherleigh and baptized
on June 27, 1787. He is next heard of as a lad
who cleaned boots and did other odd jobs in the
Rectory at Iddesleigh, where the Rector took a num-
ber of pupils. One of these became interested in
Goss and helped to educate him with a view to his
entering the Church. But Goss had “ scruples,”
and he became an usher in a school. He was an
active and inventive man. He made an orrery, i.e.
a model of the solar system, a reading-book for
teaching the aged poor to read, and a calculating
machine. What a pity he did not conduct his
hybridisation experiments on so large a scale that
the services of this machine were required !

At the age of 26 he married, under romantic
circumstances. A lady came one day to Hather-
leigh to visit an old servant, and created a sensa-
tion in this out-of-the-way village by driving up
to her door in a post-chaise. On getting out of her
carriage she stumbled and would have fallen but for
the timely intervention of a young man, who stepped
forward and saved her. This was John Goss. He
was sent for to be thanked for his services. They
were married on March 6th, 1813. His wife was
twenty-nine years older than he was. At their
marriage he was 26, and she was 55. It was during
his married life that he made the crosses with peas
(in 1820) and sent his note of them to the Royal
Horticultural Society (1822). It may be that the
carrying out of this experiment was made possible
by a leisure which he owed to his wife. He speaks of

BREEDING

198

“ living retired ” in the country. It seems that his
horticultural work was successful, for he received
a prize (purse with gold) for his work from a hor-
ticultural society. His wife died, aged 65, on
January 19th, 1823, and he went to live at Oke-
hampton, where he died on May 31, 1833, at the
age of 46, leaving his money to establish libraries in
Hatherleigh and Okehampton.

There is a sentence,* part of which has already
been quoted, in Goss’s letter to the Horticultural
Society which possesses a curiously prophetic sig-
nificance : ‘‘ Should this new variety of pea neither
possess superior merit nor be deemed singular in
its bicoloured produce, yet there is, I conceive, some-
thing in its history that will emit a ray of physio-
logical light . . .” Little can he have guessed how
bright a ray was destined to be emitted by this
singular pea with its bicoloured produce.

• Last paragraph of p. 200.

The exception to the dominance of yellow over green related
in the “Note by the Secretary” on p. 201 is apparent only.
Mr. Bateson has discovered that the Imperials have a thick
green seed-coal^ which prevents the cotyledons of the first hybrid
generation being seen. Messrs. Sutton’s King Edward has a
similar opaque green seed-coat, which prevents the colour of the
enclosed cotyledons from being seen.

199

Facsimile jrom the Horticultural Society's Transactions*

t 234 ]

XXVn. Oti the Variation in the Colour of Peas, occasioned
hy Cross Impregnation. In a Letter to the Secretary. By
Mr. John Goss.

Read October 15, 1822.

Sir,

Living retired in the country, and having a taste for gar-
dening, I have been for some years past endeavouring to raise
new varieties of vegetables. ' A gentleman in the neighbour-
hood, seeing some of the fruits of my labours, put into my
hands the Transactions of your Society : this was like the
rismg sun after the dawn, and I was enabled to see, not only
how to do my work better, but that some things which oc-
cupied my attention had by others been already accom-
plished.

I have raised some Hew varieties of Peas, and as one of
these appears to be at least a singular production, and finding
very little on this subject in your volumes, I am tempted to
give you a description of it, accompanied with a few obser-
vations.

In the summer of 1820, 1 deprived some blossoms of the
Prolific blue of their stamina, and the next day applied the
pollen of a dwarf Pea, and of which impregnation I obtained
three pods of seeds.* In the following spring, when these

• I have not been able to as<Jertain with certainty the names by which the two
parent Peas ore usually known by gardeners and seedsmen; but I believe that
the Prolific Pea, which was the female parent of the new variety, is tha Blue
Prussian, and the dwarf Pea which was its male parent, the Dwarf Spanith.

200

Facsimile from the Horticultural Society's Transactions.

Variation in the Colour of Cross Impregnated Peas. 235

w^re opened, in order to sow the seed. I found, to my great
surprise, that the colour of the Peas, instead of being a deep
blu6y like their female parent, was of a yellowish white^ like
the male. Towards the end of the summer I was equally
surprised to find that these white seeds had produced some
pods with all blue, some with all white, and many \^h both
blue and white Peas in the same pod.

Last spring, I separated all the blue Peas from the white,
and sowed each colour in separate rows ; and I now find that
the blue produce only blue, while the white seeds yield some
pods with all white, and some with both blue and white Peas
intermixed.

The edible qualities of this Pea I have not tried, having but
few. It grows two or two feet and a half high, and attains ma-
turity about the same time with its blue parent, which it much
resembles, and unfolds a large, deep green rich foliage, su-
perior to any I have seen. It seems to require a greater
depth and richness of soil than other sorts, or than I have
given it.

Should this new variety of Pea neither possess superior
merit, nor be deemed singular in its bi-coloured produce, yet
there is, I conceive, something in its history that will emit a
ray of physiological light, or at least militate against an opinion
held by Mr. Salisbury, who, in his remarks on the anomaly
of the Peach and Nectarines growing on the same branch,
says,* “ I have not a doubt of the important consequences
which ensue, when the stigma of one plant imbibes pollen
l»elonging to another, but these are only manifested in the
succeeding generation.'*

• Horticultural Society’s Transactions, Vol. I. page 105.

201

Facsimile from the Horticultural Society's Transactions.

286 Vernation in the Colour of Cross Impregnated Peas.

That Ais is incorrect, is not here “ determined by a more
able physiologist,” but by the above statement — ^by the above
fact ; for the effect was strikingly evident in the seed, which
was changed by the impregnation from a deep blue to a yel-
lowish white. And if the seed undergoes such a change, why
may not the fruit ? It is more than probable that it does ;
but, the change not being so conspicuous as the difference
between blue and white, it has escaped observation. Perhaps
the most effectual way to unravel the mystery of the anomaly,
would be to impregnate a considerable number of Peach
blossoms with the pollen of the Nectarine, and to examine
minutely if any change succeeded in the fruit.

If this communication should be thought worthy to appear
in your Transactions^ I trust Mr. Salisbury will excuse an
humble attempt to reflect a little light on an obscure part of
a science in which he himself is so luminous.

1 am. Sir,

Y ours respectfully,

Hatherleighy Devonshire, JoHN GoSS.

October 5, 1822

Note hy the Secretary,

Previous to the receipt of the above communication, one
on the same subject was transmitted by Alexander Seton,
Esq. and read at the Meeting of the Society on the 20th of
August, 1822. Mr. Seton had happened to make a similar
experiment, by impregnating the flowers of the Dwarf Im-
perial, a well known green variety of the Pea, with the pollen
of a white free growing variety. Of the flowers so treated one

202

Facsimile from the Horticultural Society's Transactions.

By Mt\ John Goss^ 237

only produced a pod, and it contained four Peas, which did
not differ in appearance from the others of the female parent.
It thus appears from the different results in this stage of the
experiments made by Mr. Goss and Mr. Seton, that the aj)-
pearance of the fruit in the first instance is sometimes affected
by extraneous impregnation, and sometimes not ; the impreg-
nation being effectual, as it was proved to have been, by the
progeny in both of these instances, and it seems desirable that
this interesting point, on which so much difference of opinion
has arisen,* partly from observation, and partly from analogi-
cal reasoning, should be subjected to further experiment.

The plants which grew from the four Peas obtained by
Mr. Seton, seemed to partake of the nature of both parents,
being taller and more diffuse than the Dwarf Imperial, and
less so than the male white parent ; but the pods resembled
those of the former, being short, and having but few Peas in
each. On their ripening it was found that instead of their
containing Peas like those of either parent, or of an appearance
between the two, almost every one of them had some Peas of
the full green colour of the Dwarf Imperial, and others of the
whitish colour of that with which it had been impregnated
mixed indiscriminately and in undefined numbers ; they were
all completely either of one colour or the other, none of them
having an intermediate tint, as Mr. Seton had expected.
The representation of one of the pods in Plate IX. Fig. 1.
conveys a very perfect idea of its appearance.

* See observations on the accidental intennixture of character in certain fruit*,
at page 6S of the present Volume.

CHAPTER Xm

THEORY TO ACCOUNT FOR THE RESULTS WHICH FOLLOW
A CROSS INVOLVING TWO PAIRS OF CHARACTERS

The theory enunciated in Chapter XI., to account
for the fact of segregation and the ratio in which the
segregated characters occur has more an indirect
practical value than a direct one. Its value in
practice is the lesson which it teaches rather than
the immediate practical efiects which it is capable of
achieving. This lesson is that any attempt to predict
the result of a given mating must be based on a
knowledge of the character-factors contained in
the germ cells of the individuals mated, and not
merely on the visible characters of these individuals
themselves. The scientific control of breeding must
depend on a knowledge of characters as represented
in the germ cells ; no amount of familiarity with
the mere external features of the animals or plants
dealt with will assist towards this control, except,
of course, in so far as these external features afiord
a clue as to the contents of the germ cells within.
The chief value, then, of the theory enunciated in
the last chapter lies in the support which it affords
to those new principles of breeding which look for
guidance to the germ, and reject the indications of
the body or soma, and are based partly on Mendel’s

203

204

BREEDING

theory already described, and partly on de Vries’
work, referred to in Chapter I. These new prin-
ciples of breeding, together with an account of an
experiment carried out to test their truth, will form
the subject of Chapter XV.

But whilst the value of the theory enunciated
in the last chapter is of this indirect nature, the
theory to be described in this, which is merely an
extension of the first-named, is of immediate practical
value, inasmuch as a familiarity with it greatly
increases the facility with which one of the most
important methods of the breeder, the combination,
in one strain, of characters existing in distinct strains,
can be carried out.

The two methods of the breeder are, first, the
method of selection, which alters and improves the
characters of breeds ; and, secondly, the method of
hybridisation which merely effects the recombination
of characters which are not themselves altered. And
it may be said that the theory to account for the
results which follow a cross involving a single pair
of characters, described in the last chapter, throws
an indirect light on the first of these methods ; whilst
the theory to account for the results which follow
a cross involving two pairs of characters, to be
described in this chapter, directly facilitates the
second of these methods.

These preliminary remarks are intended to indicate,
roughly, the application of the theory which has
been, and that which is about to be, dealt with.
Both of them, of course, have other applications as

THEORY OF RESULTS 205

well, but these remarks will, it is hoped, help to
effect a preliminary orientation of the novice.

With a view to explaining the theory to account
for results which follow a cross involving two pairs
of characters, I propose to select as an instance of
such a cross the classical one investigated by Mendel,
namely, the result of crossing a yellow wrinkled with
a green round pea shown in Plate III.

A single-square Table is sufficient for the repre-
sentation of the union of gametes which occurs in
making the cross; though a four-square one might
have been used as on p. 175. Y is written for the
factor for yellow, G for green, R for round, and W for
wrinkled. The Table is filled up as follows : —

(J gametes (J gametes

YW GR

1

03

§ gr

YG.RW

or

09

I YW

YG.RW

C8

W)

Of

t)J0

0+

This may be simply written as follows : —

YW X GR = YG.RW

It will be seen that each gamete is represented
by two letters, and not one, as in the case of the cross
between two forms differing in respect of one pair
of characters ; and each zygote by four, and not
two, as in the other case. This is a simple indication
of the fact that the two forms crossed differ in respect
of two pairs of characters instead of one.

206

BREEDING

Our next duty is to predict the result of mating
these hybrids from a knowledge of the contents of
their germ cells. The first step towards this end is
to determine the number of possible combinations of
any two of the characters, yellow, green, round
and wrinkled, except such combinations as involve
the association in one gamete of the two characters
of a pair ; that is to say, except YG and RW. Also
there cannot, of course, be gametes with the formula
YY, GG, RR, or WW : it is only zygotes, which are
of double structure, that can contain two of the same
factor. With these exceptions the number of possible
combinations of the characters Y, G, R and W are
YR, YW, GR, GW. All that we have to do, there-
fore, to predict the result of mating the yellow-round
hybrids together is to determine the number and
relative frequency of the possible combinations of
these four tjrpes of gametes. And this is mostly simply
done by means of a sixteen-square Table. The for-
mulae for the four types of S gamete are written along
the top, and four similar formulae for the $ gametes
on the left side of the Table on the opposite page.

Here, again, four horizontal squares are referred
to as a row, and four squares superimposed vertically,
as a column.

