I

;•■*'■''"-;',:,

JOURNAL OF GENETICS

CAMBRIDGE UNIVERSITY PRESS

C. F. CLAY, Manager

LONDON : Fetter Lane, EC. 4

LONDON : H. K. LEWIS AND CO., Ltd., 136 Gower Street, W.C. i
LONDON : WHELDON AND WESLEY, Ltd., 28 Essex Street, Strand, W.C. 2

CHICAGO: THE UNIVERSITY OF CHICAGO PRESS

(Agent for the United States and Canada)
BOMBAY, CALCUTTA, MADRAS : MACMILLAN AND CO., Ltd.
TOKYO : THE MARUZEN-KABUSHIKI-KAISHA

All rights reserved

JOURNAL OF GENETICS

EDITED BY

W. BATESON, MA., F.R.S.

DIRECTOR OF THE JOHN INNES HORTICULTURAL INSTITUTION
AND

R. C. PUNNETT, MA., F.R.S.

ARTHUR BALFOUR PROFESSOR OF GENETICS IN THE UNIVERSITY OF CAMBRIDGE

Volume XI. 1921

Cambridge

at the University Press

1921

AT

CONTENTS

No. 1 (April, 1921)

I'AOE

B. MiYAZAWA. Studies of Inheritance in the Japanese Convolvulus

Part II. (With Plate I (coloured)) 1

A. F. Blakeslee. a Graft- Infectious Disease of Datura resembling

a Vegetative Mutation. (With Plates II— VI) . . . . 17

R. C. PuNNETT and the late Major P. G. Bailey. Genetic Studies in

Poultry. III. Hen-feathered Cocks. (With two Text-figures and

Plates VII— XI) 37

J. A. S. Watson. A Mendelian Experiment with Aberdeen- Angus

and West Highland Cattle. (With Plate XII) .... 59
Edith R. Saunders. Note on the Evolution of the Double Stock

(Matthiola incana). (With three Text-figures) .... 69

H. L. Trachtenberg. The Analysis of the Results of Professor

Johannes Schmidt's Diallel Crossings with Trout ... 75

Ei.iZABETH Robertson. Notes on Breeding for Increase of Milk in

Dairy Cattle. (With Eight Pedigree Charts) .... 79

W. Bateson. Root-Cuttings and Chimaeras. II. (With Plates XIII

(coloured) and XIV) 91

No. 2 (September, 1921)

Frances Pitt. Notes on the Genetic Behaviour of certain characters
in the Polecat, Ferret, and in Polecat- Ferret Hybrids. (With
one Text-figure and Plates XV and XVI) 99

C. H. Ostenfeld. Some Experiments on the Origin of New Forms

in the genus Hieracium sub-genus Archieracium. (With Plates
XVII and XVIII) 117

H. Onslow. The Inheritance of Wing-Colour in Lepidoptera. V. Me-
lanism in Abraxas grossulariata (var. varleyata, Porritt). (With
six Text-figures and Plate XIX) 123

F. A. E. Crew. Sex-Reversal in Frogs and Toads. A review of the
recorded cases of abnormality of the reproductive system and an
account of a breeding experiment. (With twenty-three Text-
figures) . . . . . . . . . . .141

I. Leitch. a study of the Segregation of a Quantitative Character in
a cross between a pure line of Beans and a Mutant from it.
(With four Text-figures) 183

vi Contends

No. 3 (December, 1921)

FAQE

BuNUO MiVAZAWA. Dwarf Forms in Barley. (With Plate XX) , . 205

F. R. Parnbll. Notes on the Detection of Segregation by the exami-
nation of the Pollen of Rice. (With Plate XXI) . . .209

Rudolph Beer. Notes on the Cytology and Genetics of the Genus

Fuchsia. (With Plates XXII— XXIV) 213

J. S. Huxley. Linkage in Gammarus chevreuxi . . . . 229

R. C. PuNNETT and M. S. Pease. Genetic Studies in Poultry. IV. On

the barred plumage of certain breeds. (With two Text-tigures) . 235

E. S. Salmon and H. Wormald. A study of the Variation in Seedlings

of the Wild Hop (Hmnulus Lupulus L.). (With Plate XXV) . 241

W, Bateson and Miss A. E. Gairdner. Male Sterility in Flax,

subject to two types of Segregation. (With Plate XXVI) , . 269

H. Onslow. The Inheritance of Wing-Colour iu Lepidoptera. VI.
Diaphora mendica CI. and var. rustica Hb. (With ten Text-
figures and Plate XXVII) 277

H. Onslow. The Inheritance of Wing-Colour in Lepidoptera. VII. Me-
lanism in Hemerophila abrujjtaria (var. fuscata Tutt). (With
Plate XXVIII) 293

Kirstine Smith. Remarks on the method of calculation proposed by

Mr H. L. Trachtenberg for Diallel Crossings .... 299

Vol. II, No. I

April, 1 92 1

JOURNAL OF GENETICS

EDITED BY

W. BATESON, M.A., F.R.S.

DIRECTOR OF THE JOHN INNES HORTICULTURAL INSTITUTION
AND

R. C. PUNNETT, M.A., F.R.S.

ARTHUR BALKOUR PROFESSOR OF GENETICS IX THE UNIVERSITY OF CAMBRIDGE

CAMBRIDGE UNIVERSITY PRESS

C. F. CLAY, Manager
LONDON : FETTER LANE, E.G. 4.
also

H. K. LEWIS & CO., LTD., 1 36, COWER STREET, LONDON, W.C. I
WHELDON k WESLEY, LIMITED, 28, ESSEX STREET, LONDON, W.C. 2

CHICAGO : THE UNIVERSITY OF CHICAGO PRESS
(Agent for the United States and Canada)
BOMBAY, CALCUTTA, MADRAS : MACMILLAN & CO., LTD.
TOKYO : THE MARUZEN-KABUSHIKI-KAISHA

Price Fifteen Shillings net

Issued 21 Aprils 192 1]

THE UKIYERSITY OF CHICAGO PRESS

The Origin of the Earth. (The University of Chicago
Science Series.) By Thomas C. Chamberlin, Head of the Depart-
ment of Geology in the University of Chicago, xii + 272 pages.
12mo. Cloth, lis 6d net.
This book, by oue of the leading geologists of the world, sets forth the disclosures
that led to the rejection, one after another, of the older views of the origin of our
planet, the futile attempts then made to amend these or to build others upon the
same foundations, the final rejection of all these, and the construction of a radically
new view based on a new dynamic foundation. The later chapters of the book treat
of the early stages of the earth and of the way in which its leading processes took
their start from their cosmogonic antecedents, these being held to be essential
factors in the genesis of the planet.

Heredity and Eugenics. By John M. Coulter, William
E. Castle, Edward M. East, William L. Tower, and Charles
B. Davenport. 312 pages. 8vo. Cloth. 19s 6d net.

Heredity and Eugenics is a book intended to present the recent developments of
knowledge in reference to evolution, heredity, and eugenics, and related subjects.
Five of the leading investigators in this field have collaborated in its production,
and the result is a woi-k which will appeal to the general public, because it presents
the latest conclusions in a popular and extremely interesting manner.

Senescence and Rejuvenescence. By Charles Manning

Child, Associate Professor of Zoology in the University of Chicago,
xii + 482 pages. Svo. Cloth. 26s net.

The author of this volume, after some fifteen years of experimental investigation
of the nature and origin of the organic individual, has established certain facts which
afford a more adequate foundation for the general consideration and interpretation of
the age changes in the organic world than we have hitherto possessed.

Certain experimental methods have made it possible not only to follow the
phj'siological age changes in some of the lower animals, but to learn something of
their nature. The most important I'esult of this investigation is the demonstration
of the occurrence of rejuvenescence quite independently of sexual reproduction.
The book differs from most previous studies of senescence in that it attempts to
show that in the organic world in general rejuvenescence is just as fundamental and
important a process as senescence.

Individuality in Organisms. (The University of Chicago
Science Series.) By Charles Manning Child, x + 212 pages.
Small 12mo. Cloth. 10s net.

Professor Child's work is an attempt to state, and to present evidence in support
of, a conception of the nature of organic individuality which the author has developed
as a I'esult of fifteen years of investigation of the processes of reproduction and
development in the lower animals. In these forms organic individuality appears in
relatively simple terms, and it is hei'e that we must look for the key to the problem
of individuality in the higher animals and man. The book appeals not only to the
physiologist and to the botanist, but also to the neurologist, to the psychologist, and
even to the sociologist.

The Biology of Twins. (The University of Chicago Science

Series.) By Horatio H. Newman, Associate Professor of Zoology

in the University of Chicago. x + 186 pages. Small 12 mo. Cloth.

10s net.

In this volume many questions i-elative to twinning have been answered partly

through a study of data on human twins themselves and partly through a much

more detailed study of what has proven to be the most remarkable type of twinning

known — that in the armadillos of the genus Dasypus.

Sold in the British Empire (except Canada) by the
Cambridge University Press, Fetter Lane, London, E.G. 4

Volume XI APRIL, 1921 No. 1

STUDIES OF INHERITANCE IN THE JAPANESE
CONVOLVULUS. PART II.

By B. MIYAZAWA

(With One Coloured Plate.)

INTRODUCTION.

Since the culture experiments ending with 1918 have led me to
certain definite conclusions in respect to the hereditary behaviour of
various shades of flower-colours which appear in the F.2 generation, and
which were shortly noticed in my previous paper^ I should like to describe
below the results of these experiments. In my previous article I have
mentioned the fact that yellow plants never bear dark-red flowers, but
recently I was able to get a yellow-leaved race which nevertheless pro-
duces dark-red flowers ; and as I have conducted hybridisation experiments
between such a race and each of the two original parents, the results
so far obtained will be given also in this paper.

First of all, I will mention here briefly some results obtained by
many investigators as regards the inheritance of flower-colours which
seem to have special relations with my own studies.

Bateson and Punnett^ reported in Lathyrus odoratus several cases
where the flower-colour presented by neither parent appears in ^i or in
and after F^_. Saunders'* found in Matthiola that certain two white-
flowered individuals crossed together gave coloured F^, and such a case
was also reported by Marryat^ and by Takezaki^ in Mirahilis and Japanese
Convolvulus, respectively. Saunders" obtained purple-flowered F^ between
pink and white varieties of Salvia Horminum, and in F.2 the ratio was

1 Journal of Genetics, Vol. viii. No. 1, p. 62, 1918.

2 Beports to the Evolution Covimittee of the Royal Society. II. pp. 83 — 99, 1905.

3 L.c. I. p. 45.
^ \ ^ L.c. V. p. 46.

Osl ^ " Nippon Ikusyugakukwai Kwaiho " (Journal of the Japanese Breeders'' Association),
^I. 1, Tab. V and VI. 1916.