Let us now see what the various kinds of zygotes
formed by the union of the four kinds of gametes
are, and write the formulae for each in the square
formed by the intersection of the column relating
to the male gamete concerned, the formula of which
is written at its top, with the row relating to the

THEORY OF RESULTS

207

female gamete concerned, the formula of which is
written at its end. Let us begin with the left-hand
square of the top row. This is a union between a
(? YR gamete and a ? YR gamete ; and the result,
YY.RR, is obviously a yellow round which is homo-
zygous both for yellowness and for roundness ; in other
words, both in colour and shape. All the zygotes
on the top row are yellow rounds. The second,
YY.RW, is, like the first, homozygous for colour,
but is heterozygous for shape. The third, YG.RR,
on the other hand, is heterozygous for colour but
homozygous for shape. Whilst the fourth and last,
YG.RW, is heterozygous for both colour and shape.

(J gametes

YR

YW

GR

GW

YR

YY.RR

1

YY.RW

2

YG.RR

3

YG.RW

4

YW

1

«

YY.RW

YY.WW

YG.RW

YG.WW

1

5

6

7

8

W)

Of

GR

YG.RR

YG.RW

GG.RR

GG.RW

9

10

11

12

GW

YG.RW

3

YG.WW

14

GG.RW

15

GG.WW

16

208

BREEDING

To proceed now to the second row. The first
zygote in it, to the left, is a yellow round, YY.RW,
homozygous for colour but heterozygous for shape.
The second, YY.WW, is a yellow wrinkled, homo-
zygous for colour and, of course, homozygous for
shape, inasmuch as wrinkledness is recessive. The
third is another yellow round, heterozygous for both
colour and shape ; whilst the fourth and last,
YG.WW, is another yellow wrinkled, this time
heterozygous for colour.

The first zygote in the third row, YG.RR, is a
yellow round, heterozygous for colour and homo-
zygous for shape. The second, YG.RW, is also a
yellow round, but heterozygous in both respects.
The third, GG.ER, is a green round, homozygous, of
course, for colour, green being recessive ; whilst the
fourth and last, GG.RW, is a green round like the
last, but heterozygous for shape.

The first zygote in the last row, YG.RW, is a
yellow round, heterozygous in both respects. The
second, YG.WW, is a yellow wrinkled, heterozygous
for colour only. The third, a green round, GG.RW,
heterozygous for shape only ; whilst the last is a
green wrinkled, homozygous for both of its characters,
both of them being recessive.

This Table is now written again on p. 209 with
the visible characters of the zygotes, instead of their
formulae.

It will be seen from this Table that there
are nine yellow round, three yellow wrinkled, three
green round, and one green wrinkled zygotes,

THEORY OF RESULTS 209

which is the actual ratio in which these types occur
in the second hybrid generation from this cross.

It may also be noted that the yellow rounds
occupy the top row, the left column, and the squares

YR

YW

GR

GW

along a diagonal crossing the whole Table from the
right top to the left bottom corner ; and that repre-
sentatives of each of the four types, YR, YW, GR,
GW, occur in the squares lying on the other diagonal
of the square, namely, that crossing it from the top
left to the bottom, right-hand corner.

Much more can, however, be made out from the
o

YR YW GR GW

YeUow

YeUow

Yellow

YeUow

Round

Round

Round

Round

1

2

3

4

Yellow

Yellow

YeUow

Yellow

Round

Wrinkled

Round

Wrinkled

6

6

7

8

Yellow

Yellow

Green

Green

Round

Round

Round

Round

9

10

11

12

YeUow

Yellow

Green

Green

Round

Wrinkled

Round

Wrinkled-

13

14

15

16

210

BREEDING

Table in which the zygotic formulae are written ; for
instance, in regard to this last point, the Table on
p. 209 only gives half the truth ; it only shows that a
representative of each of the four externally distin-
guishable types occurs along the top-left to bottom-
right diagonal ; the whole truth, as revealed by the
Table on p. 208, being that these four representatives
are the only representatives of these four forms which
are homozygous in respect both of colour and of
shape. A glance at the gametic formulae at the ends
of the columns and rows which intersect to form
these squares at once gives the reason. These are the
squares formed by the intersection of columns and
rows with the same gametic formulae at their ends.

Another point brought out by this Table is that
the zygotes lying in the squares on the diagonal
passing from the top right to the bottom left comer
of the Table, contain the four yellow rounds which
are heterozygous in respect both of colour and of shape.

It will also be seen that the four kinds of zygotes
in a row to the right of a given gametic formula are
the same, and follow one another in the same order
(reading from above below in the column, and from
left to right in the row) as those in the column under
the same gametic formula. In the YR column and
row, the squares 1, 2, 3, 4 are the same as 1, 5, 9, 13.
In the YW, 5, 6, 7, 8 are the same as 2, 6, 10, 14.
In the GR, 9, 10, 11, 12 are the same as 3, 7, 11, 15.
And, lastly, in the GW, 13, 14, 15, 16 are the same as
4, 8, 12, 16.

A familiarity with the simple properties of such

THEORY OF RESULTS

2II

a sixteen-square table as this is indispensable both
to the student of heredity and to the practical breeder.

Now, with regard to the breeding properties of
various zygotic types exhibited on this Table. These
may be most conveniently exhibited in tabular form : —

Number 0/
externoUy
distinguish-
able forms

ZygoticaUy
different tyyes

Number

Breeding properties, when allowed
to self -fertilise

(1) YY.RR

1

Produces yellow rounds

(2) YY.RW

2

only

Produces yellows, about

yR

(3) YG.RR

2

75 per cent, of which are
round, and the rest
wrinkled

Produces rounds, about

(4) YG.RW

4

75 per cent, of which are
yellow, and the rest green
Produces YR, YW, GR

1

and GW in the ratio

(5) YY.WW

9:3;3:1

Produces yellow wrinkleds

3 YW -

(6) YG.WW

2

only

Produces wrinkleds, about

75 per cent, of which are
yellow, the rest green

(7) GG.RR

1

Produces green rounds only

(8) GG.RW

2

Produces greens, about

3 GR -

J

75 per cent, of which
are round, the rest
wrinkled

1 GW

(9) GG.WW

1

Produces green wrinkleds

"1

only

212

BREEDING

I have recognised all these nine zygotic types in
my own breeding experiments.

It will be remembered that the cross which gave
this variety of forms was between a yellow wrinkled
and a green round. Two new types have arisen in
the second hybrid generation — the yellow round and
the green wrinkled. The latter may be counted upon
to, and actually does, breed true at once ; and it
only takes one generation longer to fix the yellow
round. All that is necessary, to do this, is to sow
the YR seeds, the cotyledons of which constitute
the second hybrid generation, so that the plants
raised have plenty of room, and then, when they are
ripe, to look for a plant all of the seeds of which are
yellow round. There should be one such plant
amongst every nine of the YR of the second hybrid
generation, and its seeds, when sown, will produce
nothing but yellow rounds.

If, instead of crossing a yellow wrinkled with a
green round, a cross had been made between a yellow
round and a green wrinkled, the same series of forms
would have been produced in the second hybrid
generation, and in the same proportions ; but the
two new forms in it, in this case, would have been
the yellow wrinkled and the green round. The
isolation of both these forms in a pure state takes
the same time as that of the yellow round described
above. The yellow wrinkled seeds, the cotyledons
of which constitute the second hybrid generation,
are sown. Ail the seeds on the plants produced will
be wrinkled, but in two out of every three plants they

THEORY OF RESULTS

213

will be yellow and green in the ratio of three to one ;
on the remaining plant they will be yellow wrinkled
only. These, when sown, will produce nothing but
yellow wrinkled. Similarly with the isolation of the
green rounds. The green rounds in the second hybrid
generation are sown ; two out of every three plants
produced will bear round as well as wrinkled green,
the remaining plant will produce green rounds only.
These, when sown, will do likewise.

It is not, in my opinion, necessary to give any
further instances of the application of this sixteen-
square Table to practical problems. What the
breeder wants is a working familiarity with this
Table. If the reader understands it properly and
has cases which can be dealt with by its help, there
will be no difficulty in applying it to them. It is
convenient for this purpose to have this Table in a
general form, as on p. 214. ‘‘ A ” and “ a ” stand

for the dominant and recessive members of one pair ;
“ B ” and b ” for those of another pair.

There is one point in regard to the criticism of
Mendelian theory which may be emphasised here.
The fact that the 9 : 3 : 3 : 1 ratio can be deduced
from a theory of the gametic contents of the germ
cells of the hybrid does not prove that theory to be
true, because not only can the 9 : 3 : 3 : 1 ratio be
shown to follow from the mere combination of two
3 : 1 ratios, but the existence, and frequency of
occurrence, of the nine zygotic types described on
p. 211 follows from the combination of two 1:2:1
ratios, as explained on p. 100. My point is this:

214

BREEDING

It is not permissible for tbe Mendelian to argue in
this way, “ I start with my theory to account for the
phenomenon of segregation in the 1:2:1 ratio ; from
this I derive the 9 : 3 : 3 : 1 ratio ; from this, in its
turn, I derive the 9:3:4; and so on. I am like a con-

AB

Of

aB

ab

jurer who balances successive objects one on the top of
the other. Surely if the first object were insecure,
the whole series would totter. Surely there could be
no better evidence of the truth of the theory which
constitutes the foundation than the fact that the
successive predictions based on it, and then on one
another, are fulfilled.” The answer to this is that
the fulfilment of the successive predictions would

(5 gametes

AB Ab aB ab

AA.BB

AA.Bb

Aa.BB

Aa.Bb

AA.Bb

AA.bb

Aa.Bb

Aa.bb

Aa.BB

Aa.Bb

aa.BB

aa.Bb

Aa.Bb

Aa.bb

aa.Bb

aa.bb

THEORY OF RESULTS

215

certainly support the theory in question if this theory
were the foundation on which they were based. But
it is not. The basis on which these two types, of
segregation, in the ratio 9 : 3 : 3 : 1 and 9:3:4,
rest is the phenomenon of segregation in the ratio
1:2:1, and not the theory by which it is sought
to explain this phenomenon. And the fact that the
predictions made from this basis are fulfilled is only
evidence of the reality of this basis, i.e. of the occur-
rence of the 1:2:1 ratio (about which, however,
there is no doubt), and it throws no light on the
theory put forward to account for this basis.

I propose to deal now with the general biological
conclusions which have been derived from the
theoretical considerations set forth in this chapter.

It will have been observed that we have dealt with
the interpretation of the mode of inheritance of only
two pairs of characters ; and, moreover, that these
two pairs of characters were entirely independent of
one another in their transmission. But it is obvious
that a very much larger number of characters is
concerned every time fertilisation takes place; and
it is a fact that distinct characters are not always
transmitted independently. Let us deal with the
general question of the number of characters first, and
with their relation to one another, whether indepen-
dent or otherwise, afterwards.

In the opinion of those who accept Mendel’s
theory as foreshadowing, if not as, in its present
state, actually constituting a valid theory of
heredity in general, the number of characters con-

2i6

BREEDING

cerned every time fertilisation takes place is certainly
very large ; it is nothing less than the sum total of
the characters of the organism in question. Accord-
ing to this generalised Mendelian theory, the organ-
ism is made up of a number of characters which are
called unit-characters, because they are transmitted
as independent units in inheritance. These unit-
characters were, in the early days of Mendelian
speculation, considered to be associated in pairs, but
as explained in Chapter IX., the pair is now regarded
as consisting in the presence of a particular character
as its dominant member, and the absence of this
character as its recessive member. But this is a
secondary feature of the theory. The essence of it
is that the organism is built up of an obviously
immense number of separately transmissible unit-
characters, the number, limits, and nature of which
can be determined by experimental breeding. With
regard to the soundness of this theory, all we know
at present is that it applies to the relatively small
number of characters which have been dealt with in
Mendelian studies. This knowledge is sufficient to
justify its application to practical problems, if there
is reason to believe that the inheritance of the here-
ditary characters under consideration is of, or approxi-
mates to, the Mendelian type. But this knowledge
is not as yet by any means sufficient to warrant even
the hope that the future problems of heredity will
be solved by its aid. I am, however, merely con-
cerned here in enunciating this theory, and not in
estimating its truth.

THEORY OF RESULTS

217

There remains one section of the theory to be
dealt with. This relates to what has been called
gametic purity. By this is meant the idea that
the distribution of unit-characters amongst the germ-
cells is discontinuous ; in other words that a germ
cell either contains the factor for a particular unit-
character or it does not ; there is no intermediate
condition, no half-way house. It is not possible for
a germ cell to contain a mixture of yellowness and
greenness ; that was why Y6 was not a permissible
combination of letters amongst the formulae of the
gametes outside the sixteen-square Table. Also, it
is not possible for a germ-cell to have more or less of a
particular character ; it has either got it or it has
not. That was why GG was not a permissible
formula for a gamete. For a zygote it is, of course.