6 Heports to the Evolution Committee of the Royal Society, 1905. II. p. 50.

T-H Jouin. of Gen. xi 1

>-

2 Inheritance in the Japanese Convolvulus. Part II

9 purple : 3 pink : 4 white. Tschermak^ also got similar results in
Pisum sativum. All these cases shew that the cooperation of the two
factors is necessary for the production of anthocyanin.

The fact that the diluting or modifying factor is present in the flower
of Matthiola was confirmed by Saunders-; Bateson and Punnett^ also
reported a case in Lathyrus that can be explained by the presence of
the diluting factor.

Baur^ found in Antirrhinum that zygotes heterozygous for L in the
presence of D are pale magenta in flower-colour, whereas those homo-
zygous for L are intermediate magenta in the same respect. Marryat^
discovered in Mirabilis the occurrence of heterozygous forms with flower
colours which may be best represented by the following formulae :

CCMM crimson

CcMM magenta

CcMm magenta rose

CCMm orange red

CCmm yellow

Ccmm pale yellow

A case where the effects of the one factor, T, are not manifested
unless the other, L, is at the same time present in the zygote was
studied by Wheldale" in Antirrhinum; namely, no magenta colour
appears in the flower tube unless there is magenta colour in its lips.

Bateson^ studied the relation between the colour and the shape of
flower in Lathyrus, and found that in that having the hooded standard
the colours of the standard and wings agree with each other, but in that
having the erect-standard they do not.

EXPERIMENTS.
I. i^2 GENERATION.

As the characters of both parents as well as the F^ plant are already
described in Part I* of this paper I will not repeat them here, but as
to the flower-colour of each F^ plant, though it was described in the

1 Zeitschriftfilr induktive Abstammungs und Vererbtingslehre, 1912, Bd. vii. pp. 81 — 234.

2 Proceedings of the Royal Society, Vol. lxxvii. 1905, pp. 236 — 238.

3 Reports to the Evolution Committee of the Royal Society, 1906. III. p. 38.

■* Zeitschrift fiir induktive Abstammungs und Vererbungslehre, 1910, Bd. iii. pp. 34 — 98.
•^ Reports to the Evolution Committee of the Royal Society, 1909. V. pp. 42 — 46,
fi Proceedings of Royal Society of London, 1907, Vol. lxxix. B, pp. 288 — 305.
7 Journal of Genetics, Vol. viii. No. 1, pp. 61, 62.
s 7..C. pp. 61, 62.

B. MiYAZAWA 3

previous paper^ I should like to repeat it here shortly. In F2 not only
do we find flowers of white, dark-red and magenta colour, exactly similar
to those of the two original parents and the ^1 plant, respectively, but
also we have those of scarlet colour ; and besides, in each of these colours
— dark-red, magenta, and scarlet — there are three gradations of tone,
sharply distinguishable from each other, which I will call light, medium
and deep, respectively. (See PI. I, figs. 1 — 9.)

The details of the segregation of leaf- and flower-colours in F2 arc
shewn in Table I.

TABLE I.

Actual numbers

Leaf-colour Flower-colour

of individuals

Expected

, light mageuta

164

166-500

magenta

74

83-250

deep magenta

72

83-250

light scarlet

48

55-500

„

scarlet ...

24

27-750

^reen <( ^eep scarlet

24

27-750

light dark-red

111

111-000

dark-red

69

55-500

deep dark-red

58

55-500

I white

217

222-000

/ light magenta

95

83-250

magenta

45

41-625

deep magenta

49

41-625

light scarlet

32

27-750

Yellow ^

scarlet
deep scarlet

14
14

13-875
13-875

light dark-red

0

—

dark-red...

0

—

deep dark-red

0

—

\ white

74

74-000

Totals

1184

Here it may be remarked that in the above Table the results of two
reciprocal crosses as well as those of the Fo experiments repeated in
1917 are all summed up, because I have found no essential difference
among all these results.

In Table I we see that in green plants the ratio between magenta,

scarlet, and dark-red is nearly equal to 3 : 1 : 2, and that in yellow plants

that between magenta and scarlet is nearly equal to 3:1, and moreover,

it will be seen that of the three gradations of the tone of each colour

the ratio of light : medium : deep is 2:1:1. In order to explain such

results I have adopted the following genetic formulae for the two

parents :

^=ggddBBMM,

B = GGDDbbmm.

L.c. p. 62.

1—2

4 Inheritance in the Japanese Convolvulus. Part TI

The meaning of the factors is as follows :

G, the factor for green colour in the leaf

D, that for dark-red flower-colour when the accompanying G is in
homozygous condition ; it produces, however, scarlet colour when G is
present as Gg, or when it is altogether absent (i.e. gg) and B is also
absent.

B, that for blue colour; in the presence of D its effects are not
manifest when G is present as GG ; it produces magenta colour both in
homo- and heterozygous condition, when G is in the condition Gg or gg.

M, that for modifying the tone of flower-colour both iu homo- and
heterozygous condition, i.e. the medium grade of magenta, scarlet, or
dark-red colour is produced when M is present and the plant is also
homozygous for D ; the light grade of each of these colours is produced
when M is present and the plant is at the same time heterozygous for D ;
and finally the deep grade appears when M is altogether absent.

The genetic constitution of each F^ plant is naturally GgDdBbMm,
and when it is self-fertilised we expect 81 zygotes of different genetic
constitutions, as is shewn in Table II. Thus we may acknowledge that
my hypotheses above stated will well explain all my results. It is
necessary to add here that though the meaning of the factor D is some-
what more broadened in this than in my first paper, inasmuch as not
only is its action intimately bound with G but also with B, nevertheless
my view in respect to the relation between green leaf and dark-red
flower-colour remains quite unchanged.

II. Back-crossing.

In 1916 the back-crossing of one F^ plant by both of the two
parents, A or B was done.

1. FiX^ (= GgDdBbMm xggddBBMM).

The results are shewn in Table III.

TABLE III.

Light miigenta White Totals

Green ... 33 48 81

Yellow ... 45 38 83

Totals ... 78 86 164

\ 41 41

41 41

Expected

In this case, as was described in my previous paper, it will be seen
that the ratios between green and yellow plants and between light
magenta and white are 1 : 1, respectively; and besides, coloured flowers

B. MiYAZAWA

TABLE II.

Genetic

Number of

Flower

constitutions

individuals

colour

GgDdBBMM

(4)

GgDdBBMm
GgDdBbMM

{^s! (36)

Light magenta n^

GgDdBbMm

(16)

GgDDBBMM

(2))

GgDDBbMM
GgDDBBMm

(8))

Magenta

GgDDBbMm

GgDDBBmm

(2)

GgDDBbmm
GgDdBBmm

(4)
(4)1

(18)

Deep magenta

GgDdBbmm

(8)1

GgDdbbMM
GgDdbbMm

i^!! <-'

Light scarlet

GgDDbbMM
GgDDbbMm

SI <»•

Scarlet

GgDDbbmm
GgDdbbmm

<2)l (6)
(4) J ^^>

Deep scarlet

GGDdBbMM

(4)\

GGDdBBMm

(4)

GGDdBbMm
GGDdbbMM

l2)r(24)

Light dark-red

GGDdBBMM

(2)

J-

GGDdbbMm

(4)/

GGDDBBMM

(1)^

GGDDBbMM

(2)

GGDDBBMm
GGDDbbMM

(2)
(1)

-(12)

Dark-red

GGDDBbMm

(4)

GGDDbbMm

(2).

GGDDBBmm

(1)^

GGDDbbmm

(1)

GGDDBbmm
GGDdBBmm

(2)
(2)

r(12)

Deep dark-red

GGDdbbmm

(2)

GGDdBbmm

(4) J

GGddBBMM

(1)^

GGddbbMM

(1)

GGddBBmm

(1)

GGddBbMM

(2)

GGddBBMm

(2)

GGddBbMm

(4)

GGddBbmm

(2)

GGddbbMm

(2)

GGddbbmm
GgddBBMM

(1)
(2)

y (48)

White ^

GgddBBMm

(4)

GgddBbMM

(4)

GgddBBmm

(2)

GgddbbMM

(2)

GgddBbMm

(8j

GgddBbmm

(4)

GgddbbMm

(4)

Ggddbbmm

(2),

Leaf
colour

y Green

Genetic
constitutions

ggDdBBMM
ggDdBBMm
ggDdBbMM
ggDdBbMm

ggDDBBMM
ggDDBbMM
ggDDBBMm
ggDDBbMm

ggDDBBmm
ggDDBbmm
ggDdBBmm
ggDdBbmm

ggDdbbMM
ggDdbbMm

ggDDbbMM
ggDDbbMm

ggDDbbmm
ggDdbbmm

ggddBBMM
ggddBbMM
ggddBBMm
ggddBBmm
ggddbbMM
ggddBbMm
ggddBbmm
ggddbbMm
ggddbbmm

Number of
individuals

Flower
colour

Leaf
colour

(2))
(4)
(4)
(8)1

(1))
(2)
(2)
(4))

(1))

(2)

(2)1

(4)1

(2)1
(4))

(1)1
(2))'

(1)1
(2)j

(2)^

(2)

(1)

(1)

(4)

(2)

(2)

(1)

(18) Light magenta ~i

(9) Magenta

(9) Deep magenta

(6)
(3)
(3)

Light scarlet

Scarlet

Deep scarlet

(16) White

Yellow

6 Inheritance in the Japanese Convolvulus. Part II

are light magenta exclusively. The reason for this will readily be seen
when we compare the genetic constitutions of the F^ and A plants.

2. i\ x5(=GgDdBbMm X GGDDbbmm).

The results are shewn in Table IV.

TABLE IV.

Light Deep

Light Deep Light Deep dark- Dark- dark-

magenta Magenta magenta scarlet Scarlet scarlet red red red White Totals

Green 8 6 12 4 7 9 9 11 15 0 81

Yellow 0 0 0000 0 0000

„ . ^ \ 5-063 5-063 10-125 5-063 5-063 10-125 10-125 10-125 20-250 0
Jiixpectea |^ 0 0 0 0 0 0 0 0 0

In this case all plants have green leaves and we find magenta, scarlet,
and dark-red flowers in all their respective three tones. If we add
together individuals of each colour belonging to the same tone it will
be seen that the ratio magenta : scarlet : dark-red is 1:1:2, and in each
colour there are two plants of deep tone for each one of the two other
tones. Here again the facts are clearly in accordance with hypotheses
put forward.

III. Fs, Fi, AND F^ GENERATIONS, ETC.

I have made a series of breeding experiments on various families in
Fs, Fi, and F^ ; moreover, various crosses were made among the offspring
of the hybrids both with one another and with the original parents, the
offspring in each case being grown on. All these experiments have fully
borne out my interpretation above given, and since I have never en-
countered any contradictory case, I will not describe here details of all
these results, but simply some few selected examples.

1. Green plant with magenta flower.
The results are shewn in Table V.

^3. No. 31 j^Xv
^..No-Bl-lj^-o".

TABLE

Magenta

40
12
22
13

V.

Scarlet

8
7

7
2

Dark-red

32

0
14

0

Totals

^^\ 99
19 i ^^

Totals e'ff"

yellow

Expected

62
25

\ 58-875
i 29-437

15
9

19-625
9-813

46
0

39-250
0

'slji"

B. MiYAZAWA 7

It is apparent from the above Table that in green plants the ratio
magenta : scarlet : dark-red is 6:2:4, and that in yellow plants the
ratio magenta : scarlet is 3 : 1, so that we must consider that the genetic
constitution of the plant used is GgDDBbMM.

2. Oreen 'plant with deep magenta flower.
The results are indicated in Table VI.

TABLE

VI.

Deep
magenta

Deep
dark-red

Totals

N-9-23 feZv

51
21

23

0

lij 8=

E-P-*^<1 IS^w

47-50
23-75

23-75
0

It will be seen from the above Table that the ratio green, deep
magenta : green, deep dark-red : yellow, deep magenta is 2:1:1. Thus
we may consider the genetic constitution of the plant used to have been
GgDDBBmm.

3. Green plant ■with light scarlet flower .
The results are as follows :

TABLE VII.

Light Deep Light Deep

scarlet Scarlet scarlet dark-red Dark-red dark-red White Totals

No 60 1^'^^^'' 20 10 11 7 6 4 20 78) ,

^o- ou I ^,^jjp^ 7 4 4 0 0 0 8 23^"^

Ti,,„,„, , \18-938 9-469 9-469 9-
Uixpeciea | g.^^^ ^.^^^ ^.^^^ ^

\18-938 9-469 9-469 9-469 4-734 4-734 18-938

0 0 6-313

From the above Table we see that in green plants light scarlet :
scarlet : deep scarlet : light dark-red : dark-red : deep dark-red : white are
in the ratio 12:6:6:6:3:3:12, and in yellow plants light scarlet:
scarlet : deep scarlet : white are in the ratio 6:3:3:4. Thus the genetic
constitution of the plant used must have been GgDdbbMm.

4. Green plant luith scarlet flower.
The results are shewn in Table VIII.

TABLE VIIL

Scarlet Deep scarlet Dark-red Deep dark- red Totals

F. xB No IS \?>'^^^^ 41 13 18 8 80) ,„^

i-ixii, No. 13 lygiio^ 20 5 0 0 25^^^^

T?^„_.„j \ 39-375 13-125 19-688 6-563

^^P^*'*^*^ ) 19-688 6-563 0 0

8 Inheritance in the Japanese Convolvulus. Part II

This Table shews that the ratio scarlet : deep scarlet : dark-red : deep
dark-red is 6 : 2 : 3 : 1 in green plants and also, that scarlet : deep scarlet
is .3:1 in yellow plants. Accordingly the genetic constitution of the
plant used must have been GgDDbbMm.

5. Green plant with light dark-red floiver.
The results are indicated in Table IX.

TABLE

IX.

Light dark-red \

Dark-red

White

Totals

No. 14

45

20

14

79

„ 55

28

9

11

48

„ 14-2 ...

48

33

25

106

„ 19 (1918)

65

38

36

139

Totals

186

100

86

372

Expected ...

186-00

93-00

93-00

From the above Table we see that the ratio light dark-red : dark-
red : white is 2 : 1 : 1. Let us now consider what should be the genetic
constitution of the plant used. From the explanation of the results in
F2 (Table II) we know that there should be six different genetic con-
stitutions in plants with light dark-red flowers. But our hypothesis
assumes that when G is in homozygous condition in the presence of D,
the effect of B is not manifest, so that in this case, the result is quite
the same, whether B is present as BB, Bb or bb. Consequently, the
genetic constitution of the four families above mentioned must corre-
spond to either one of GGDdBBMM, GGDdBbMM or GGDdbbMM;
and in order to determine which of these three is the actual one, the
hybridisation experiments with other families would be necessary.

6. Yellow plant with light magenta flower.
The results are indicated in Table X.

TABLE X.

Light magenta

Magenta Light scarlet

Scarlet

White

Totals

No. 39—1...

47

26 19

6

40

138

Expected ..

51-750

25-875 17-250

8-625

34-500

From the above Table it is quite evident that the ratio light
magenta : magenta : light scarlet : scarlet : white is 6 : 3 : 2 : 1 : 4, so that the
genetic constitution of the plant used must have been ggDdBbMM.

B. MiYAZAWA 9

7. Yelloio plant with magenta flower.
The results are shewn in Table XI.

TABLE XI.

Magenta Deep magenta Scarlet Deep scarlet Totals
No. 39— 7 ... 64 22 22 10 118

Expected ... 66-375 22-125 22-125 7-375

In the above Table we see that the ratio magenta : deep magenta :
scarlet : deep scarlet is 9:3:3:1. Thus the genetic constitution of the
plant used must have been ggDDBbMm.

8. Yellow plant with light scarlet flotuer.
The results are shewn in Table XII.

TABLE XII.

Light scarlet

Scarlet

White

Totals

No. 22 4— 6 ...

32

14

15

61

„ 22 4 10 ...

17

14

8

39

Totals

49

28

23

100

Expected

50

25

25

Thus the ratio light scarlet : scarlet : white is 2:1:1, so that the
genetic constitution of the plants used must have been ggDdbbMM.

9. Yelloiu plants with dark-red flowers.

As was reported in my previous paper\ I found in ^4 a family (44 — 6)
which, in spite of the fact that it segregates into green and yellow
plants, yet breeds true to dark-red flowers ; evidently there must exist
some special reason for this fact, and the experimental results conducted
on this family in F^, are as follows :

No. 44—6-

The above Table shews that all the families have dark-red flowers in
spite of their yellow leaf-colour. What then, we may ask, is the cause
which has led to such contradictory results ? Now let us consider in

1 Journal of Genetics, Vol. viii. No. 1, p. 73.

TABLE

XIII

Leaf colour in

F,

Leaf colour in F^

Dark-red

2

Yellow

Yellow

81

• 4

) J

,,

25

10

Green

Green

40

■ 7

\ Green

22

>>

) Yellow

6

10 Inheritance in the Japanese Convolvulus. Part II

detail the ancestors of these families. The seeds of these families were
taken from the individuals which segregated in the following manner
ini^3:

Scarlet Deep scarlet Dark-red Deep dark-red Totals

^'°-" Irene", '\ I I % ??!'»

The family 44 — 6 was the offspring derived by self-fertilisation from
a plant having green leaf and dark-red flower. Consequently if the
hereditary behaviour of these characters were normal we should have in
the offspring leaves which are constantly green and flowers which are
either constantly dark-red or will segregate into dark-red and deep dark-
red. The results, however, have shewn that notwithstanding the fact
that the flower-colour is constantly dark-red the leaf-colour segregates
into 3 green and 1 yellow. So that it may be considered that there
occurred some permanent variation among the factor or factors in the
F^ plant used for self-fertilisation. But we cannot decide how and where
such change has occurred, and any hypothesis on this point would be
useless unless founded on the facts actually obtained, so that this case
will remain the subject of my future study.

IV. Hybrids between J. or 5 with a race G which has

YELLOW leaf AND DEEP DARK-RED FLOWER.

In all cases described till now there were found no yellow plants
with dark-red flowers, except the case given above. Such a plant, how-
ever, is found among certain races of Japanese Convolvulus, and it seems
to have been in cultivation more than 60 years ago, as it is described in
a book entitled Santo Ittyo published in 1854, in which many coloured
figures are found. My material is characterised by having the " hukurin "
part on the margin of the flower. Although I have described the flower-
colour of this race as deep dark-red, I found that certain differences are
discernible between B and C, inasmuch as the colour of the latter is a
little darker than that of the former. They are, however, so similar to
each other that this slight difference would not be noticeable without
direct comparison.

1. A xC and G x A,
a. Fi generation.

In both reciprocal crosses I obtained exactly the same results, the
leaf being yellow and the flower light magenta. The " hukurin " appeared
in quite the same manner as in the parent G.

B. MiYAZAWA 11

h. Fo generation.

All the individuals had yellow leaves. As in the case oi A x B we
obtained here also flowers of magenta, scarlet, and dark-red colour, and
moreover there were found the three usual tones in each. All plants
with coloured flowers had the "hukurin" on the corolla. It must be
said that all the tones of flower-colour in G x A are somewhat different
from those of F2 plants in A x B, and moreover, that the magenta and
scarlet colour are slightly more blue-tinged in the former than in the
latter.

The results of both reciprocal hybridisation are shewn collectively in
Table XIV.

TABLE XIV.

AxC

AxC

Totals

Expected

Light magenta

.33

36

69

6'J-563

Magenta...

16

20

36

34-78ii

Deep magenta

14

23

37

34-782

Light scarlet

13

16

29

34-782

Scarlet

7

10

17

17-391

Deep scarlet

7

9

16

17-391

Light dark-red

15

19

34

34-782

Dark-red ...

11

11

22

17-391

Deep dark-red

8

9

17

17-391

White

44

50

94

92-750

Totals

168

203

371

It will be readily seen from the above Table that the results are
altogether similar to those of J. x B.

% BxC.

a. Fi generation.

Fi plants had green leaves with deep dark-red flowers.

b. Fo generation.

The flower of all Fo plants was deep dark-red, and the leaf-colour
segregated into green and yellow. As previously mentioned, there was
a little difference between the flower-colours of B and G, and this differ-
ence appeared in F.^, but as it was extremely difficult to distinguish
them clearly I did not undertake to do this work. I will only mention
here, however, that also in yellow plants the flower-colour similar to
that of B parent has been observed. The results are indicated in
Table XV.

In the above Table we see that the ratio green, "hukurin" : green,
fully-coloured : yellow, "hukurin" : yellow, fully-coloured is 9 : 3 : 3 : 1, so

12 Inheritance m the Japanese Convolvulus. Part II

that it may be seen that the leaf-colour and the "hukurin" are inherited
in a simple Mendelian fashion.

TABLE

XV.

Number 1 Number 2

Totals

Expected

p,.nnr, S white-margined ..,
ureen | fyHy.coloured ..,

48
18

31

9

79
27

75-938
25-313

v„ii„,„ i white-margined ..
^^"°^^ \ fully-coloured ..,

14
6

7
2

21

8

25-313

8-438

Totals 86 49 135

From the results oi A xG, C x A, and i? x C it is clear that the
genetic constitution of the G parent is ggDDbbmm with respect to the
leaf- and flower-colour. Yet if we will speak more strictly there may
exist some differences as regards the factor D between B and C, but I did
not undertake to study these differences.

It is quite clear from the above results that G has a factor to produce
the "hukurin" on the corolla; moreover that this factor is present in
A but not in B has been fully confirmed.