There has been a good deal of misunderstanding as
to what is meant by gametic purity, and consequently
a good deal of unintentionally irrelevant criticism of
Mendelian theory. Let me indicate what the precise
significance of the Mendelian teaching on this point
is. One of the most practically important of the
corollaries which follow from this doctrine of gametic
purity is that the extracted dominants or recessives
of the second hybrid generation are identical in
respect of a 'particular character with their pure
dominant or recessive grandparents which were
mated to make the cross. The form which this
sometimes takes in the mouths of objectors is, “ The
extracted recessive or dominants are identical with
the pure dominant and recessive grandparents.

2i8

BREEDING

etc.” Now this is demonstrably untrue, and no one
familiar with Mendelian phenomena ever made such
a statement. Pure and extracted forms are only
supposed to be identical in resfect of 'particutar
character or characters.

For instance, the albinos which appeared in the
second hybrid generation of my experiment, though
to all appearances identical with the pure albinos,
proved themselves, when mated with waltzers, to be
very different from them in their breeding properties.
For amongst the offspring of extracted albinos mated
with waltzers there appeared pink-eyed and even
albino mice, forms which are never produced when
pure albinos are mated with waltzers ; and this fact
was hailed by objectors to the Mendelian theory as a
conclusive proof of the falsity of the doctrine of the
purity of the gamete. “We have met you on your
own ground,” they declared, “ and have defeated
you. Here were two albinos which ought, according
to your theory, to be identical. We tested the con-
tents of their germ cells according to your teaching,
and we have demonstrated that they are funda-
mentally different.”

But a familiarity with the Mendelian account of
reversion, to be given in the next chapter, will at
once explain how such a result could occur and be
in perfect accord with Mendelian theory as now
held. There were, in all probability, amongst the
albinos in my second hybrid generation three zygotic
types corresponding to the three zygotic types of
white-flowered peas in the second hybrid generation

THEORY OF RESULTS 219

of the experiment illustrated on Plate IV., and
interpreted on p. 226.

I mention these facts and their possible inter-
pretation merely to show the reader that results
which, in the mind of a casual observer, may appear
to be incompatible with the doctrine of gametic
purity do not in fact bear on it one way or the
other.

In this chapter attention has been paid to two
pairs of characters in which neither of the characters
of one pair exert any efiect on either of those of the
other. The theory to be enunciated in the next
chapter is based on the fact that the dominant
member of one pair exerts a very profound effect
on the dominant member of another pair ; it is a
necessary condition of its visible existence.

CHAPTER XIV

THE THEORY OF REVERSION

The Mendelian theory by which it is sought to
explain the phenomenon of reversion will be dealt with
in this chapter. By way of illustration, the result
of crossing a white-seeded with a grey-seeded pea,
and that of crossing a white -flowered with a pink-
flowered variety of pea will be taken (see Fig. 27
and Plate IV.).

The two pairs of characters in the colour of the
seed-coat with which we are concerned here are
purple-spot (P) and absence of purple-spot (p), and
grey (G) and absence of grey (g). According to
the theory, the character P has this remarkable
property, that the zygote which contains it cannot
manifest the purple-spot unless the character G be
present also. The formulje for the gametes of the
white-coated and grey-coated pea crossed are, for
the grey-skinned, Gp, namely, presence of grey and
absence of purple-spot; and for the white-skinned,
gP, namely, absence of grey and presence of purple-
spot, which, however, cannot be manifested in the ab-
sence of the grey-coat. Suppose the white-coated were
used as the female parent, the union of the gametes
and its result would be represented as follows : —

Gametes of white Gametes of grey
seed- coated $ parent seed-coated (J parent Zygote produced
gP X Gp = GgPp.

THEORY OF REVERSION 221

The purple-spot appears in the zygote produced
because the factor for grey-coat and purple-spot
co-exist in this zygote. The reason of its absence from
either parent will be seen from their zygotic formulae ;
the grey-coated is GGpp, i.e. grey but no purple-
spot ; the white is ggPP, i.e. purple-spot but no grey.

Germ-cells of four kinds, according to the character-
factors which they contain, would be produced by
the hybrid, namely GP, Gp, gP, gp, and the number
and kinds of zygotes produced by their random union
can be discovered, as in the case of the 9 : 3 : 3 : 1
ratio, by plotting their possible combinations on a
sixteen-square Table. This Table conforms precisely
to the general type given on p. 214.

(J gametes

GP

Gp

gP

gp

GP

GGPP

GGPp

GgPP

GgPp

1

2

3

4

GGPp

GGpp

GgPp

Ggpp

B

5

6

7

8

c3

W)

Of

gP

GgPP

GgPp

ggPP

ggPp

9

10

11

12

gp

GgPp

Ggpp

ggPp

ggpp

13

14

16

16

222

BREEDING

The external appearance of these four zygotic
types is given in the subjoined Table : —

Purple-

spotted

Grey

1

Purple-

spotted

Grey

2

Purple-

spotted

Grey

3

Purple-

spotted

Grey

4

Purple-

spotted

Grey

Purple-

spotted

Grey

Grey

5

6

Grey

7

8

Purple-

spotted

Purple-

spotted

White

White

Grey

9

Grey

10

11

12

Purple-

spotted

Grey

White

White

Grey

13

14

15

16

From this it will be seen that there are nine
purple-spotted greys, three greys and four whites.
This Table is precisely analogous to that for cotyledon
colour and shape. The purple-spotted grey characters
correspond to “ yellow round ” in that, in both of
them, two dominant characters are present in the
same zygote, either in the homozygous or hetero-
zygous condition. The greys correspond to the
yellow-wrinkled because they possess only one
dominant character, namely grey (in the DR or DD
condition) ; and one recessive, absence of purple-
spot.

THEORY OF REVERSION 223

The three whites — and here we come to the really
interesting part of the sixteen-square Table illus-
trating the union of gametes of reversionary hybrids
— ^the three whites in squares 11, 12 and 15 correspond
to the green rounds, and bear the same relation to
the greys as the green rounds bore to the yellow
wrinkleds ; that is to say, they possess the recessive
factor for the character in respect of which the grey
was dominant, namely, absence of grey ; and they
possess the dominant factor for the character in
respect of which the grey was recessive, namely,
purple-spot, which, however, is not seen in the seed-
coat because this is white and not grey — ^the presence
of grey, as already explained, being a necessary
condition for the manifestation of the purple-spot.
The white in the corner (in square 16) corresponds to
the green wrinkleds, possessing, as it does, two reces-
sive characters ; and it differs from the other whites
in that it is a pure white, i.e. it does not possess
the factor for purple-spot as the three other
whites do.

The parallelism between the Tables giving the
9:3:4 ratio and that giving the 9 : 3 : 3 : 1 is com-
plete in every particular. For instance, in the
former, as in the latter, the four zygotes homozygous
in respect of both their characters lie along a diagonal
passing across the Table from the top-left to the
bottom-right corner, whilst the four zygotes hetero-
zygous in both their characters (all purple-spotted
greys in this Table, as all yellow rounds in the other)
lie along the other diagonal.

224

BREEDING

The theory put forward to explain the results of
crossing a pink-flowered with a white-flowered pea
will now be briefly set forth on the same lines. This
case is given as well, because, for practical purposes,
it is more convenient than the seed-coats ; in the
first place, the characters, pertaining as they do to
the flowers, form more suitable objects for demonstra-
tion in the garden than the seed-coat characters ;
though these are very useful as permanent exhibits.
In the second place, for the purposes of experimenta-
tion, characters of the flower are much more conveni-
ent than those of the seed-coat, because in the case of
the former the character of the individuals being
mated can be seen at the time the cross is being
made — indeed, it cannot but be seen; whilst in
the case of the seed-coat characters the actual
characters of the plants mated cannot be determined
until they are dry and ripe, except indirectly and
imperfectly by the known fact that a white flower
is associated with a white seed-coat, and a purple
one with a grey or purple-spotted grey seed-coat.

The instances of flower colour and of seed-coat
colour also follow one another in a natural sequence,
inasmuch as the interpretation of the former is made
more readily intelligible by a familiarity with the
theory which is put forward to explain the latter.
For in the case of the flower colour it is unlikely that
the theory that the purple was due to the simultaneous
presence of two colour factors (namely, blue and pink)
belonging to distinct pairs, would have suggested
itself if the similarity of the composition of the

THEORY OF REVERSION 225

second hybrid generations in the two cases (9 purple,
3 pink, and 4 white in flower colour, and 9 purple-
spotted grey, 3 grey, and 4 white in seed- coat) had
not given the clue. In the case of the seed- coat
this interpretation is relatively obvious, inasmuch
as the two dominant characters, the purple-spot and
the grey background, can be seen as distinct and
separate things when present in the same zygote ;
whereas in the case of flower colour the one is super-
imposed on the other, the blue on the pink, or vice
versa, in such a way as to afiord no indication that
the resultant blend is compounded of two distinct
things.

The two pairs of factors involved in this case
are blue (B) and absence of blue (b), and pink (P)
and absence of pink (p). Blue bears the same relation
to pink as purple-spot does to grey — ^namely, that
the blue factor is unable to manifest itself in the
absence of pink. It can only come into being in a
zygote in which pink also exists, so that blue never
appears as such, because it can only exist in the
presence of pink, and the two together make purple.
But pink can exist in the absence of blue, just as
grey can exist in the absence of purple-spot.

If blue were not dependent on the presence of
pink for its development, the second hybrid genera-
tion, the types composing which are shown in
Plate IV, would have been 9 purple, 3 pink, 3
blue, and 1 white. Also the cross would have been
between a pink and blue, or between a purple and
white.

226

BREEDING

But to return to the case as it is. The union of
the gametes of the pink and white parents of the
cross may be represented as follows : —

Gametes of
pink-flowered
parent

Pb

Gametes of
white-flowered
parent

X pB

Zygote

produced

PpBb

The union of the gametes produced by this hybrid
would be represented as follows : —

(J gametes

c3

bO

Of

PB

Pb

pB

pb

PB

PPBB

PPBb

PpBB

PpBb

1

2

3

4

Pb

PPBb

PPbb

PpBb

Ppbb

5

6

7

8

pB

PpBB

PpBb

ppBB

ppBb

9

10

11

12

pb

PpBb

Ppbb

ppBb

ppbb

13

14

16

16

The external appearance of the sixteen zygotes
shown in this Table is given on page 228.

THEORY OF REVERSION 227

The relation between the constituent factors and
the resulting external appearance can be more readily
perceived by comparing the Table on p. 229 with
that on p. 228.

The factor for fink is written O

The factor for absence of fink <l

The factor for blue •

The factor for absence of blue C

In this Table, as in that representing gametic
unions in seed-coat colour, the four whites occupy
the four bottom right-hand squares. The three whites
in the squares 11, 12 and 15 carry the factor for blue,
as the Table on p. 229 shows. There are two types
of them. One of them. No. 11, is homozygous for the
blue factor ; the other two. Nos. 12 and 15, are hetero-
zygous for the blue factor. The only white carrying
no blue is the one in square 16. That is to say,
there are three zygotic types of whites altogether.

Now, it must not be supposed that the corre-
spondence between the theory, as set forth on this
Table, with the actual result which it was invented
to explain, is proof that the theory is true. It cannot
be denied that the expectation based on this theory
is the occurrence of four white, three pink, and nine
purple-flowered plants amongst every sixteen, on the
average, in the second hybrid generation. Nor can it
be denied that these three things occur in these propor-
tions. But a great deal more than this is wanted before
the truth of the theory can be admitted. The three
zygotic types of white, for instance, must be shown

228

BREEDING

to exist. In this case the theory is in advance of
our knowledge ; it is not yet known whether these
three types do exist. But this can be easily deter-
mined by mating all the whites which occur in such
a second hybrid generation with pure pinks. Let

Purple

Purple

Purple

Purple

1

2

3

4

Purple

Pink

Purple

Pink

6

6

7

8

Purple

Purple

White

White

9

10

11

12

Purple

Pink

White

White

13

14

15

16

us see what will be the result of mating the three
types of white with pure pinks the formula of
which would be PPbb.

The single white homozygous for blue with the
formula ppBB (No. 11)* will be considered first. The
union is of this type. The pink produces only
one kind of gamete, Pb, and the white also
produces only one kind, pB ; the result of their

• See Table on p. 226 or p. 229.