V. Some facts observed during the experiments.

1. Relation between the colour of flowers and that of other
parts of plant.

Many observations have already been made by several authors as
regards the relation between the colours of flowers and those of other
parts of plant. The phenomenon most commonly known is that where
positive relations exist between them. Gregory^ observed in Primula
sinensis that anthocyanin may be entirely absent when the stem is green
and that pale flower-colour is associated with faint colour in young leaves
as well as green or faintly coloured stem ; besides, he has discovered the
following associations : deep flower-colour and deep stem-colour, true
red flower-colour and true red stem, blue flower-colour and blue stem.
Nohara^ observed in his study of inheritance on Oxalis corniculata that
the purple colour in the eye of the corolla and in the leaf are associated
to each other, but the leaf-purple can appear without being associated
with the eye purple. I^ have reported in the garden varieties of Rhodo-
dendron ohtusum, Rh. ledifolium var. purpureum, and Rh. indicuni var.
macranthum which are cultivated in Japan, that the more intense the

1 Journal of Genetics, 1911, Vol. i. pp. 73—132.

- Journal of College of Agriculture, Tokyo, 1915, Vol. vi. pp. 165 — 181.

'^ Journal of the Scientific Agricultural Society, Tdkyo, 1914, No. 145, pp. 1 — 6.

B. MiYAZAWA 13

reddish-brown colour of winter leaves are, the deeper the flower-colour
is ; besides I have reported that white flower-colour is associated with
gi'een leaf, and striated flower with striated leaf.

The above statements shew that the colours of flower and leaf or
stem are due to one and the same factor. There are some cases, however,
where the colour of flower and that of other plant-organs seem not to be
determined by one factor. Shull^ found in the hybrids between Oenothera
ruhricalyx smdrubriJiervis and those between ruhricalyx and Lamarckiana
that pigmented buds of ruhricalyx is invariably associated with a low
degree of pigmentation in stems and rosettes. Besides these, many other
examples were reported in Primula sinensis, Helianthus and Lathyrus, etc.

In the Japanese Convolvulus white flowers may be associated with
green or yellow stem, and coloured flowers with pigmented stem. In my
experiments the stem-colours in plants with deep-coloured flowers were
found to be deeper than in those of light-coloured individuals, and more-
over the stems of plants with dark-red flowers were dark-reddish-brown
and those of scarlet flowers were reddish-purple. But I could not discern
the differences of stem colours between plants with magenta and those
with scarlet flowers. Thus the relation between the colours of stem and
flower is similar to what we see in Primula sinensis.

2. On streaked flowers.

Though there are many investigations on the inheritance of streaked
flowers in Antirrhinum, Mirahilis, etc., I will mention here simply the
facts observed in Japanese Convolvulus. In and after the F^ generation I
have often observed that just one half of a petal to the throat produced
a colour entirely different from that in other parts. Such phenomenon
did not occur in all flowers of a plant but only in one or two out of fifty
or more. Thus I have seen a light magenta part in a light scarlet flower,
a magenta part in a scarlet flower, and a deep magenta part in a deep
dark-red flower. These are the instances where the dominant colour was
produced in a recessive coloured flower, though I have once seen a deep
scarlet part in a deep magenta flower, i.e. recessive colour in dominant
coloured flower.

Emerson^ studied the occurrence of anomalous seeds of maize as
regards their pigmentation. In one case, the seed was half colourless

1 Journal of Genetics, 1914, Vol. iv. pp. 83—102.

2 Zeitschrift fiir induktive Abstammungs und Vererbungslehre, 1915, Bd. xiv. pp.
241—259.

14 Inheritance in the Japanese Convolvulus. Part II

and half purple ; in the other, half purple and half red. His opinion on
the occurrence of this phenomenon is that it is due to a somatic mutation,
that is a change in genetic constitution rather than a segregation • of
genetic factors, and this somatic mutation may be a gain of at least one
new factor, the loss of a factor, or the permanent modification of a factor.
Moreover he brings against the segregation hypothesis the following
considerations. "If a dominant character appears as a bud sport, in
material known to be homozygous with respect to a recessive character
that is alleloraorphic to the dominant character in question, it seems
clear that a somatic mutation is the responsible agent.... It would be
interesting to know whether recessive bud sports actually occur much
more frequently in heterozygous than in homozygous material. If this
is found not to be the case, it will have an important bearing upon the
problem of whether bud sports are mutations or segi'egations, for the
latter would occur only in heterozygous material."

Above we have described three cases of Japanese Convolvulus where
the dominant character has arisen in material known to have the recessive
characters. Accordingly it is impossible to consider this phenomenon
merely due to the segregation of factors, but it will be necessary to think
that there occurred some change in the somatic cells which have had
recessive characters. I did not cultivate the offspring of the plant which
produced a deep scarlet coloured part in a deep magenta coloured flower,
so that I cannot decide whether the original colour was in a homozygous
or a heterozygous condition. If in the latter condition we may consider
the phenomenon due either to the segregation of factors or to somatic
mutation, but if in homozygous condition the latter would seem to be
the real cause of the phenomenon.

3. The relation between leaf-colour and the growth habit of plant-body.

The parent B, which has green leaves, grows vigorously. It has a
big and long stem with long internodes, and side branches appear at
the fifth or sixth leaf axil from the base. A on the other hand has a
slender stem with side branches appearing at the second or third leaf-
axil and flowers opening earlier than B. The F^ plant has characters
similar to those of B, at least as regards the above stated points. In
the F<i generation also these characters were found to be associated with
the leaf-colour, though I did not undertake to measure. It is quite clear
that the weak gi-owth of yellow-plant is due to small amount of green
pigment in them.

JOURNAL OF GENETICS, VOL XI. NO. 1

PLATE

^.4£^

I 1

'^

n 4

\ i

//

f

v'^

"L. N. MiDl'SIMA

Cambridge Univertity Prts»

B. MiYAZAWA 15

Summary.

1. The light magenta colour in F^ is produced when both G and D
are in heterozygous condition ; and the bluing-factor B and the modifying
factor M are brought in from the parent A.

2. Reciprocal hybrids are similar to each other in all respects.

3. D produces dark-red colours when G is present in a homozygous
condition, but it produces red-colours (magenta and scarlet) when G is
present in a heterozygous condition or altogether absent. Such an inter-
relation between G and D was found only in the hybrids between A and
B, and it does not exist in other hybrids, though (7, for instance, has
colours closely related to those of B.

4. The fact that D has such a character may be seen closely from
the results in which all three families of the offspring of (10 — 1) x (16 — 9)
which is green, white and yellow, deep scarlet, respectively, produced
dark-red colours.

5. The effects of B are not manifested in the individuals which are
in a homozygous condition with respect to G.

6. The magenta colour appears in plants which have the con-
stitution DB, either Gg or gg being present at the same time. On the
contrary, the scarlet colour appears only in plants which are in the
condition Db.

7. The white colour appears in the individuals when D is altogether
absent, and then G, B, and M may be in any condition. So that there
are various genetic constitutions among white-flowered plants.

8. The interrelations between D and M are as follows :

DdM light colour

DDM medium colour

DDmm) , ,

^ , \ deep colour

Ddmm | ^

9. The magenta colour is dominant over both scarlet and dark-red,
and scarlet is dominant over dark-red.

10. There may exist homozygous plants, at least with respect to
flower-colour, with medium and deep tones of magenta, scarlet, and
dark-red ; but we could find no individuals with light tones of each
colour.

EXPLANATION OF PLATE I.

Figs. 1, 2, 3. Light, medium, and deep magenta.
Figs. 4, 5, C. „ ,, ,, scarlet.

Figs. 7, 8, 9. ,, ,, ,, dark-red.

A GRAFT-INFECTIOUS DISEASE OF DATURA
RESEMBLING A VEGETATIVE MUTATION.

By a. F. BLAKESLEE,
Station for Experimental Evolution, Cold Spring Harbor, N.Y.

(Received May 15. 1920.)

(With Plates II— VI.)

I. Introduction.

At the New York meeting of the Botanical Society of America in
1916 the writer presented a paper entitled "A Recurrent Mutation in
Datura Suggesting Vegetative Segregation." As reported at the next
meeting of the Society, more recent experiments have shown that the
peculiarity is in the nature of a disease since it can be transmitted by
grafting upon normal stock. The variation involved, among changes in
other morphological characters, a suppression of spines on the capsules.
As observed by earlier geneticists, it had been assumed to have resulted
from imperfect dominance of spininess when the normal type was
crossed with a smooth-capsuled variety. In as much as the disease is
transmitted through both seed and pollen and causes profound mor-
phological changes in the individuals affected, it at first appeared to
be genetic in nature. It seems appropriate therefore to publish the
present paper in a journal of Genetics rather than in one devoted to
phytopathology with the thought that results obtained in Datura may
be of possible value to geneticists in explaining peculiarities in their
breeding experiments with other species.

The Jimson Weed, Datura Stramonium^, is historically of interest
to students of heredity, since before the appearance of Mendel's work

1 A number of species have been established based on forms which differ genetically
from D. Stramonium by single factors. Thus D. Tatula is the purple spiny form,
D. laevis of Naudin and D. Bertolonii of Godron is the white, smooth-capsuled form.
The contrasting pairs of genetic characters so far discovered are spiny and smooth
capsules, purple and white flowers associated with purple and green stems, and many and
few nodes between the cotyledons and the first fork. We have preferred to include under
the name D. Stramonium all the forms which differ by single Mendehzing factors as well
as various named biotypes which differ from D. Stramonium at best in minor characters
chiefly of leaf and habit.

Journ. of Gen. xi 2

18 Graft- Infectious Disease of Datura

on jjeas, Naudin (7) and Godron (6) had used this species in studies of
inheritance and with it had discovered the reappearance of the con-
trasting parental types in the i^a generation. De Vries (8) and later
Bateson and Saunders (2) used the same species soon after the re-
discovery of Mendel's Law in studies on the inheritance of spininess
and colour. The present writer with Avery (4 and 5) has also accumu-
lated additional data on the inheritance of these and other characters.

Plants showing the disease have been called Quercina (abbreviated
to Q.) on account of peculiarities in leaf structure. The first Quercina
plant was found in cultures at the Conn. Agric. College in the fall of
1915. It was considered a spineless mutant of the purple-stemmed
form. Two open-pollinated capsules had set, and two flowers later
opened, which, since their stamens were devoid of pollen, were arti-
ficially pollinated by other plants. The seed thus obtained was the
origin of a series of breeding experiments with the Quercina mutant as
it was then called.

II. Description of Quercina Plants.

In describing the distinguishing characteristics of Quercina plants
it will be convenient to start with seedlings from the open pollinated
capsules obtained in 1915. As will be shown under breeding experi-
ments, Quercina plants do not breed true, but throw a small proportion
of normals along with Quercina seedlings. The latter usually can be
recognized at an early stage, sometimes as soon as the cotyledons have
expanded. PI. II, fig. 1 shows above three Q. plants in 3-inch pots in
contrast with three below from the same pedigree which were considered
normal at the time the photograph was taken. The early leaves of Q.'s
tend to be narrower than normals, more or less twisted, and with
indentations at the margin. PI. II, fig. 2 shows the same plants at a later
date but similarly arranged. The increasing indentation of the leaves
in comparison with the normals is evident. The plant at the lower left-
hand corner which previously was classed as normal has now shown its
Q. nature by the character of its later leaves. At times, Q. seedlings
appear normal till a relatively late stage, but in general a seedling
which has the character shows it early.