THEORY OF REVERSION 229

union is, therefore, always heterozygous purples.
Thus

pB X Pb = PpBb

The two whites heterozygous for blueness. Nos. 12
and 15, with formula ppBb, will produce two kinds

O

m

a»

a

c3

tJD

Of

(L

(L

gametes

o# oc (L m a c

of gametes, pB and pb, so the result of mating them
with pink can be most conveniently foretold by
means of the four-square table on p. 230, the external
characters of the four zygotes produced being shown
in the Table alongside.

The result of this union is, therefore, equal num-
bers of purples and pinks.

230

BREEDING

Gametes of Pink
Pb Pb

<D

ia pC

<u

a

zi

o

pb

PpBb

PpBb

Ppbb

Ppbb

Purple

Purple

Pink

Pink

The last remaining white, No. 16, which produces
one kind of gamete only (pb), carries no blue at all, so
that when mated with pink it will produce no purples,
but only pinks. Thus

pb X Pb = Ppbb

The total result of mating these three zygotic
types of white-flowered peas of the second hybrid
generation with pure pinks may be represented in
tabular form, thus : —

White Parent

Pink Parent

Percentage of
Purples produced

Percentage of
Pinks produ(^

ppBB

PPbb

100

ippBb

PPbb

60

50

j ppBb

PPbb

50

50

ppbb

PPbb

!

!

100

That is to say, equal numbers of purples and pinks
will be produced on the average, if a large number
of crosses are made ; and, if the theory expressed in

THEORY OF REVERSION 231

the sixteen-square Table on p. 226 be true, they must
be distributed as shown in the last two columns of
the foregoing Table. Nothing short of the carrying
out of an experiment of this kind on a large scale
will suj0&ce to prove the truth of the Mendelian theory
of reversion in this case.

Assuming that it will stand this test — and I have
reasons for believing that it will — we see that we
have here a consistent theory of reversion for the
first time. Reversion, according to this theory, is
due to the meeting in one zygote of the two factors
necessary for the production of the ancestral character.
These factors had, presumably, at some period
become separated and lodged the one in one and the
other in the other of the two strains which, when
they are mated, produce the reversionary hybrids.
In the case of flower colour in Pisum it is not
difficult to make a suggestion as to how this
occurred. The wild Pisum, of which I have grown
plants from seed kindly given me by Mr. Arthur
Sutton, has a purple flower. But this is not a hetero-
zygous purple but homozygous (i.e. of the formula
PPBB), because it breeds true to purpleness and
never produces pinks and whites. Where, then, did
the pink come from ? This question may be answered
by making only one assumption, which is amply
warranted by the frequency of analogous instances
of the same occurrence throughout the vegetable
kingdom. This assumption is that a white-flowered
variety of the pea arose by mutation from the purple-
flowered by the sudden and simultaneous loss of both

232

BREEDING

of the factors necessary for the production of purple.
It would have the formula ppbb. If mated with the
pure purple it would give heterozygous purples,
PpBb ; and if these were then self-fertilised and
set seed, they would produce purples, 'pinks, and
white. Another possibility, and perhaps a more
probable one, is that the homozygous purple sud-
denly mutated to the heterozygous condition ; the
self-fertilised seed of such a plant would produce
whites, purples and pinks. In this case it is not
necessary to assume, first, that a new form appeared
by mutation from the purple, and then was crossed
back with it. This question, however, of the origin
of Mendelian characters will be dealt with in Chapter
XVII. Ail that I am concerned with now is the idea
that the two factors necessary for the production of
purple in the power of Pisu'm did become separate
at some period, and one of them (P) lodged in a
pink-flowered strain, and the other (B) in a white-
flowered one, and that the reappearance of this
purple as the result of a cross is due to the re-union
of these two factors in the zygote resulting from a
cross between the two strains mentioned.

CHAPTER XV

SOME NEW PKINCIPLES OP BREEDING BASED ON
Mendel’s theory

How the theory of heredity of the future will differ
from that of the past can, perhaps, be most easily
made clear by describing an experiment as to the
result of which the expectation based on the Mendelian
theory and that based on the older notions which
it is likely to supplant are diametrically opposed to
one another.

The reader will not need to be reminded that the
result of crossing a yellow-seeded with a green-seeded
pea is a yellow, which produces a second hybrid
generation consisting of 75 per cent, yellow and
25 per cent, green. There is this proviso, however,
that this is only known to be true when the green
and yellow, with which the cross is made, belong to
pure green and pure yellow strains. But this should
make no difierence according to the Mendelian theory :
a yellow, provided its zygotic formula is YY, and a
green, provided its zygotic formula is G6, should,
whatever their ancestry, give rise to yellows which
will produce 75 per cent, yellows and 25 per cent,
greens, when self-fertilised. The expectation based
on the Mendelian theory is in violent opposition to
what the old notions of heredity would lead us to

233

234

BREEDING

expect, for according to these notions the ancestry
of two individuals mated, play a very large part
in determining the result of the mating.

The cross made was this. Instead of mating a
pure yellow cotyledoned strain with a pure green one,
I mated it with an extracted green of the fifth hybrid
generation ; that is to say, with the green printed
in italics in the following pedigree : —

Yellow X Green

I

Yellow

1

Hybrid

Pure

Yellow

Yellow

1

Green

1 .

Hybrid

Pure

Yellow

Yellow

1

Green

1

Hybrid

Pure Yellow

Yellow

1

Green

1

Hybrid

1

Pure Yellow

Yellow

Green

This green was mated with a pure yellow and, accord-
ing to the Mendelian theory, the result of the union
should be the same as that of a mating between a
pure yellow and a pure green. Before considering the
actual result, it will be well to look at the ancestries
of these two kinds of mating set forth, side by side, in
the form of pedigrees : —

SOME NEW PRINCIPLES 235

Ancestry of cross between
pure yellow and pure
green

GREEN Yellow

I I

GREEN Yellow

I I

GREEN Yellow

I I

GREEN YeUow

I I

GREEN Yellow

I I

Green x Yellow

I

Yellow

75% Yellow 25% Green

When the two parents of a generation are alike, their
character is only written once. The two forms crossed in each
case are printed in italics. The ancestors in respect of which
their pedigrees differ are printed in capital letters.

The result of the type of mating at the right is
well known, and is written there to emphasise the
fact that the result of the mating to the left is not
known.

Anyone unfamiliar with Mendelian phenomena,
and with only these two pedigrees to go upon, would
unhesitatingly declare that the result of the two
matings would not be the same. He would predict
with some confidence that, the result of the mating
to the right being what it is stated to be, the result
of the mating to the left will differ from it in showing
a much lower percentage of greens in the second
hybrid generation. “ I cannot but believe,” he
would argue, “ that the great weight of yellow

Ancestry of cross between pure yellow
and extracted green of the Fifth hybrid
generation

YELLOW X Green Yellow

I I

YELLOW Yellow

YELLOW Yellow

I I

YELLOW Yellow

YELLOW Yellow

. I I

Oreen x YeUow

I

?

BREEDING

236

ancestry behind the green in this case will make itself
felt by increasing the proportion of yellows and
diminishing that of greens in the second hybrid
generation produced.”

The actual result is entirely in accord with
Mendelian expectation. Our interest, of course, is
focused on the second hybrid generation. The first
is yellow, as in the normal cross. The ratio of greens
in the second hybrid generation is 24*88 per cent.,
as close an approximation as is ever obtained to
25 per cent. This ratio, 24*88 per cent., is based
on what is, I believe, the largest second hybrid
generation that has been observed. It consisted of
139,837 individuals (in this case, of course, seeds),
the actual number of yellow ones being 105,045,
and of greens 34,792.

To anyone brought up solely on Mendelian prin-
ciples— and it is possible that this book may, some
day, fall into the hands of such an one — this result will
not seem in the least unexpected. But I would remind
the reader that this is written at a time when the
number of people who have derived their notions of
heredity solely from teaching given in post-Mendelian
times (namely, since 1900) is very small, and that I
myself derived my ideas of heredity from teachers who
held exactly the same views on heredity as those
expressed in the last paragraph but one.

The conformity of the result with Mendelian
expectation is seen directly we pay attention, not to
the somatic characters of the individuals in the
pedigrees on the preceding page, but to the contents

SOME NEW PRINCIPLES 237

of their germ-cells. The ancestry of the extracted
green, marked with an asterisk, is given below in
a pedigree which includes the collateral ancestors,
and displays the zygotic constitution, and conse-
quently the germinal contents, of all the ancestors.
Outside the circles, which represent zygotes, are
printed to the right the somatic characters of the
individuals, whether yellow (Y) or green (G) ; outside
the circles, to the left, the ratios in which they
occur ; and inside the circles their zygotic formulae : —

75*12 Yellows, 24*88 Green

BREEDING

238

Only the ratio of the greens to the yellows is inserted
in the bottom line of the pedigree, because the con-
stitution of the yellows has not been determined.
Nor does it concern us here.

The result of this experiment, therefore, shows
that, if we confine our attention to one pair of cha-
racters at a time, the attempt to predict the result of
a given union must be based on a knowledge of the
factors representing these characters in the germ-cells
of the two individuals mated, and that the only know-
ledge of the characters of their ancestors which is
required is such as is necessary to discover what these
factors are.

Besides supporting this general thesis, the result
of this experiment tests another corollary of equal
practical interest which follows from the Mendelian
theory. It tests the truth of the assertion that an
“ extracted ” individual in any generation is as pure
in respect of the character which it bears as an
individual in a pure strain characterised by this
feature. For instance, it tests the truth of the
statement that an extracted green in the fifth hybrid
generation is identical in respect of its greenness
with a pure green, and it proves that this statement
is correct. Not only is the colour of such an extracted
green apparently identical with a green of a pure
strain, but its breeding properties are identical, as
shown by the result of mating it with a pure yellow.
This conclusion is merely, however, an instance of
the special application of the general conclusion set
forth in the last paragraph.

SOME NEW PRINCIPLES 239

There are two points of view from which this
conclusion may be regarded — ^the general biological
one and the strictly utilitarian one. If we regard
it from the former point of view, we have to ask
ourselves whether we are prepared to believe that the
activities of the soma play no part in determining
the structure of the soma in subsequent generations.
For my own part, I am by no means prepared to
state that I believe any such thing. In the first
place it is extremely doubtful whether any more
than a few hundred characters, the vast majority of
which exist only in a state of domestication, are
inherited in Mendelian fashion ; and, in the second
place, we have to remember that the changes to which
evolution is due have either been imperceptibly
small or, if not small, have been separated by immense
numbers of generations, and that, in either case, they
have extended over such prodigious periods of time
that it is improbable that any observations likely to
throw any light on their nature will be made in the
number of centuries during which public opinion will
remain stable enough to continue these observations
on the same lines, and still less in the number of
years during which a single man can devote his
attention to them.

If on the other hand the conclusion, stated at
the end of the last paragraph but two, be regarded
from the utilitarian point of view, comfort may be
derived from the reflection that most of the characters
of animals and plants which are of economic import-
ance seem to be inherited in Mendelian fashion ; and

240

BREEDING

also from the fact that the equipment and time at
the disposal of the biologist, though at present abso-
lutely inadequate for the purpose of grappling with
evolution as a whole, are probably sufficient for the
task of effecting such improvements in domesticated
races of animals and plants as may be required.

This being the case, our next step is to deal more
fully with the application of Mendelian principles to
the practice of breeding. I have already insisted on
what, in my opinion, is the most important service
which Mendel’s work has done for the science of
heredity, namely, the establishment of the principle
that the contents of the germ- cells, and not the
outward characteristics of the animals and plants
dealt with, must be our guide in breeding.

This principle can as yet only be applied when
the inheritance of the characters, which are being
bred for, fits, or may be expected to fit, into some
recognised scheme of heredity, such as the Mendelian.
If the mode of inheritance of the character in question
is demonstrably not Mendelian, and does not corre-
spond to any of the types recognised by de Vries
in the vegetable kingdom, it must be carefully
described and, if possible, interpreted ; though the
latter is not important. What is urgently needed is
an accurate description of the various ways in which
the characters of domesticated animals and plants
are inherited. It will be time enough to interpret
them when our knowledge of them is a great deal
more perfect than it is at present. For instance,
there is one question of the greatest importance to

SOME NEW PRINCIPLES 241

the breeder on which the present available infor-
mation is very scanty. It is the simple question of
the inheritance of bulk. In the pea, for instance, there
is a race with a seed as large as that of Victoria
Marrow, and also a race with a seed half its size,
as in some of the continental field peas. Now these
two characterisers, “ large ” and “ small,” do not
constitute the two members of a Mendelian pair ;
the cross between them is a blend, and so far as
data at present available show, segregation does
not occur when the hybrids are self-fertilised. If
one of these “ small ” races possessed, as one of them
happens to do, an economically valuable character,
and if it is desired to combine this character with
large ” size in a single strain, this would be an easy
matter if “ large ” and small ” constituted the
members of a Mendelian pair. But as — or let us say,
supposing — they are not, we cannot proceed to effect
this combination until the mode of inheritance of these
two characters has been determined. The most im-
portant point to discover in such an inquiry would be
to determine which, if any, of the types in the second
hybrid generation bred true. This generation might
consist of a complete series of gradations between
a “ small ” identical with the pure “ small ” at one
end, and a “ large ” identical with a pure ‘‘ large ”
at the other end ; and it might possibly be that these
two extremes of the series bred true, that the blends
in the middle of the series never did, and that the
forms intermediate between the blends and the
extremes bred true in some individuals and not in
Q

242 BREEDING

others. This is the kind of thing which would have
to be found out.