A series of Q. leaves are shown on PL IV, fig. 5 below in contrast
with a series of normal leaves above from the same line. The marked
indentation of the leaves shown in the photograph is not an extreme
condition. The leaves on some Q.'s may be eroded so far as to leave

A. R Blakeslee 19

little more than the mid-ribs. If the Globe mutant, which has broad
leaves, becomes Q., the leaves become narrower than typical Globes,
but are still broader than normal narrow-leaved types. An increased
indentation, however, is characteristic of all Q. leaves, whatever may
be the genetic type to which the plant belongs.

The stature of Q. plants is somewhat less than that of normals.
The branches are more slender and the root system is less well
developed.

Typical Q. flowers are recognizable at an early stage. A series with
buds is shown on PI. Ill, fig. 3 in comparison with normal flowers below.
The corolla of Q. flowers is split between the lobes as far down as the
insertion of the filaments as shown in the opened flower on the extreme
right from which the calyx has been removed. The corolla segments
are incurved and twisted around the base of the flower. Frequently in
the expansion of the bud they are unable to free themselves from the
calyx. The third and fourth Q. buds from the left are typical of such
" blind " flowers which may fall from the plant without opening. The
infolding of the lobes is characteristic of early stages, and hence, when
a normal bud would have its corolla nearly reaching the tip of the bud,
the incurved lobes leave an empty space inside the apex of the calyx
at the stage shown by the second bud from the left. The flowers of
purple Quercinas are darker coloured than normals and the stem colour
in such forms is also increased in intensity.

The stamens in the bud (Fig. 4) are found to be shrivelled, and
in typical Q.'s produce nothing more than a few small grains which
when tested have been found to be functionless.

The stigma in normal plants is made up of two thick lobes, facing
outward (Fig. 4). In Q. flowers the stigmatic surface is chiefly on the
inside of the lobes and runs part way down two sides of the style.

The most conspicuous peculiarity of Q. plants, however, is the
suppression of spines on the capsule. This may be complete, giving
capsules as smooth as those of the inermis variety. The form of the
capsule, however, is that characteristic of the variety or mutant aflected,
the globe mutant plants, if also Q., having flattened globose fruits.
Seeds from Q. capsules are distinctly smaller than those from normals
and their percentage of germination is less. The cotyledons are fre-
quently unable to extract themselves ft-om the seed coats without
assistance and plants may die for this reason. Older seedlings also
appear to be less vigorous than normals.

2—2

20 Graft- Infectious Disease of Datura

III. Origin of Quercinas in the Field.

The description which we have given of Q. plants is based on in-
dividuals which had come from Q. seed. The condition, however,
breaks out in the field, generally late in the season on plants which
have been produced by normal parents. The infection may be first
noticed weakly expressed in a single branch, and gradually spreads as
new growth occurs. Fig. 6 is a photograph of such a branch which
has a normal spiny capsule at the lowest fork shown. The later
capsules show a gi-adual suppression of spines to the one which is
entirely smooth. Occasionally an infected capsule is found with one or
more of the four valves smooth and the rest spiny as shown in Fig. 7
(insert, in Fig. 6). The two flowers of the infected branch, above and at
the right, have the characteristic separation of the corolla into segments.

The presence of the disease may be first observed in the change in
form of leaves and flowers. Purple flowers become darker in colour
and often slightly mottled with lighter patches, before the infection
has brought about a splitting of the corolla. As the season advances
more plants show Q. branches. When normal plants have flowered
and formed capsules, Q. plants are still flowering; for the Jimson
Weed is normally self-pollinated in the bud, and Q. plants, from their
lack of pollen, are dependent for setting capsules upon occasional off-
pollination which is especially rare in such plants on account of the
failure of many flowers to open.

Table I shows the number of Q. plants found in normal and Q.
pedigrees in the field for the year 1916. In 1917 in the same field there
were 1'67 °/^ spontaneous occurrences of Q.'s out of 5230 individuals
planted in the ordinary manner. No Q.'s from seed were planted in
the plot this year.

With only a single possible exception have Q.'s ever been observed
to occur spontaneously in the greenhouse, where Q. plants as well as
normals were under cultivation. Under greenhouse conditions plants
are not grown as large as in the field. This fact would give a lessened
opportunity for infection, but in the many thousand plants infection
would have been expected in a few if the means of transmission were
present. It is perhaps similarly because of the lack of means of trans-
mission that Q. infections have been rare in our plots near the green-
houses, while a mile away in our hill plots they have been common.
Thus, while in 1917 there were 1*7 "/„ in the 5230 plants in the regular
hill plots, there were no infections in over 2000 plants in the plots

A. F. Blakeslee 21

near the station, including pedigrees planted close and far apart.
There were at this time also 88 Q. seedlings planted in the station
plot.

In that year representatives of certain of the pedigrees were grown
in a " spread " plot with the individuals spaced five feet apart. Plants
in these spread plots had an extremely high percentage of incidence
(26*4 °/^ out of 167 individuals), while plants from the same pedigrees
elsewhere in the field, but spaced the customary one foot apart, showed
only the usual proportion of Q. plants (I'G °/„ out of 643 individuals).
The larger number of Q.'s in the spread plots may have been due to
the relatively enormous size of these plants which were not reduced in
growth by their neighbours. Each plant therefore gave a greater area
exposed to infection, and infection once taking place was carried to all
the numerous branches of the plant. Infection on the branch of a
plant in the more crowded plots, on the other hand, was apparently
unable to infect more than the few branches organically connected
with it since plants adjacent to Q.'s and with their branches touching
seem no more likely to acquire the infection than if adjacent only to
normals. What the carrier of the virus may be is at present unknown
but evidence beside that just given leads to the belief that the Q.
disease is not communicated by mere contact.

Apart from its appearance in our cultures, we have found Q. plants
upon two occasions. Once in Staten Island where several such plants
were discovered, and once at Woodbury, Long Island, where a single
Q. plant was found among a group of 182 normals. One out of five
seedlings from the latter plant was a Q., but none of the four seedlings
from the Staten Island material turned out Q. The writer has in-
spected relatively few Daturas growing wild. It is probable that a
more careful search would show the disease not uncommon in wild
plants.

IV. Breeding Experiments.

(a) Inheritance of Q. through Female Gametes.

While it was still thought that Q. was a genetic mutant, a con-
siderable amount of breeding work was carried on combining it with
all the available types of Jimson Weeds. No varietal immunity has
been discovered within the species. From the four capsules produced
by the first Q. plant discovered in 1915, 120 seedlings were produced,
of which 53 were normal and 67 Q.'s. Breeding work with typical
plants of the latter class was can-ied on in the greenhouse in 1915-16.

22 G^^aft- Infectious Disease of Datura

Since the typical Q.'s are devoid of pollen it has been necessary to
cross them with normals. Table I (p. 32) shows the results of such crosses
from field records of 1916. Q.'s are produced in 79 7o of the offspring
of Q, female parents. The remaining normal plants from Q. parents do
not carry the infection. The incidence of spontaneous infection in the
field was even somewhat less for them (0"95 °/^) than for normals free
from Q. ancestry (1'26 °l ^.

Q. plants from Q. parents usually show their Q. nature at an early
seedling stage. In 1916 the seedlings from Q. parents were labelled
"Normal," "Questionable," or "Quercina" at the time of setting into
the field from 3-inch pots. When the adult plants were inspected in
the fall, their condition was compared with the earlier records on the
pegs that remained. The results are tabulated in Table II (p. 32).

Plants classified as Q.'s in seedling stage are Q.'s when adult. The
tw^o exceptions noted out of 671 plants, if not the result of experi-
mental error, may be due to faulty classification in the earlier record.
It will be noted that in seedlings the Q. infection is determined by
leaf abnormalities alone. Apart from these two possible exceptions we
have not observed that the plants once infected ever recover. The
infection often manifests itself, however, only after the plants have
been transplanted to the field, as is shown by the 17 ""/^ of apparently
normal seedlings which later became Q.'s. There is usually little
difficulty in distinguishing Q. plants out of infected seeds from field
infected individuals, in which the character is evident later in the season
and does not involve the older leaves. About as many " Questionable "
seedlings turned out normals as there were normal seedlings which
turned out Q. If all the questionables were included with the Q.
seedlings they would have given in consequence a close approximation
to the total Q.'s found in the adult stage.

Some of the plants put into the field died before final records were
taken, and some of the labels were lost. The table includes only
individuals with complete records. Of 77 plants not included in the
table which died in the field, 24 were described on the pot labels as
Q.'s, 51 as Questionables, and two only as normals. In other words
3*4% of the seedlings classified as Q.'s died in the field, 12*8 7o of
those classified as Questionables, and only 0*7 °/^ of those classified as
normals. The questionables were probably thus classified on account
of not having developed far enough to allow a sure identification of
their character. They were thus weaker plants when set out, hence
their greater mortality. If we class all the Questionables with the

A. F. Blakeslee 23

Q.'s we have a mortality of 6*8 °/^ in Q.'s against 0'7 °/^ mortality in
normals. Whatever the figures used, the infected plants are seen to
be less likely to survive than normals. As will be noted later, the
per cent, of germination of Q.'s is lower than that of normals, hence
the final records fail to indicate the total amount of seed infection.
I'he breeding evidence shows, however, that the Q. infection is carried
by the female gametes and affects at least 79 °/^ of the seed produced
by Q. parents.

(6) Investigation of Tiuo Field-infected Plants.

In the 191G field cultures two plants (A and B), which had become
spontaneously infected and which showed the infection in different
degrees in individual branches, were chosen for detailed study. The
problem was to ascertain if any relation existed between the pro-
portion of Q. offspring and the strength of the Q. character measured
by the relative number of spines on the capsules from which the
seedlings were derived. A smooth capsule was graded 0, and one fully
armed was graded 4 ; 1, 2 and 3 were intermediate grades from very
slightly spined to a condition with spines somewhat reduced below
normal. Final records were taken of the potted plants between four and
five months after planting. A considerable number of individuals were
not graded on account of late germination, or on account of doubtful
determination at the time it was necessary to clean out the green-
house for other cultures. If such doubtful plants had been retained to
a recordable size they would probably have increased the percentages
of Q.'s. All plants that germinated however, have been included in
figuring the percentages of germination. The capsules of plant B
were all open pollinated, those of plant A were part open pollinated
and part pollinated by hand. Tables III and IV (p. 38) give details for
the two plants, and Table V (p. 34) gives summaries. Although some
individual inconsistencies occur, it will be observed that the capsules
which show evidence of infection give a higher percentage of Q. seedlings
and a lower percentage of germination than capsules which appear
normal.