But if the characters bred for are inherited in a
Mendelian manner, there is no difficulty as to how
to proceed in practice, and, moreover, no question
but that the only guide which will conduct the breeder
to the end which he wishes to achieve is a knowledge
of the character-factors concerned, in the germ-cells
of the animals or plants under consideration. For
instance, it is clear that in the case of the blue colour
in Andalusian fowls, or the red-roan in cattle, the
attempt, based on the confinement of one’s attention
to outward bodily characteristics alone, to raise a pure
breed possessing these colours, by only mating animals
which exhibited them, is doomed to failure. True, the
Mendelian cannot tell the breeder how to raise a
strain which will breed true to these colours ; but
he can tell him how to raise 100 per cent, in each
generation where the breeder, acting on the principles
which he had learnt from his ancestors, only suc-
ceeded in raising 50 per cent.

With regard, now, to this principle of conducting
breeding on the basis of a knowledge of the contents
of the germ-cells. It may be asked, ‘‘How are we
to determine the nature of these contents ? ” The
answer in most cases is “ By breeding from the
individuals themselves ; to determine what are the
characters in respect of which they are homozygous
and what those in respect of which they are hetero-
zygous.” Indeed, the difference between modern and
old-fashioned principles of breeding may be sum-

SOME NEW PRINCIPLES 243

marised in the statement that whilst the old-fashioned
breeder looked for guidance to the pedigree of the
animals and plants he was breeding, the modern
one looks to their gametes ; the former looks to
their ancestry, and the latter to their offspring ;
the former looks backwards, the latter forwards. A
criticism which may be made of the modern method
is that it amounts to nothing more than predicting
what will happen by finding out what does happen.
But this is not, in fact, a fair criticism. In the antithesis
in the last paragraph ancestry was opposed to off-
spring and not to posterity — to one generation only
and not to many. The breeder, according to the
old principles, required a knowledge of the parents
and of as many of the ancestors as possible, and
based his prediction on this knowledge without going
beyond it. The breeder, according to the new prin-
ciples, only requires a knowledge of the offspring,
and he only needs this as an indication of the germinal
contents of their parents ; he does go beyond the
characters of the offspring to the germinal contents
of their parents, and bases his prediction on this,
and not on the characters themselves.

We have now to deal with another part of the
Mendelian doctrine, and with its application to
practice. This is the idea that the organism is
built up of independently heritable unit-characters.
I do not propose to consider here the question how
far it is likely that this conception will be found to
be applicable to all the characteristics of organisms.
That is a question to which, it seems to me, it is not

244

BREEDING

profitable to attempt to supply an answer until
information with regard to a much larger number
of characters of wild animals and plants than we
know about at present is available. But that this
conception does apply to a very large number of the
characters of domesticated races is sufficient for the
practical breeder. The general realisation of the
fact that the unit which has to be handled in experi-
mental breeding is not the individual animal or plant,
but the independently heritable unit-character, marks
the beginning of a new epoch in the history of the
practice of breeding. Once it is clearly understood
that dominance and recessiveness attach to char-
acters and not to individuals, one serious obstacle
to progress has been removed. And, in my opinion,
the most important advance which is rendered possible
by the removal of this obstacle will be effected by
a systematic inquiry into the nature and mutual
relationships of the characters, whether Mendelian
or not, which go to make up the organism. By
“ nature ” I mean whether they conform to the
Mendelian or to some scheme (which must then
become the subject of investigation) in the manner
of their inheritance ; and by “ mutual relationships ”
I mean whether they are transmitted independently
of other characters in inheritance or in association,
and, if so, how close the association is. This pheno-
menon of the association of character is a very
important one in practice, and may be illustrated by
characters of the culinary pea. The colour and
shape of the cotyledons are, as the reader knows.

SOME NEW PRINCIPLES 245

absolutely independent of one another. A pea which
is yellow is not for that reason any more or less Ukely
to be round. At the other end of the series there
are two such characters as the purpleness of the
flower and the purpleness of the axil, the angle
where the flower-stalk joins the main stem. These
two characters cannot be separated by hybridising ;
it is impossible to breed a purple-flowered pea with a
green axil, or a purple-axilled one with a white
flower. The two characters are apparently indis-
solubly connected. The comment may be made, on
this, that they are not two characters, but one. And
my answer to this is that I have no objection to
expressing the fact that two characters are indis-
solubly associated by calling them one character,
except that it is better to refer to two apparently
indissociable characters which we can conceive of as
being separated, or desire to be separated, as two
characters, partly because it may turn out that they
are separable by as yet undiscovered methods, and
partly because by naming them one character we
renounce the attempt to separate them.

Between these two extremes of association there
are intermediate forms. A great deal of work has
been done in the hope of interpreting the various
ty^)es of association on Mendelian lines. These,
however, do not concern us here, and it will suffice
to describe a single instance of two characters belong-
ing to distinct pairs which are imperfectly associated
together. One of these characters is purpleness of
flower in the culinary pea ; the other is the occurrence

BREEDING

246

of the first flower at a relatively high node on the
plant, which is the chief condition determining
lateness. The allelomorph to purpleness in the
flower is white, and the allelomorph* to highness oi
node of first flower is lowness. In the second hybrid
generation, raised from a hybrid produced by crossing
a high ” purple with a “ low ” white, there is
evidently a strong tendency for purple to be associated
with ‘‘ high ” and white with “ low.” With a very
low percentage of exceptions, all the purples are
“ high ” and all the white “ low ” ; but there are
just a few “ lows ” amongst the purples, and a few
“ highs ” amongst the whites, which show that the
two characters are not associated so closely as,
for instance, purpleness of flower and purpleness
of axil.

In conclusion it may be said that the urgent
need of the present is an exhaustive enquiry, under-
taken by independent observers, into the mode of
inheritance of the characters of domesticated animals
and plants. It will be time enough to systematise
this knowledge when there is more of it.

I.e., the other character of the same pair.

CHAPTEE XVI

THE INHERITANCE OF SEX

In the previous chapter the maimer in which the
practice of breeding will probably be modified by
the dissemination of the main conclusion which flows
from Mendel’s work was indicated. In the present
chapter a theme of mere academic interest will
engage our attention, namely, the application of
Mendelian principles to the question whether there
is a fundamental and constitutional — to be absolutely
precise, a zygotic — difference between the sexes.

One of the most recent, and certainly the most
interesting, of the applications of Mendelian prin-
ciples consists in the attempt to interpret the pheno-
mena of sex by means of these principles. But
before this theory is described it is necessary to
present certain cases in which the normal course
of Mendelian inheritance is disturbed by the inter-
ference of sex.

A horned breed of sheep — the Dorset Horned —
in which the horns are well developed in both sexes,
was crossed* with a hornless race, such as the Suffolks,
both sexes in which are hornless. The result was that
all male lambs of the first hybrid generation developed

* Note on the “ Inheritance of Homs and Face Colour in Sheep.” —
Joum. Agric. Sct.^ vol. i., pt. 3, p. 364.

247

BREEDING

248

horns, whilst the female remained hornless. The result
of reciprocal crosses was identical. It looks as if the
horned character were dominant in males, but recessive
in females. But this view is shown to express the
truth imperfectly by the appearance in the second
hybrid generation of horned ewes and hornless rams.
The explanation of these phenomena, so far as
it is an explanation, is found to be that a ram is
horned both when he is heterozygous (DR) and
homozygous (DD) for the horned character, whilst
a ewe is horned only when she is homozygous for it.
That is why all the ewes in the first hybrid generation
were hornless. In other words, a ram will develop
horns with only one dose of the horned character,
which is present in the heterozygous condition of
that character ; but a ewe needs the two doses which
are only present in the homozygous condition of
that character. There appears to be something in
the constitution of the female which inhibits the
development, in this case, of the horned character
when this is only present in the diluted heterozygous
condition. This phenomenon is known as sex-limited
inheritance, because the course of descent of a
character is afiected by the sex of the individuals
which bear it.

Another instance of it is afforded by the pheno-
mena of colour-blindness in man. This peculiarity
occurs both in men and women, but is very rare
amongst the latter. Colour-blind men may transmit
the peculiarity to their sons, but very rarely do so
to their daughters. These imaffected daughters may,

INHERITANCE OF SEX

249

however, transmit it to their sons. In a word, men
cannot transmit the peculiarity without having it,
whilst women can. The mode of transmission and
the interference of sex would appear to be exactly
the same as in the case of the horned character in
sheep. A woman is only colour-blind when she is
homozygous for the character ; a man is colour-blind
both when he is homozygous and when heterozygous
for it. The normal women who transmit it are
heterozygous for it. If this theory is true all the
sons of colour-blind women should be affected, because
even if she is mated to a normal male all her ofispring
will be heterozygous for the peculiarity, and that
is sufficient to make her male ofepring develop it.
Records in the possession of Mr. Bateson show that
the seventeen sons (who lived to be tested) of seven
colour-blind women were all colour-blind. The bear-
ing of these facts of sex-limited inheritance on the
Mendelian theory of sex will shortly appear.

Another familiar instance of the results of sex-
limited inheritance is afforded by the difference
between the male and female offspring resulting from
a cross between a black and an orange cat. The
general rule would appear to be that the male offspring
are orange, whilst the female are tortoiseshell. This
is why male tortoiseshells are practically unknown,
and why the females occur sporadically, inasmuch
as it is not possible to raise a race of tortoise-
shells because there are no toms to mate the cats
with.

The idea that sex was a Mendelian character

250

BREEDING

was first suggested by Professor Castle.* According
to Professor Castle, sex was represented by a factor in
the germ-cells exactly as a Mendelian character is.
(The factor for maleness will be written (?, and that for
femaleness $.) He supposed that both male and female
were heterozygous for sex, i.e. both were of the zygotic
formula (J $. He accounted for the equality in the num-
bers of the sexes by supposing that, in about half the
zygotes, maleness was dominant, and in the other half
femaleness was dominant. He also supposed that,
when these heterozygotes mated, the unions between
like gametes were infertile, because it was his theory
that each sex had the formula cJ ?, and it was there-
fore necessary to explain the non-existence of the
zygotes cJcJ and $$ which, according to the ordinary
Mendelian scheme, would result from the union of
<J 9 and cJ $, thus : —

Gametes of male

1 <?
«4-l

B Q

a +
O

<?<?

<??

?<?

It will be seen that this theory involves two serious
assumptions ; first, the alternativeness of dominance

♦ Castle, W. E. “ The Heredity of Sex.” Bull, Mus. Comp. Zool,
Harvard, 903. Vol. si., No. 4.

INHERITANCE OF SEX

251

m the two characters in the heterozygote, to account
for the approximate equality in the numbers of the
sexes ; and, secondly, the sterility of unions between
gametes bearing like characters.

Castle’s form of the theory has now been super-
seded by a much simpler one which is due to Mr.
Geoffrey Smith. It was briefly referred to in the
remarks on p. 181, relating to the possible kinds of
union between any two of the zygotic forms, DD,
DR and DD, where it was seen that there are only
two kinds of imion in which the parental types are
reproduced in equal numbers, namely, DR x RR, and
DR X DD. Of these two the former is the only one
in which the two parental types differ from one
another in all cases ; that is to say, RR is always
difierent from DR in outward appearance, whilst
DD is indistinguishable from DR except in those
cases, like the Andalusian fowl, in which the DR
zygote has a character peculiar to itself. Inasmuch,
therefore, as the male differs by external characteristics
from the female, and inasmuch as the result of their
union is usually the production of males and females
in approximately equal numbers, it is natural that
the suggestion that one sex was of constitution DR
and the other of constitution RR should be made.

The theory is this ; Femaleness and maleness
constitute the two characters of a Mendelian pair.
Femaleness is dominant ; that is to say, femaleness
is due to the presence of a factor the mere absence
of which constitutes maleness. But femaleness never
occurs in the homozygous condition

252

BREEDING

always in the heterozygous ? ^ (DR), whilst maleness
occurs in the homozygous (recessive) condition c?<J
(RR).