A number of the capsules of plant A showed unequal distribution
of spines on the four valves of the type shown on PI. IV, fig. 7. Seeds
from individual valves were sown separately with the result shown in
Table VI (p. 34). Valves A and B form one carpel, and C and D form
the other, of the two-celled ovary. In some cases only a part of the
seeds was sown separately from the individual valves, and therefore

24 Graft- Infectious Disease of Datura

the numbers in Tables III and VI are not in all cases identical. A
correspondence between the percentage of Q. seedlings and the amount
of spininess of valves from which they were derived is scarcely evident.
We conclude that the percentage of infection of the seed is only
very roughly proportionate to the external evidence of the disease on
the part of the plant from which they were derived and that the
correspondence largely if not entirely disappears in reference to the
divisions of individual capsules. In other words, if a capsule from
a plant Avhich is becoming Q. is fully armed and thereby shows no
external evidence of infection, its seedlings are relatively low in Q.'s.
On the other hand, if the capsule shows evidence of reduction of spines
recognized by the grades employed, the seedlings are relatively high in
Q.'s. So long as the capsule appears to be infected at all, the grade of
spining has no close relation to the percentage of Q.'s in its offspring.

(c) Inheritance of Q. through Male Gametes.

It has been shown that the Q. infection is strongly transmitted
through the female gametes. The question arises as to whether it can
be carried by the pollen. A difficulty lies in the fact that typical
Q. plants are devoid of pollen. Occasionally, however, as a plant be-
comes infected in the field, some of the flowers may not be so strongly
affected as the rest and in consequence may produce a varying amount
of pollen. This was the case with five flowers of plant A already dis-
cussed. Table VII (p. 34) shows the proportion of Q. and normal offspring
when the infection is carried by the male gametes. The pollen from
flowers J. 11, A 12 and A 15 was only slightly potent in transmitting
the disease, as might be expected from the grade of their capsules, while
the flowers A 7 and A x were more active. Unfortunately the record
on flower Ax \^ not complete. It is evident that these flowers which
throw few Q.'s when selfed do not actively transmit the disease through
the pollen, while A 7 transmits the disease actively both when selfed
and when the infection is carried through the pollen. The high pro-
portion of Q.'s from the self of A 15, the pollen of which was inactive,
may have been due to an infection of the pistil which had not yet
involved the stamens. Why pedigree 16655 should run so high in
Q.'s is not clear unless a single anther of the flower was strongly in-
fected and happened to be used in making the cross. It cannot be
determined from our present records whether pollen was taken from
individual anthers or mixed in making these crosses.

The experiments show that the male gametes can carry the Q.

A. F. Blakbslee 25

infection. Since Q. pollen infects the seed, it doubtless would infect
the tissues of the plant when carried to flowers of normal plants.
Insect pollination might be one means of the spread of the disease
from lightly infected individuals with functional pollen.

V. Inoculation and Grafting Experiments.

The spontaneous appearance of Q. branches on field-grown plants
and the spreading of the character to new branches of affected in-
dividuals suggested that the peculiarity was in the nature of a disease.
In testing this supposition a series of inoculation and grafting experi-
ments were undertaken.

(a) Inoculation Experiments.

One of the commonest diseases of tobacco, another member of the
Solcmaceae, is " mosaic " or " calico " as it is called from the patches of
lighter colour on the leaves. The causal organism is not known, but
the infection is readily communicated from one plant to another by
mere contact of healthy plants with infected leaves. Mosaic diseases
of unknown nature have been discovered in other groups. It was
thought that the Q. complex might be due to such a mosaic type of
disease. Accordingly attempts were made to transmit the infection to
healthy plants by rubbing their leaves with those of Q. plants, and by
needle inoculations with the expressed juice of Q. leaves, but without
success. In one experiment two cions from normal plants were left in
the expressed juice from Q. leaves for 24 hours, and two other cions
for 48 hours, before grafting them on to normal stock. Neither of these
four cions took the disease nor caused the normal stock to turn Q.
Transmission of mosaic diseases by contact in other species is often
difficult of accomplishment and it is possible that more extensive trials
would have given occasional success with the Q. disease of the Jimson
Weed. However, mere contact of leaves cannot be a usual means of
infection in nature, since in the field normal plants, adjacent to Q.'s and
with their leaves touching, seem no more likely to be infected than
those adjacent only to normals. The disease is easily transmitted,
however, by grafting.

(b) Grafting ExpeHments.

Grafting of Q. branches on normal stock or the reverse union in-
variably has given infection. This has often resulted even when the
Q. cion has died and dropped off before a perfect union has been
established. PL V, fig. 8 is a photograph of a green-stemmed normal

26 Graft- Infections Disease of Datura

plant. Its two main branches were cut off a short distance above the
fork and on the left branch was grafted a purple Q. cion, recognized by
its eroded leaves, and on the right branch was grafted a purple normal
cion. The infection passed down the one branch and up the other into
the normal cion, causing the production of a capsule with a reduced
number of spines and a later capsule entirely smooth. The buds in
the stock below the fork also grew out with Q. characters, the one on
the left shows a single capsule with reduced spines and one entirely
smooth while the one on the right shows a single capsule likewise
entirely smooth. The infection is usually manifest as soon as new
leaves are formed and gradually spreads leaving the older parts, pro-
duced before the infection, free from abnormality and causing the new
growth to be more and more distinctly Q. in appearance.

A number of other Solanaceous species, including several different
genera, were tested by grafting with Quercina Jimson Weeds. None
was found to be so susceptible to the disease as the Jimson Weed,
however.

Datura meteloides could be infected by grafting, but the disease was
slow in showing. The disease could be brought back to the Jimson Weed
by grafting on it a cion of infected Z). meteloides. In PI. V, fig. 9 a cion
of this species is shown grafted on the normal Jimson Weed at the left
(marked Nor.) and a cion from the same plant was grafted on the
Q. plant on the right (marked Q.). The infection of the J), meteloides
shows as a mottling of the leaves with lighter patches and a puckering
of the leaf surface. This was the condition three or four months after
the graft. A few months later the leaves of the cion were free from
abnormal colour and puckering of their surfaces, but had their margins
deeply eroded. On such plants the flowers are devoid of pollen. Three
of such infected flowers were pollinated from a normal plant of B. mete-
loides and selfed seed from the pollen parent were taken as a control.
In comparison with the 10 individuals in the control, the 291 seedlings
from the Q. capsules of D. meteloides were obviously stunted and their
leaves of a paler colour. By the time of flowering, however, all of the
291 plants had recovered and showed no trace of infection except one,
which had lacerated leaves, flowers without pollen, and somewhat eroded
corollas. This plant. No. 18182 (1) is shown at the left in PI. VI, fig. 10
beside a normal plant from the same pedigree. The Q. disease is therefore
not so readily transmitted through the seed in Datura meteloides as in
D. Stramonium, although the former species acquires the disease slowly
through graft infection.

A. F. Blakeslee 27

Most of the species tested seem immune to the infection and unable
to transmit the virus. The Petunia is a typical example. In addition
to controls, eight attempts were made to infect this species by grafting
with Q. Jimson Weeds, but without success. In one case, shown in
PL VI, fig. 11, a Petunia cion was grafted on a Q. Jimson stock (Q.).
A month later a cion from a normal Jimson (N.) was grafted on to the
Petunia, leaving a distance of 16 cm. between the two grafts. Five and
a half months after the last graft was made the Jimson cion at the top
of the plant was still normal in appearance, and had produced two
spiny capsules while the original stock was still Quercina. Twenty-six
seeds from the second of these capsules were planted and produced
20 seedlings, all normal. It may be concluded that the Petunia neither
acquires the disease, nor is able to transmit the virus through as much
as 16 cm. of its stem.

An experiment siruilar to that just described for the Petunia was
tried with the tomato. The virus failed to pass through 3*5 cm. of the
tomato stem and infect a normal Jimson graft above it. It is possible
that with a shorter distance the experiment would have succeeded,
since, in one out of three trials, a cion taken from a graft on a Q. plant
gave the disease to a normal Jimson when grafted on it. In this case,
however, the cion was cut off from very near the junction between the
tomato and the Q. stock, and it is barely possible that some of the tissue
of the latter was the cause of the infection.

Both the two cions of Jerusalem cherry grafted on Q. stock produced
normal flowers with pollen and fruit, but their leaves, although normal
in shape and otherwise like the controls, were more or less blotched with
yellowish patches. Of two grafts from these infected cions on to normal
Jimsons, one transmitted the disease and the other did not.

The egg-plant gave some evidence of being susceptible to infection.
Two cions were grafted on to Q. Jimsons. One grew but slightly and
hence had little opportunity to show the disease. The other cion pro-
duced a vigorous growth, with many flowers with pollen and finally
fruited. The leaves were normal in shape, but were puckered in a manner
shown by the earlier stage of infection in Datura meteloides.

The only instance not already mentioned where a species showed
any evidence of being susceptible to Q. infection through grafting was
with D. Geratocaula. Of two cions grafted on to Q. Jimson stock, one
produced a flower without pollen and also two slightly abnormal buds
which fell off before opening. These latter buds had their corollas
more or less slit. Later flowers from this plant, however, were normal.

28 Graft-Infectious Disease of Datura

Except for the abnormalities of these three flowers in a single plant the
cions were normal in appearance.

In Table VIII are listed the species tested by grafting with Q.
Jimson stock. Controls on normal Jimsons were tried in all cases.

At least one other mosaic-like disease has been found infecting
Jimson Weeds. This has been called " Z." It was first noticed in two
adjacent plants in the field cultures of 1917. PL VI, fig. 12 is a photo-
graph of an infected individual. Infected plants are obviously diseased.
The leaves are light in colour, mottled, more or less eroded, and strongly
puckered resembling somewhat the badly diseased leaves of beans
attacked by mosaic. The leaves may be reduced to merely the midribs.
The capsules are deformed, with spines reduced or entirely absent. The
buds are generally elongated, the flowers " confused " with corollas
often split or otherwise malformed and with numerous accessory carpels
frequently developed. Infection develops rapidly in the plant and soon
renders it valueless for records or for the production of seeds. The disease
was first serious in the field cultures of 1919. From its spread from
centres of infection it was obviously transmitted by means of contact.
Attempts to communicate the disease to normal plants by rubbing them
with infected leaves were successful with some exceptions. No extensive
experiments have been made with this disease to discover what other
Solanaceous species are susceptible. It has not yet been found possible,
however, to transmit it to the tobacco {Nicotiana Tahacum) either by
rubbing the leaves together or by grafting.

From smooth capsules of a Z plant 89 seeds were sown and gave
77 seedlings which remained normal. The experiments indicate that
the Z disease is infectious by contact of leaves but is not carried by
seed.

VI. Early Records of Quercinas.

It has not seemed profitable to search all the early records for de-
scriptions that would indicate plants with Q. infection. Naudin(7)
and Godron (6), as well as Bateson and Saunders (2), undoubtedly had
them under observation. It may be stated that the writer has grown
many thousand plants of the Jimson Weed, including large numbers of
individuals that were heterozygous for inermis capsules, but has never
observed any except Q.'s which showed a mosaic arrangement of the
spines on the fruits, with some valves smooth and others more or less
spiny.

A. F. Blakeslee 29

Naudin (I.e., p. 49) reports in an F^ between an inermis variety
(7). laevis) of D. Stramonium, and the armed type, that while most of the
plants had spiny fruits, others had fruits with reduced spines. Many of
the capsules on three out of the 40 plants in this generation were
very spiny on part of the surface while totally smooth on the rest.
Naudin believed that they united thus by distinct and separate com-
partments the distinctive traits of the parental types {D. laevis and
D. Stramonium). He calls this " liyhridite disjoints " and cites in this
connection the condition in the graft chimaera Cytisus Adami. He
attempted to secure offspring from the smooth part of such a mosaic
capsule, but due to the poor maturity of the seed only four seedlings
were obtained. Of these, one was inerm,is, and may have been an
extracted recessive or a strong Q, In a series of F2 plants from the same
original cross, Naudin found six individuals out of 38 which again
showed more or less well marked his " hyhridite disjoints " and which
were also presumably Q.'s.

Godron (I.e., p. 14) reports finding capsules partly spiny and partly
smooth. He objects to Naudin's interpretation that the separation of the
fruit into smooth and spiny portions is due to their origin from smooth
and spiny parents, since he says that he has found this condition when
both parents had spines.

Bateson and Saunders (I.e., p. 23) after discussing the intermediate
colour of the flowers of the Fi between white and purple flowered forms
say:

" The occurrence of intermediate forms was also occasionally notice-
able in the fruits. Among the large number of capsules examined, there
were some of the mosaic type, in which part of the capsule was prickly
and the remainder smooth, while others, suggesting a blend, were more
or less prickly all over, but the prickles were much reduced in size, and
often formed mere tubercles. These mosaics occurred as rarities both on
prickly individuals and on smooth ones still more rarely."

Further evidence pointing to the Q. nature of these abnormal cap-
sules is given in their following statement :

" Such intermediate fruits were most often found towards the end of
the flowering season."

In a footnote (I.e., p. 24) they call attention to a plant with a single
undersized smooth fruit in the Fi from the cross, purple inermis x white
armed, where onl}'^ armed plants heterozygous for inermis should be
expected. If their suggestion of possible experimental error for the
appearance of this plant be not the correct explanation it may have

30 Graft- Infectious Disease of Datura

been a strong Q. which, both from competition with more vigorous
normal individuals and from the difficulty of securing pollination, might
readily be reduced to the production of but a single capsule. A single
capsule on a normal plant in the field, however, would be rare under
usual cultural conditions.

To one familiar with the manifestations of the Q. disease, it seems
likely that the examples just given above from the literature are in-
stances of infected plants rather than cases of blending or of so-called
" mosaic " inheritance. The term " mosaic," however, may be applied to
the disease, if not to the type of inheritance, since it resembles in many
ways the mosaic of the tobacco, beans, and other forms for which a
causative organism has not been discovered.

VII. Discussion.

Mosaic diseases are not uncommon but their nature is obscure. A
number have been described for the Solanaceae (cf. Allard (1)), and the
mosaic or calico of tobacco is familiar to all who have grown the plant.
The disease of the Jimson described in the present article differs from
the others of the family in that it is carried by both seed and pollen and
appears not to be transmitted artificially by mere contact or inoculation.

That grafting communicates the infection to normal plants relates
the disease to infectious chlorosis of Ahutilon Thompsoni and other
forms investigated by Baur (3). It differs fi'om such instances in that
no vegetative function of the plant is obviously impaired. The strong
Q.'s, which come from infected seed, although smaller than the normals,
are as vigorous as many of the mutants. It also differs from infectious
chlorosis of the mallows in that it is carried by seed.

The profound morphological changes brought about in the leaves
and especially in the flowers and fruit are such that Q. individuals would
be considered worthy of specific if not of generic separation if 100 %
of the seedlings instead of only 79 % came true to the Q. complex. As
the facts stand, however, there is much in the behaviour of Q. plants
which suggests genetic phenomena.

A. F. Blakbslbe 31

VIII. Summary.

1. A form called Quercina was discovered in the Jimson Weed
{Datura Stramonium) occurring spontaneously like a mutation.

2. Quercina plants are distinguished from normals by greater denta-
tion of the leaves, slitting of the corolla, absence of pollen, partial or
entire suppression of spines on the capsules, and certain other characters
associated with less vigorous growth.

3. In a single year's test about 1| 7o of the normal plants in the
field took on the Quercina character by the last of the season. The
Quercina character generally shows itself weakly in a single branch and
gradually spreads to all the new growth.

4. The Quercina complex is transmitted by seed to about 79 "/^ of
its offspring when pollinated from normal plants. The remaining 21 "j^
normal offspring do not produce Quercina seedlings in the next gener-
ation.

5. There is a rough correspondence between the strength of Quercina
character in the parent and the number of Quercina plants in its
offspring.

6. In plants which were becoming Quercina, pollen was obtained
which transmitted the character.

7. Quercina cions grafted on to normal plants of the Jimson Weed
cause the new growth of the stock to take on the appearance of
Quercinas.

8. The cause of the Quercina characters is a disease transmitted by
grafting.

9. Certain other species of the Solanaceae were found to be susceptible
to the disease by gi-afting though to a less degree.

10. It has not been possible to infect plants artificially by rubbing
with diseased leaves nor by inoculation of expressed juice from Quercina
plants.

11. Another disease of the Jimson Weed is briefly described which
is highly infectious by contact.

12. Instances in the literature are cited where the same disease
apparently has been recorded but mistakenly attributed to a vegetative
segregation or " mosaic" inheritance brought about by smooth and spiny
varieties in the ancestry.

32

Gh'aft-Infectiou^ Disease 0/ Datura

TABLE I.

Quercina Jimson Weeds (Datura Stramonium). 1916 Records.

Number of
Origin of Cultures Seed Parents

Total
Offspring

Norma

Quercinas

Percentage
(.Quercina

(1)

Washington, D.C.
Inbred 5 generations :

(a) Purple Armed (2 lines)

13

.596

584

12

2-01

(b) White Armed (1 line)

4

260

258

2

0-77

(2)

Erfurt, Germany.
Inbred 4 generations :

White Inermis (2 lines) ...

6

390

388

2

0-51

(3)

Various Crosses (6 lines) within
(l)and(2)

83

2845

2813

30

1-05

(4)

Staten Island (4 lines)

28

770

748

22

2-86

(5)

Bronx, N.Y. (4 lines)

4

694

689

5

0-72

(6)

Cold Spring Harbor, N.Y. (1 line)

1

246

246

0

0-00

(7)

All lines with Normal ancestry
(15 lines)

139

5801

5726

73

1-26

(8)

Normal Plants from Quercina
Parents (1 line)

51

1781

1764

17

0-95

(9)

Normal Plants (16 lines)
Grand Total

190

7582

7490

90

1-19

(10) Quercina (1 line) crossed with
Normals (5 lines)

34

1348 283 1065

79-01

TABLE IL

Comparison of Seedlings and Adults from Quercina Parents. 1916 Records.

Seedlings

Adults ...
Totals

INos.

I Percentage

I Percentage

Nos.

Normal

273
21-1

Questionable

349
27-0

83-1
227

Q.
16-9
46

N.

12-3
43

Q.

87-7
306

Quercina

671
51-9

N.

0-3
•2

Q.

99-7
669

In Adult stage, 79-0% = Quercina ; 21-0 °/o = Normal.

A. F. Blakeslee

Table hi. offspring from Individual Capsules of Plant No. 1645 (91).

Capsule

Grade of

Number of

Percentage total

Percentage

Number

Capsule

seeds plaotei

germination

N.

Q.

Q.

A 2

4

100

89-0

55

4

6-8

A 3

4

100

40

2

1

33 3

A 4

4

100

260

23

0

0 0

A 6

3

100

60 0

5

45

90-0

A 7

2

100

42-0

0

32

100 0

A 8

4

100

92 0

62

5

7-5

A 10

4

100

89-0

26

27

50-9

.4 11

4

100

95-0

64

2

3 0

A 12

4

100

77-0

52

3

5-5

-1 13

4

100

97-0

57

20

26-0

,J 14

4

100

85-0

60

0

0-0

A 15

4

100

92 0

38

23

37-7

^ 17

4

100

86-0

19

31

62-0

A 18

4

100

90-0

76

1

1-3

J 21

4

2

50-0

0

1

100-0

J 22

0

50

30-0

2

5

71-4

A 24

2

334

63-2

2

122

98-4

^ 25

0

100

43-0

6

19

76 0

A 26

4

100

97-0

69

0

0-0

^27

2

178

86-5

19

139

88-0

.J 28

2

216

25-0

15

28

65-1

^30

2

319

83-6

24

60

71-4

A 31

1

78

75-6

11

23

67-7

J 32

1

102

59-8

12

28

70-0

A 33

1

176

92-0

47

46

49-5

J 34

3

100

75-0

4

24

85-7

.4 35

3

80

50-3

4

6

60-0

A 36

3

157

92-0

60

20

25-0

TABLE IV. Offsp

ring from

Individual Capsules

of Plant iV

0. 1618 {{

Capsule

Grade of

Number of

Percentage total

Percentage

Number

Capsule

seeds planted

germination

N.

Q.

Q-

B 1

2

85

92

8

28

10

26-3

B 2

2

69

91

3

21

9

300

J5 3

0

100

90

0

34

28

53-8

B 4

1

100

85

0

38

22

36-7

B 5

2

69

75

4

12

0

0-0

B 6

1

100

73

0

27

3

100

B 7

2

100

94

0

35

31

470

5 8

1

100

82

0

43

1

2-3

/i 9

0

22

81

8

9

6

400

£10

4

100

97

0

51

0

0-0

Bll

0

25

60

0

10

0

0-0

5 12

1

100

84

0

12

26

68-4

BIB

4

100

95

0

62

5

7-5

5 14

4

64

89

1

44

0

0-0

iil5

4

100

93

0

58

0

0-0

ii 16

1

11

100

0

3

7

70-0

B18

3

100

90

0

26

32

55-2

B19

3

75

93

3

52

1

1-9

£20

2

100

93

0

32

6

15-8

B22

0

100

95

0

31

22

41-5

£23

0

50

92

0

12

24

66-7

£24

0

17

58

8

5

0

0 0

£25

4

100

89

0

37

6

14-0

£26

3

27

92

6

11

11

50-0

£27

2

100

79

0

21

34

61-8

£ 28

2

77

85

7

25

22

46-8

£ 29

4

89

100

0

54

0

0 0

£30

2

63

17

5

6

2

25-0

Journ. of Gen. xi

34

Graft- Infectious Disease of Datura

TABLE V. .Summaries of Tables III and IV.