Only one instance of the kind of evidence, based
on Mendelian investigations, by which this theory is
supported will be given here.

The common currant moth {Abraxas grossulariata)
has a variety lacticolor, which is characterised by
the fact that the black spots on its wings, though
they resemble those of the parent species in distri-
bution, are very much smaller. This variety is only
known as a female.

A series of crossings between the parent species
and its variety were made by Mr. Doncaster and
Mr. Raynor, and gave the following remarkable
results : —

1. Lacticolor $ X grossulariata (J gave in FI* <Js and $s

all grossulariata.

2. FI grossulariata $ X FI grossulariata $ gave grossulariata (Js

and $s, and lacticolor ; no (Js of lacticolor appeared.

3. Lacticolor $ X FI grossulariata ^ gave all four kinds,

grossulariata (Js and $s and lacticolor (Js and $s.

The lacticolor males were the first males of the variety

that had ever been seen.

4. Perhaps the most curious result of all. FI grossulariata

5 X lacticolor (J gave equal numbers of lacticolor (Js

and grossulariata $s.

Mr. Doncaster attempted to account for this
remarkable series of results by adopting Professor
Castle’s idea that both sexes were heterozygous ; and
he elaborated a consistent explanation of his results.

* A symbol, introduced by Mr. Bateson, to denote the first hybrid
generation.

INHERITANCE OF SEX

253

His theory involves five distinct assumptions, which
will not be gone into here because a much simpler
explanation, suggested by Mr. Bateson and now
accepted by Mr. Doncaster, has obtained general
credence. Mr. Bateson’s theory involves two assump-
tions, which are as follows, in his own words : — ^

1. “ That the female is heterozygous in sex,
femaleness being dominant, and the male a homo-
zygous recessive (see pp. 251-2).

2. “ That when in Fl the two dominant characters
femaleness and the grossulariata factor co-exist, there
is spurious allelomorphism or repulsion between
them, such that each gamete takes one or other of
these factors, not both.”*

A word in explanation of spurious allelomorphism.
True allelomorphism expresses the relation between
the dominant and recessive members of the same
pair — ^yellow with green, or round with wrinkled.
Spurious allelomorphism is a relation between factors
belonging to distinct pairs which gives the appearance
of their belonging to the same pair. If a relation
of this kind, similar to that which exists between the
two dominant characters, femaleness and the gros-
suLariata character, were to exist between the two
dominant characters, yellow and round, in the hetero-
zygote produced by crossing a yellow wrinkled with
a green round, yellow would repel round, so that they
could not both exist in the same gamete, just as in
true allelomorphism yellow and green cannot exist
in the same gamete ; yellow could only exist in the

* “ Mendel’s Principles of Heredity,” p. 175.

254

BREEDING

same gamete with wrinkled, and a relation of spurious
allelomorphism would thus be established between
these two characters.

The explanations of the results of the four kinds
of mating made by Doncaster would then be repre-
sented by the following schemes. The dominant
and recessive member of the pair of characters
relating to colour, namely grossidariata and lacticoloTy
are written D and R respectively ; and the dominant
and recessive members of the pair of characters
relating to sex are written ? and c? respectively.
The gametes are written outside the square, the zygotes
inside the squares, and underneath the zygotic form-
ulae the actual character and sex of the zygote : —

1. Lacticolor $ x grossvlariata $ produces 50 per
cent, cjs and 50 per cent. ?s, all grossidariata. Thus : —

Lact. $

«0

CO

I

R?

DR?<J

DR(?c?

gross. $

gross. (J

DR?(?

gross. 9

gross.

2. Fl grossvlariata ? x Fl grossvlariata <J pro-
duces 50 per cent, gross. 26 per cent, gross. and
25 per cent. lact. ?: —

INHERITANCE OF SEX

255

I

Rc?

FI Gross. $

Dd

R$

DDdd

DR ?d

gross, d

gross; $

DBdd

RR?d

gross. <J

lact. $

3. Lacticolor $ x Fl grossulariata $ gives 25 per
cent, gross. ?, 25 per cent, gross. (J, 25 per cent.
loot. $, and 25 per cent. lact. ^ : —

Lact. $

'i

o

Rc?

R ?

R d

DR?d

DRdd

gross. $

gross. $

RR$d

RRdd

lact. $

lact. cJ

4. Fl grossulariata $ x lacticolor <? produces
50 per cent, gross, d and 50 per cent. lact. $ : —

256

BREEDING

FI Gross. §

1

R(?

R?

RR?<?

gross. ^

loot. $

DR<?<?

RR?<?

gross. (J

loci. $

Doncaster has since found that the result of crossing
pure wild grossulariata $ with lacticolor is precisely
the same as that of mating No. 4 on p. 252 ; that is,
as the result of crossing Fl grossulariata with (J
lacticolor. This proves that the wild grossulariata
female is heterozygous in regard to this character,
half her germ- cells bearing the grossulariata factor,
the other half the lacticolor factor.

Here, then, is an instance in which the simplest
interpretation of the phenomena is the supposition
that the female is heterozygous for sex, femaleness
being the dominant character, whilst the male is
pure recessive. The value of this case as evidence
for the theory that the female is a hybrid in respect
of sex and the male is pure recessive does not merely
lie in the fact that this is the simplest way of account-
ing for the facts ; for this is also the simplest way of
accounting for the fact that males and females are
produced in equal numbers in the human species,

INHERITANCE OF SEX

257

and this circumstance would be slender evidence for
the truth of this theory. The peculiar interest and
value as evidence of cases like that of the currant
moth lies in the fact that in them the character of
sex has, as it were, become entangled with and
entered into permanent relations with a varietal
colour- character, the consequences of which insigni-
ficant little event has been to withdraw the veil
which has all this time hidden the real difference
between the sexes from our view.

Bearing in mind the presence and absence hypo-
thesis of the nature of the two characters of a
Mendelian pair, it will be observed that, acco^ing
to the theory we are considering, the female who
is heterozygous for sex is characterised by the 'presence
of the factor in virtue of which she is female, whilst
the male lacks this factor altogether. Even the
female, being a hybrid, has only one dose of this
factor, whatever it is. The male is male in virtue
of the fact that he has none at all. Half of the
germ cells of the female contain this factor ; the
other half lack it. All the germ- cells of the male
lack it.

Now this factor, which is present in the female
alone, has been identified with the thing which checks
the development of certain characters which attain
to their full development in the male, in cases of sex
limited inheritance. In the case of the horns in
sheep there is something in the female hybrid resulting
from a cross between a horned and hornless individual
which inhibits the development of horns, which attain

R

BREEDING

258

their normal size in the male hybrids from the
same cross. And this inhibiting something is, as I
have said, identified with the dominant factor which
is present in the zygotic constitution of the female
and absent from that of the male. Those of my
readers who are familiar with Otto Weininger’s
theory, expounded in his “ Sex and Character,” of
the difference between the two sexes will detect a
very close parallel between it and the Mendelian
view.

Evidence derived from widely different and abso-
lutely independent lines of inquiry tends to support
the Mendelian theory. But there is a qualification.
The various lines of evidence point to the con-
clusion that one sex is hybrid and the other pure ;
but they do not all point to the same sex as being
hybrid and pure respectively.

Let us examine them separately. We will deal
first with the only one, of those I propose to consider,
which, like the Mendelian theory, points to the
female as the hybrid. This is contained in a book
entitled “The Causation of Sex,” by Dr. R. Dawson,
which appeared recently. The theory, however, was
first enunciated at the Obstetrical Society in 1900,
so that there is no possibility of the author having
heard of the Mendelian theory of sex, even if the fact
that he does not refer to it were not sufficient evidence
of this. Dr. Dawson thinks that the male plays no
part in determining sex in the human species, that his
germ-cells are all indifferent in respect of sex ; the
female, on the other hand, gives rise to two kinds of

INHERITANCE OF SEX 259

ova, one kind which, when fertilised, will give rise to
males, and another which, when fertilised, will give
rise to females. The former, the male-producing,
arise in the right ovary, and the latter, the female-
producing, in the left ovary. So that about equal
numbers of the two kinds are given off. When a
woman is once mature, an ovum is produced alter-
nately from each ovary, one at a time, at each
ovulatory period; a male-producing one from the
right ovary ; next month a female-producing one from
the left ; next, a male from the right ; then a female
from the left again, and so on. So that when a woman
has had one child under such circumstances that the
date of the ovulatory period at which the egg, from
which the child developed, was extruded from the
ovary can be calculated, the sexes of any future
children required may be arranged. The evidence
adduced for the allocation of the male- and female-
producing ova to the right and left ovaries respectively
cannot yet be considered to have proved the point ;
nor is the point necessary for our purpose. What
is of intense interest is the fact that clinical obser-
vations conducted in complete ignorance of Mendelian
speculation should have led to the enunciation of a
theory of sex which is practically identical with the
Mendelian one, The two theories, at any rate, have
these points in common. According to both, one sex,
the female, produces two kinds of germs, male-produc-
ing and female-producing, in equal numbers, whilst the
male only produces germ- cells of one kind. The sex,
therefore, is determined by the female. Dr. Dawson’s

26o

BREEDING

theory and the Mendelian one are, of course, expressed
in different terms, but in their essential features they
are identical. And, I think, it is very difficult to
regard this circumstance as a mere coincidence. It
is not likely that either of them is closely in touch
with actuality, but that up to a certain point they
agree is an indication that a clue has been found.

We have now to deal with pieces of evidence, very
different from one another and from those already
considered, which strongly support the view that
one sex is a hybrid, or, to use a word of wider
signification, bipotential, and the other pure, or
unipotential, but point to the female as pure, or
unipotential and the male as hybrid or bipotential.

The first of these, indications is derived from
Mr. Geoffrey Smith’s observations on crabs in the
Bay of Naples. The female of this species of crab is
characterised by the breadth of her abdomen and
the presence, on it, of long filamentous appendages
to which the eggs are attached when they are laid,
and by the delicacy of her pincer-claws. The male
is characterised by the narrowness of his abdomen,
the possession of a pair of hard copulatory appen-
dages instead of a number of flexible filamentous
ones, and by the much greater size and strength of
his pincer-claws.

This crab is occasionally infected by a parasite,
called Saccvlina, which is also a crustacean, though
it would never be recognised as such, for it hangs,
like a full wallet, from the body of its host, whilst
inside it consists of a ramifying system of fibrils which

INHERITANCE OF SEX 261

extend root-like throughout the body. The parasite
may kill the crab ; if it fails to do this, and the crab
recovers, the effects which it produces are sufficiently
profound. The internal root-like part attacks the
genital organs and destroys them. If the crab
recovers the genital organs are regenerated. What
is remarkable about the case is that the effect differs
according to the sex of the crab affected. If it is
a female that was affected very little effect beyond
such a slight diminution in the size of her filamentous
limbs as could be attributed to the shock of the
infection and strain of recovery is observed when
she has recovered.

But if it was a male that was affected, the changes
brought about are very much more profound, and of a
very remarkable nature. They consist not only in
the assumption by the male of the external features,
already described, which distinguish the female, but
in the actual development by him, in the place of
the destroyed testis, of a genital organ which contains
eggs. The pincers of the male lose their robustness
and become reduced to the size characteristic of the
female. So do the copulatory appendages. This
change might, however, be ascribed to shock and
general debility consequent on it ; but the other two
changes which occur cannot be so interpreted. Not
only does the abdomen lose the pointed shape charac-
teristic of the male, but it actually develops the long
filamentous swimmerets characteristic of the female.

In a word, this parasite is capable of making a
male crab undergo a series of changes which, in bad

262

BREEDING

cases, may end in its becoming a female ; but it is
incapable of making the female undergo even the
smallest change in the direction of becoming a male.

The general conclusion derived from these facts
is that the male contains within his organisation the
potentiality of both sexes, only one of which — ^the
male — is manifested under ordinary circumstances;
though under the exceptional ones just described
it may be suppressed and the female given the oppor-
tunity of appearing ; whilst the female contains only
the potentiality of that sex to which she ostensibly
belongs. This conclusion, stated in Mendelian terms,
is that the male is a heterozygote, and the female
a homozygote, which agrees with the Mendelian
theory of the constitution of the sexes in so far as
it asserts that one of them is heterozygous and the
other homozygous, and differs from it in that it
makes the male heterozygous and the female homo-
zygous.

In the case of the next piece of evidence bearing
on the constitution of the sexes we come to very
close quarters with the processes which determine
sex, and the course of heredity in general — to the
germ-cells themselves. The weakness in Mendel’s
theory to account for his simplest results is that the
existence of factors in the gametes is no more than
an assumption. The assertion that they exist rests
on an a priori basis. They have never been seen.
In the case to be considered now, the statement that
one sex is heterozygous for sex, and that half of its
germ-cells carry a thing which determines a particular

INHERITANCE OF SEX 263

sex and the other half do not, is based on the existence
in half of the germ- cells of things which can actually
be seen.

It has been discovered that in certain insects
half the spermatozoa contain a supernumerary, or,
as it is called, accessory, chromosome. This chromo-
some is not merely an additional one, but it behaves
differently from the other chromosomes in the division
of the cells. In the insect Anasa, for instance, a
form allied to the Bug, the number of chromosomes
in one half of the spermatozoa is nine ; in the other half
ten ; the number in all the eggs is ten. The result
of the union between a spermatozoon with nine
chromosomes and an egg (with ten) is a male ; and
of the union between a spermatozoon with ten and
an egg (with ten) is a female. This is known to be the
case because the number of chromosomes in the
somatic cells of males is nineteen ; whilst that in
the somatic cells of females is twenty.

If a case such as this is described in Mendelian
terms, the male, half of whose germ- cells contain an
element which the other half lack, must be regarded
as heterozygous (DE), whilst the female, all of whose
germ-cells are alike in respect of this element, must
be regarded as homozygous, but inasmuch as all her
germ -cells are alike in the 'possession of an element
she must be regarded as a dominant homozygote,
namely DD, not EE.

It will be noted that the general conclusion as
to the nature of the difference between the sexes
based on the behaviour of the accessory chromosome is

BREEDING

264

almost identical with that based on the changes
effected by the parasites in the crabs. In both cases
the male is bipotential and the female unipotential.
This is all the more remarkable when the profoundly
different nature of the phenomena from which the
conclusions are derived, in the two cases, is borne in
mind.

It would be unprofitable in the present stage of
inquiry, when every inference ought to be regarded as
sub judice, to attempt to base a verdict on such
conflicting and insufficient evidence as that which
is at present obtainable. For the present it will
be sufficient to sum up the evidence which has been
considered.

All four sets of evidence examined point to the
conclusion that one sex is bipotential and the other ‘
unipotential. The first two, the clinical and the
Mendelian, point to the female as the heterozygote.
The last two point to the male as the heterozygote.
In estimating the relative value of the evidence in
these two sets of evidence it should be borne in mind
that in the case of the Mendelian and the clinica],
the conclusion is only reached by a long and devious
argumentative route ; it is as if the secret of the
constitution of the sexes were at the end of a winding
passage, and the glimpse we get of it by the Mendehan
and clinical methods is such a one as we might get
by the adjustment of mirrors in the passage so as to
see the far end of it. The conclusion is liable to be
falsified in proportion to the number of distinct steps
(where error may creep in) in the argument by which

INHERITANCE OF SEX 265

it is reached ; just as the image of the end of the
passage will be distorted in proportion to the number
of mirrors it has to be reflected in on its way to the
eye. In the case of the efiiects produced by the
parasites in the crabs and in that of the accessory
chromosome, we are, as it were, much farther along
the passage, at much closer quarters with the actual
constitution of the sexes, and can look directly in.
Compare the crabs with the currant moth ; in
the former actual sexual changes can be observed ;
in the latter we could learn nothing about the nature
of sex at all if the sex-character had not become
entangled in the gamete with a varietal character
of pigmentation. Compare the long chain of argu-
ment, some of the links of which are not very strong,
which leads to the conclusion that a woman produces
equal numbers of female- and male-producing ova
with the demonstrable fact that in certain insects
half the spermatozoa contain an element which the
other half lack.

A comparison between the conclusion based on
the Mendelian and cytological evidence in connection
with that derived from the crabs is profitable because
it may help to divest the problem of certain pre-
conceptions which, in my opinion, tend to obscure
its real nature.

The diflS.culty created by the fact that the Men-
delian evidence points to the male as the homozygote
and the cytological to the female may be cleared up
in one of two ways : It may be proved that the
Mendelian indication or (less likely) the c}dological,

266

BREEDING

one is wrong ; or it may be that one sex is hetero-
zygous in one species and another in another. But
there is another point in which the conclusions derived
from these two sources differ. According to the
Mendelian conclusion the homozygous sex, which
happens to be the male, is a recessive homozygote
(RR), whilst according to the cytological the homo-
zygous sex, which happens to be the female, all of
whose germ-cells possess the “ extra ” chromosome
possessed by only half of the spermatozoon, is a
dominant homozygote (DD). Now, if the Mendelian
theory, and especially the presence and absence
hypothesis, means anything at all, both these con-
clusions cannot be true. DD contains two “ doses ”
of the sex factor ; RR contains none. They differ
more from one another than a male from a female,
as interpreted by Mendelian theory. DR differs
from RR only by one “ dose.” It may be that the
homozygous sex is DD in some cases, and RR in
others. But this impasse may be due to the initial
error of founding a theory of sex, as the Mendelian
theory of it has been founded, on a basis of material
particles. And it may be that the antithesis is really
not between heterozygous and homozygous, i.e. be-
tween the presence and absence of particles, but, as
the case of the crabs seems to indicate, between an
unstable condition and a stable one.

CHAPTER XVII

THE ORIGIN OF THE MENDELIAN MODE OF INHERITANCE

It is a very important matter to determine whether,
as some hold, Mendelian heredity is a fundamental
process the true understanding of which will render
all other phases of heredity intelligible, or whether,
as others believe, it is an anomalous phase of heredity
which is restricted to a small group of characters
belonging chiefly to domesticated animals and plants.
For if the latter is the case, the investigation of
Mendelian phenomena can throw no light on the
problem of the relation between successive generations
of organisms, nor help to elucidate the nature of the
processes to which evolution is due. The chief
exponent of the view that Mendelian inheritance is
merely an anomalous phase of heredity is Dr. Archdall
Reid, whose views on this subject are expressed in
his recent book, ‘‘ The Laws of Heredity,” to which the
reader is referred for a full exposition of this view.
Perhaps the simplest way of expressing this view is
to say that, according to it, we have begun to read
the series of phenomena and of hypotheses, which
extends from the first to the sixteenth chapter in
this book, at the wrong end. We have taken the
simple Mendelian phenomenon as the starting-point,
and have finished up by interpreting the inheritance
and nature of sex in the light of it ; whereas, according

267

268

BREEDING

to Dr. Reid and his adherents, the mode of inheritance
by which sexual characters are transmitted is the
real starting-point, and Mendelian phenomena are
anomalous occurrences caused by certain varietal
characters taking to be inherited by the sexual mode
of transmission, and occupying a mere cul-de-sac in
the stream of evolution. “ It has been suggested,”
says Dr. Reid, “ that the inheritance of sexual
characters is Mendelian ; we shall be nearer the

truth if we say that the inheritance of Mendelian
characters is sexual.” By this he simply means

that the sexual and not the Mendelian mode of
inheritance is the phenomenon in the light of which
the other must be interpreted. According to his view
both sexes are bipotential, each sex contains the
opposite one, in fotentia ; this he calls alternative
latency and potency ; there is no such thing as

alternative inheritance, which is the distribution of

characters to the germ- cells in such a way that one
germ- cell may receive a character in its entirety and
another not at all. The illusion of alternative in-
heritance is created by the occurrence in about
half the individuals, of a species, of a particular sex
in a patent and the other sex in a latent state ; and, in
the other half, of the same sex in a latent, but the
other sex in a patent state. This view resembles
that of Castle in that, according to it, both sexes
have the same constitution — that is, they contain
the potentiality of both sexes — and only differ from
one another in the matter of which sex happens
to be uppermost, or patent.

MENDELIAN HEREDITY 269

Now, the indications that one sex, to state it in
the broadest terms, is bipotential and the other
unipotential is based on minute evidence of a strictly
a 'posteriori nature, whilst the conclusion that both
sexes are bipotential is based on evidence which its
author claims to be “ massive,” and reached by the
a priori ladder, which he believes to be trustworthy.
For my own part, the minute weighs more heavily
with me than the massive. At any rate, I propose
to regard it as more probable that one sex is
bipotential and the other unipotential than that both
are bipotential. And in any case, this is a secondary
point at issue. The essential one is : Which came
first, the Mendelian or sexual mode of inheritance ?
And by the sexual mode I mean, now, the theory
that one sex is bipotential and the other unipotential,
and not Dr. Reid’s alternative latency and potency.
The Mendelian may here interpose that I am holding
his brief. I am, in the sense that I am prepared to
accept an interpretation of sex on the basis of the
Mendelian one. But I also incline to the view that the
sexual preceded the Mendelian mode in so far as it
seems to me likely that the mode of inheritance from
which both are difierentiations was, in the first place,
only used (if the expression may be allowed) by
sexual characters, and was of the type DR x RR,
or, more generally, heterozygote x homozygote,
or xy X XX ; and that all the phenomena which we
regard as characteristically Mendelian — the existence
of pure dominants and recessives, the uniformity of
first hybrid generations, the segregation in the ratio

270

BREEDING

1:2:1 in the second hybrid generation, and so on —
are, as it were, a secondary display set off by certain
varietal characters becoming inherited by the sexual
mode described in this paragraph. The heterozygote,
together with the ordinary (or homozygous) type
(xy + xx) first came into existence in connection
with the inheritance of sexual characters. And it
will readily be admitted that an extremely simple
way of bringing about the production of the two
sexes in equal numbers would be the institution of
one sex (A), half of whose germ- cells contained the
factors for one sex, whilst the other half contained
those for the other ; and of another sex (B) all of
whose germ-cells were indifferentiated as regards sex.

The original mode of inheritance, of which the
Mendelian is a later development, and which has pos-
sibly persisted without much modification as the sexual
mode, may, therefore, be expressed in the formula —

DR X HR = 60 % DR -f 50 % RR
Or, more generally —

xy X XX = hO % xy bO % XX

The familiar

DD X RR

DD

1 DD 2 DR 1 RR

I L_

1 i!)D 2 dr 1

RR

RR

is a secondary development, and only appeared when

MENDELIAN HEREDITY 271

nbn-sexual characters came to be inherited by the
above-outlined sexual mode of inheritance. So long
as the only character which existed in the DR con-
dition was one of the sexes, the DR condition could
never unite with anything but the RR condition,
which was that in which the other sex existed. If
that of the $ was the bipotential one, only one kind
of mating could take place, DR ( $) x RR ( But
directly a non-sexual character acquired the bipoten-
tial condition the fat was in the fire. It became
possible for the DR condition to mate with the DR
condition, and this led to the origin of the pure
dominant and recessive types, DD and RR, which,
as being perfectly constant, would be greedily
seized on by the breeder so soon as such types came
under his intelligent observation. And the existence
of these two pure types, DD and RR, suggested them-
selves as suitable material for hybridisation experi-
ments, probably hundreds of years after they were
isolated, to Mendel, who tabulated the results of his
crosses and discovered the phenomenon (illustrated
on the frontispiece of this book) which we are apt
to regard as so fundamental. Mendelian segregation
in the ratio 1:2:1 only became possible when it
became possible for a DR to mate with a DR ; and
this only became possible when the DR condition
ceased to be the exclusive attribute of, say, femaleness,
and became an attribute of, say, yellowness. Men-
delian segregation in the ratio 1:2:1 could never
take place before this event, because a DR could not
mate with a DR, both being females ; but once it

272

BREEDING

had happened, the way for Mendelian segregation
was made clear by the possibility of a union between
two DRs. A female cannot mate with a female ;
but a yellow can mate with a yellow.

The only alternative to this view of the origin of
the Mendelian phenomena is the supposition that
the first types to originate were the pure dominant
and recessive homozygous forms. Now, apart from
the fact that this is rendered improbable by the
probability of the common origin of the Mendehan
and sexual modes of inheritance from a mode which
most closely resembles, even if it differs from, the
sexual one, it can be shown that at least in one case
this cannot have been the way in which these pheno-
mena arose. And it is well to remember that the
question we have to ask ourselves when we are
considering the value of evidence for a particular
theory is not : how many facts are consistent with
it ? but : is there a single one which is inconsistent with
it ? The statements of forty of the friends of the
Irishman who was accused of shooting his landlord, to
the effect that they had not seen him doing it, were
consistent with his statement that he had not done it.
But he was hanged, because the fact that one man had
seen him do it was not consistent with his statement.

One of the varieties of the Snapdragon, which
ordinarily has green leaves, is a form with much
yellower, variegated, leaves; this variety is spoken
of as aurea (golden). The inheritance of these
characters has been investigated by Dr. Baur. He
found that the greens bred true but that the varie-

MENDELIAN HEREDITY 273

gateds, when self-fertilised, produced altogether 573
variegated and 286 green. Now this is a ratio of two
to one, instead of the three to one which would be
expected in an ordinary case of segregation. It looks
like a 1:2:1 ratio shorn of one of its ones. This
is what, in fact, it actually is. Dr. Baur examined the
seedlings raised from the seeds of a variegated plant
very shortly after the germination of the seeds, and
found 77 green, 160 variegated, and 51 seedlings
which were almost white and entirely destitute of
green. These latter possess no chlorophyll, and
consequently die. They evidently represent the
homozygous condition of the character, the hetero-
zygous condition of which is the variegated form.
This case, therefore, may be represented thus, assum-
ing that the green is recessive : —

DR X DR

(variegated) | (variegated)

[1 DD] 2 DR 1 RR

{chtorophyU-less forms (variegated) (green)

whi<)h 'perish)

This particular case of the Mendelian pheno-
menon can, therefore, never have originated in the
union of pure dominant and recessive homozygous
forms, DD and RR, thus : —

DD X RR

I

DR

1 DD

k

2 DR

1 RR

274

BREEDING

because the dominant form does not exist— or,
rather, gets no nearer maturity than the seedling
stage.

The general trend of the evidence seems to be
that the Mendelian mode of inheritance originated
from — or, at any rate, in close connection with — that
by which sexual characters are inherited. At any
rate, the similarity between the two modes is very
close.

Glossary

Allelomorph, a member of a Mendelian pair of characters. The
adjectival form allelomorphic is useful, because by its em-
ployment the statement that the character “ round ” bears
that relation to the character “ wrinkled ’’ which the two
characters of a Mendelian pair bear to one another, can
be condensed into the statement that round is allelomorphic
to wrinkled.

Bisexual crosses, according to Prof, de Vries, are those between
two individuals each of which bears a particular character,
one of them possessing it in a latent and the other in a
patent state. Compare with “ unisexual.”

Chromosome, a term applied to certain minute bodies, in the
kernel or nucleus of the animal and vegetable cell, which
appear at definite periods in the division of the cell, are
constant in number for each species of animal or plant,
and are characterised by the fact that they stain very deeply
with certain dyes. The number of chromosomes in a germ
cell is half the number of them in a body cell of a given
plant or animal.

Cotyledon, a term applied to the first leaf or leaves of the plant.
One group of plants, of which the wheat is an example, is
characterised by the possession of only one ; they are called
monocotyledons. Another group, of which the pea is an
example, has two ; these plants are called dicotyledons.

DD (See Homozygote).

DR (See Heterozygote).

Dominant characters are those borne by the first generation

275

276

GLOSSARY

from a cross between two indi\nduals possessing respectively
the two characters of a Mendelian pair.

Duplex eyes are those which have a layer of brown pigment in
front of the iris (See Iris).

“ Extracted ” is the term applied to the individuals, in the
second hybrid generation from a cross, which resemble the
parent forms that were mated to make the cross. An
“ extracted ” green is a green which appears in the second
or a subsequent hybrid generation from a cross between a
yellow and a green pea.

Fi, a symbol, introduced by Mr. Bateson, as an abbreviation
for the first hybrid generation.

F2, a symbol for the second hybrid generation.

F3, a S5mbol for the third hybrid generation. And so on.

Factor^ a name for the thing in a germ cell which makes that
germ cell develop a particular character, such as tallness
as opposed to dwarfness.

Fasciated plants are those in which the nodes do not, as in the
normal plant, succeed one another regularly up the stem,
but occur in profusion at the top.

Gamete, a name for the reproductive cell, whether male or
female, in both animals and plants.

Germ cell, the equivalent of Gamete.

Heterozygote, a zygote (q.v.) resulting from the union of two
gametes bearing dissimilar factors — one a dominant, the
other a recessive one. DR is the symbol for heter-
ozygote.

Homozygote, a zygote (q.v.) resulting from the union of two
gametes bearing similar factors, which may be either both
dominant, producing a dominant homozygote or DD, or
both recessive, producing a recessive homozygote or RR.

Internode, the section of the stem of a plant between two nodes.

GLOSSARY

277

Iris, the coloured part of the human eye surrounding the
pupil.

Mutation, the origin of species by discontinuous variation (See
Variation, discontinuous).

Prepotency, the property said to be possessed by certain in-
dividuals, especially amongst stallions, of transmitting
their qualities to their offspring, whatever female they are
mated with.

Pure, a term applied to those individuals which when mated
together produce, exclusively, individuals indistinguishable
from themselves.

Recessive characters are those which, in a cross between in-
dividuals the two characters of each of which bear one of
the same Mendelian pair, entirely disappear in the first
hybrid generation.

Reversion, the production, on crossing, of a supposed remote
ancestor of the two forms crossed.

Segregation, the reappearance in definite ratios, in the second
hybrid generation, of the characters of two forms crossed ;
and of the first hybrid generation (where this differs from
the dominant character).

Soma, the body, as opposed to the germ cells.

Somatic, of, or pertaining to, the body as opposed to the germ
cells.

Unisexual crosses, according to Prof, de Vries, are those between
two individuals, one of which bears a particular character
entirely lacking in the other (See Bisexual).

Variation, the production by animals and plants of those differ-
ences the accumulation of which is supposed to have resulted
in the various forms of living things. Continuous variation
is the name given to the occurrence of such differences as
always exist in any sample of a given species. These varia-
tions are not supposed to be transmitted to the next genera-
tion. Discontinuous variation is the term applied to the

278

GLOSSARY

differences which are said to be exhibited after long intervals
of time by all species of living things. Instances of such
differences occurring in the Evening Primrose {CEnothera
Lamarckiana) are given in Chapter I. These differences are
transmitted to the next generation when two individuals
bearing them are mated.

Zygote, the result of the union of two gametes (See Gamete).

Index

AbraxcLS (the currant moth), in-
heritance of sex in, 252
Absorptive capacity of seeds of
peas, 128

Accessory chromosome as sex deter-
minant, 263

Albinism, inheritance of, in mice,
73

Allelomorph, 246
Allelomorphism, spurious, 253
Anasa, accessory chromosome in,
263

Andalusian fowl, inheritance of
colour in, 33

Angora character, inheritance of,
in cats, 191
Antirrhinum, 272

Bateson, W.; 4, 190, 249, 253
Baur, E., his work on Antirrhinum,
272

“ Bisexual ” crosses of de Vries, 139
Bunyard, E. A., his work on
cotyledon colour in pea?, 131
Butler, Samuel, 188

Oastle, Professor, his theory of sex,
250

Oats, tortoiseshell colour in, 249
Cattle, roan colour in, 80
Chdidonium, 137

Oobbett, William, his book on
gardening, 143

Oolour- blindness, inheritance of, 248

Oombs of fowls, 104

Counters, device for illustrating
Mendel’s theory with, 172

Grabs, effect of parasite SaccvUna
on sex of, 260

Grossing, practical instructions for,
in peas, 146

Currant moth, inheritance of sox in,-
252

Darwin, Charles —
his theory of evolution, 3
his theory of pangenesis, 187

Dawson, R.; his theory of sex,
258

DD, DR, RR, possible matings
between these zygotic types,
179

De Vries —

his work on mutation, 5
his theory of the nature of
Mendelian characters, 10, 139

Directions for carrying out experi-
ment with peas. Oh. X.

Dominance —

not an essential feature of Men-
delian inheritance, 38
distinguished from prepotency,
96

Dominant characters —
the number of generations it
takes to combine two, 104

279

28o

INDEX

Dominant Oharaoters —

not necessarily the older char-
acter in the pair to which they
belong, 125

Doncaster, his work on sex in
currant moth, 252
Dumpling, how does the apple get
into the ? 188

Duplex character in human eye, 41

Enemies of peas, 143
“ Extracted ” recessive, or domi-
nant, “ purity ” of, 238
Eye-colour —
in man, 40
in mice, 73

“Faetor,” 162

FI, symbol for first hybrid genera-
tion, 252
Fowls —

Andalusian, 33
combs of, 104

rumpless condition of certain, 125

Gamete, 176
Gametic purity, 217
Goss, John, 193

Hdix hortemis, absence of colour
dominant in, 138
Heterozygote, defined, 178
Homozygote, defined, 178
Horns —

in cattle and sheep, 137
in sheep, 248

Horticultural Transactions, fac-
simile from, 199-202
Hurst, 0. 0., his work on human eye-
colour, 40

Hybrid, how to distinguish between
it and a dominant, in practice,

46

Immunity to “rust” in wheat, 138

Keeble, F. W., his work on stature
in peas, 133

“ Laws of Heredity, The,” Dr.
Archdall Reid’s book entitled,
267

Lock, R. H., 70, 159

Maple pea, 110

Mendel —

directions for repeating his ex-
periments, Gh. X.
varieties of peas bearing the
characters investigated by him,-
158

Mendelian phenomenon —
the, 14-16 and Frontispiece
origin of the. Oh. XVII.

Mice, inheritance of colour in^
73

Mutation, 4

Nageli, Oarl, his correspondence
with Mendel, 190

Nettle, inheritance of character of
leaf of, 137

(Enothera —
brevislylisy 6
gigo8, 6 (Figs. 2 and 3)
gigas-nanella, 10
loRvijolia, 6

Lamarckiana, 5 (Figs. 1 and 3)
nanella, 9 (Fig. 5)
rubrmervia, 8 (Fig. 4)

Pangenesis, Oharles Darwin’s theory
of, 187

INDEX

281

Pea, culinary —
stature of, 12 (Fig. 6)
faaciated variety of, 21 (Fig. 8)
texture of pods of, 24 (Fig. 9)
colour of pods of, 26 (Fig. 10)
colour of seed-coats of, with
simple dominance of grey, 27
(Figs. 11 and 12)
colour of seed-coats of, giving
reversionary purple spots in
first cross, 114 (Fig. 27)
colour of seed-coats of, maple and
purple spot, 110 (Fig. 26)
colour of flowers of, 118 (Plate
IV.)

colour of cotyledons of, 53
(Plate I.)

shape of cotyledons, 66 (Fig. 21)
colour and shape of cotyledons,
93 (Plate m.)

1:2:1 ratio in. Oh. II. and
Frontispiece

3:9:4 ratio in, 114 (Fig. 27)

9 : 3 : 3 : 1 ratio in, 93 (Plate III.),
110 (Fig. 26)

Presence and Absence —
hypothesis, 126

its application to cotyledon shape,

130

its application to cotyledon colour,

131

evidence relating to, 137
its application to theory of sex,
257

Purity, gametic, 217
Purple spotted pea, 110

Ratio —

1:2:1
in peas, 16
in fowls, 34

Ratio —

1:2:1 {cordinued)
in successive hybrid genera-
tions, 70-71
in mice, 76
9:3:3:1

in peas ; cotyledon characters,
97

seed-coat characters, 112, 115
3:9:4

in peas ; flower-colour, 119
similarity of 3 : 9 : 4 to 1 : 2 : 1,
121

Recessive characters, the number
of generations it takes to com-
bine them, 104
Reid, Dr. Archdall, 267
Reversion, 74

Roan colour in cattle, 80, 178
“Rust” in wheat, 138

SaccvUna, effect of, on sex of crabs,
260

Seed, the nature of a, 52
Segregation, the essential feature
of Mendelian inheritance, 38, 80
Selectionist, point at issue between
the, and mutationist, 4-5
Sex —

Dr. Archdall Reid’s theory of
the inheritance of, 26^

Dr. Dawson’s, 258
the Mendelian, 251
Sex-limited inheritance, 248
Simplex eye, 41

Smith, Mr. Geoffrey, his theory of
sex, 251
Soma, 239

Specimens illustrating Mendel’s
results, method of mounting,
159

282

INDEX

Spurious allelomorphism, 253
Starch-grains of —

round and of wrinkled peas, 128
round x wrinkled peas, 134
Sturges, the Rev., 39
Susceptibility to “ rust,” 138
Symbols for types of individual in
Mendelian heredity, 176

Tortoiseshell colour in cats, 249

Unfixable characters, their signifi-
cance, 39

“ Unisexual ” crosses of De Vries,
139

Unit characters, 216
Uvea, 43

Variation —
continuous, 4
discontinuous, 4

Varietal characters, Mendelian in-
heritance of, 10

Varieties, Prof, de Vries’ definition
of, 9

Weininger, Otto, his theory of sex,
258

Weismann, 187
Zygote, 176

Printed by Cassell & Company, Limited, La Belle Sauvage, London, E.C.
10.11.11

N