Number

Number

Average Percentage

Average Percentage

of Capsules

of Seeds Planted

Germination

Quercina

Grade

A

iS

^and JS

A

B ^ and £

A

JS

A and B

A

B

A and B

0

2

6

8

150

314 464

36-5

79-6

58-1

73-7

33-7

53-7

1

3

5

8

356

411 767

75-8

84-8

80-3

62-4

37-4

49-9

2

5

8

13

1147

663 1810

59-9

78-6

69-3

84-6

31-6

58-1

3

4

3

7

437

202 639

69-8

91-9

80-6

65-0

35-7

50-4

0—3

14

22

36

2090

1590 3680

60-4

83-7

72-1

71-4

34-6

53-0

4

14

6

20

1302

553 1855

76-4

93-9

85-2

23-9

3-6

13-7

TABLE VL

Offspring from Individual Valves of C apsides of Plant No. 1645 (91).

Valve A B C D

Capsule

Grade

N.

Q.

Grade

N.

Q.

Grade

N.

Q.

Grade

N.

Q.

A 24

1

2

37

3

2

40

1

3

31

3

2

14

^27

1

1

8

2

2

24

2

3

32

1

7

23

A 28

3

7

5

2

0

9

2

3

9

0

5

5

A 30

2

1

13

0

5

10

2

6

11

2

1

20

A 31

2

5

8

3

3

0

0

2

8

0

1

7

A 32

0

3

13

0

1

5

1

o

6

1

5

4

A 33

4

7

9

0

18

9

0

11

15

0

11

13

A 36

3

17

6

3

14

8

3

16

3

3

13

3

Totals

Grade 0 = 59*9 7oQ- = Grade 1 = 83-9 %Q.: Grade 2 = 85-7 % Q. :
Grade »= 51-6 °/oQ.: Grade 4 = 56-3% Q.

1

TABLE VII. Transmission of Quercina

through

Pollen.

Pedigree

Type

Grade of

Percentage

Number

Parent

of Pollination

Capsule

Normal Quercina

Quercina

16512

Al^Al

Q. X Self

2

0

32

100-00

16624

16122 (4) x^ 7

N.xQ.

4

11

86

88-66

16680

16122 (17) x^ 7

N.xQ.

4

9

61

87-28

16679

16122 (17) X Self

N. X Self

4

89

0

0-00

16511

^11x^11

Q. X Self

4

64

2

3-03

16659

16122 (13) X 4 11

N.xQ.

4

25

0

0-00

16670

16122 (15) X 4 11

N.xQ.

4

14

0

0-00

16510

^ 12 X .4 12

Q. X Self

4

52

3

5-45

16612

16122 (2) X ^ 12

N.xQ.

4

35

0

0-00

16628

16122 (5) X ^ 12

N. X Q.

4

86

6?

6-52

16626

16122 (5) X Self

N. X Self

4

50

0

000

16639

16122 (7) X ^ 12

N.xQ.

4

90

5?

5-26

16636

16122 (7) X Self

N. X Self

4

43

0

0-00

16655

16122 (11) x^ 12

N.xQ.

4

30

31

50-82

16658

16122 (13) x^ 12

N.xQ.

4

46

0

000

16673

16122 (16) X A 12

N.xQ.

4

91

0

0-00

16441

1634 (11) x^ 12

Q.xQ.

—

2

71

97-26

16501

1645 (65) X 4 12

Q.xQ.

—

6

61

91-04

16509

J 15 x^ 15

Q. X Self

4

38

23

37-71

16617

16122 (3) X 4 15

N.xQ.

4

92

0

000

16615

16122 (3) X Self

N. X Self

4

44

0

0-00

16729

16169 (63) x 4 15

N.xQ.

4

24

0

0-00

16742

16169 (69) X 4 15

N.xQ.

4

91

0

0-00

16665

16122 (14) X Ax

N.xQ.

4

2

91

97-85

16662

16122 (14) X Self

N. X Self

4

43

0

0-00

16710

16169 (50) X Ax

N.xQ.

4

0

10

100-00

16709

16169 (50) X Self

N. x Self

4

29

0

0-00

A. F. Blakeslke 35

TABLE VIII.

(irafts
Species Tested with Q. Results

1. Datura Stramonium (Jivason Weed) many Always infections (c£. text)

2. Datura meteloides 3 Infection, but slow in appearing; trans-

' ferable back to Jimson (cf. text)

3. Datura ceratocaula 2 No infection evident, except three slightly

abnormal flowers (cf. text)

4. Brotcallia demissa 5 No infection ; cions fruited

5. Capaicum annuum (Ve^i^ex) ... 2 No infection ; one cion fruited

6. Lycopersicuniesctilentum(Toma,to) 3 No infection; cions fruited; one out of

three grafts back to Jimson caused infec-
tion

7. Nicotiaiia ajfinis 1 No infection

8. Nicotiaiia Tabacum (Tobacco) ... 3 No infection ; cions fruited

9. Petunia hybrida 8 No infection ; cions fruited (cf. text)

10. Solanum ' Pseudo-capsicum (Jeru- 2 Slight infection ; back graft to Jimson

salem Cherry) caused infection in one case out of two

(cf. text)

11. Solanum citrulUfolium 2 No infection ; cions with flowers and spiny

burs

12. Solanum Dulcamara 3 No infection ; cions fruited ; back graft to

Jimsons caused no infections

13. Solanum Melongena (Egg-plant) 2 One cion showed slight infection ; one

cion fruited (cf. text)

14. Solanuvi tuberosum (Potato) ... 2 No infection

EXPLANATION OF PLATES.

PLATE II.

Fig. 1. Seedlings in 3-inch pots; upper row Quercinas, lower row normal controls.
Fig. 2. Later stage of same seedlings shown in Fig. 1 similarly arranged. Lower left con-
trol now seen to be Q.

PLATE III.
Fig. 3. Buds and mature flowers of Jimsons; at extreme right calyx removed ; upper row

Q., lower row normal.
Fig. 4. Dissections of pistils and stamens ; left row Q., right row normal ; figure slightly
enlarged.

PLATE IV.
Fig. 5. Upper row successive leaves from a normal plant ; lower row, the same from

a comparable Q. plant.
Fig. 6. Branch of a field-infected Q. plant showing degrees of spininess from normal armed

to almost entirely smooth.
Fig. 7. (Insert, in Fig. 6.) Capsule with one side smooth and other side spiny from field-
infected Q. plant.

PLATE V.
Fig. 8. Green-stemmed stock with purple Q. cion grafted on left fork and purple normal
cion grafted on right fork. Infection has caused production of smooth capsules on
the normal cion and on the two branches of the stock below the fork.
Fig. 9. Cions of Datura meteloides grafted on to normal Jimson at left and on to Quercina
Jimson at right. D, meteloides cion on Q. stock shows infection in crumpUng of
leaves.

3—2

36 Graft- Infectioiis Disease of Datura

PLATE VI.

Fig. 10. Seedlings from Quercina Datura meteloides ; Q. at left, normal at right.

Fig. 11. Stock of Quercina Jimson supporting a cion of Petunia which bears at its apex

a normal Jimson cion. The Petunia has failed to transmit the infection from the

Q. stock to the normal Jimson cion.
Fig. 12. Jimson Weed attacked by a contact infectious disease (Z) of the mosaic type.

Youngest leaves are reduced to the midrib.

BIBLIOGRAPHY.

1. Allard, H. a. 1916. "A specific mosaic disease in JVicotiana viscosujn distinct

from the mosaic disease of tobacco." Joum. Agr. Research, Vol. vii.
pp. 481—486.

2. Bateson, W. and Saunders, E. R. 1902. Reports to the Evolution Committee of

the Royal Society, Rep. I. pp. 21 — 32.

3. Baur, E. 1906. "Ueber die infektiose Chlorose der Malvaceen." Sitzungsh.

Kgl. preuss. Akad. d. Wissenseh. Bd. i. pp. 11 — ^19.

4. Blakeslee, a. F. and Avery, B. T. 1917. " Adzuki Beans and Jimson Weeds."

Journal of Heredity, Vol. viii. pp. 125 — 131.

h. , 1919. "Mutations in the Jimson Weeds." Journal of Heredity,

Vol. X. pp. 111—120.

6. GoDRON, D. A, 1873. Des hyhrides et des metis de Datura (Nancy), pp. 1 — 75.

7. Naudin, Ch. 1865. " Nouvelles recherches sur I'hybridite dans les vegetaux."

Nouv Arch. Mus. Tom. i. pp. 41 — 54.

8. DE Vries, H. 1900. " Das Spaltungsgesetz der Bastarde." Ber. Dent. Bot.

Gesell. Bd. xviii. pp. 83—90.

JOURNAL OF GENETICS, VOL XI. NO. 1

PLATE

Fig. 1-

Fig. 2.

JOURNAL OF GENETICS, VOL XI. NO. 1

PLATE

I

Fig. 3.

.?=*•

.^^

-^

m

Ui||||lj!l||l|||!l|||l|l||||l||||l||||ll|||lllimillllilllill|llil|llll{lllljtllllllll|ll

.ewcWemsI ' 3l ' 4! ' sr ' el 7I el • 9I

Fig. i-

iiiiii!:

JOURNAL OF GENETICS, VOL. XL NO. 1

PLATE IV

to

JOURNAL OF GENETICS, VOL. XL NO. 1

PLATE V

JOURNAL OF GENETICS, VOL XI. NO. 1

PLATE VI

Fig. 10.

Fig. 11.

Fig. 12.

GENETIC STUDIES IN POULTRY.
III. HEN-FEATHERED COCKS.

Bv R. C. PUNNETT, F.R.S.,
AND THE LATE Major P. G. BAILEY, R.F.A.

(With two Text-figures and Plates VII— XI.)
CONTENTS.

PAGE

Introduction ........... 37

Experimental data 38

Intermediates 41

Colour of Intermediates 44

The hypothesis of complementary factors 48

Castration experiments (performed by Dr F. H. A. Marshall) . . 49

General discussion 51

Note on the sex-proportions and ratio of henny to normal cocks . 54

Tables 55

Explanation of Plates 56

References 57

Introduction.

In certain breeds of poultry, notably in the Sebright Bantam, in
Henny Game, and in certain strains of Campines^, the cock is nor-
mally feathered like the hen. The hackles, saddle feathers, and the
feathers of the tail in such henny cocks are of the form of the cor-
responding feathers in the female, a peculiarity which will be found
illustrated in any of the standard works on poultry. In 1911 we started
to investigate the genetics of hen-feathering in the male bird, making
use of the Silver Sebright Bantam as our hen-feathered breed. It
happened that Prof Morgan in America started some similar experi-
ments in the same year. As he has recently published a sumptuously

1 Heuny cocks may occasionally occur in other breeds. Lamon and Slocum figure a
henny Silver Wyandotte and a henny Silver Spangled Hamburgh ; and it is well known
that they may occur in Pencilled Hamburghs (see Plate IX).

38

Genetic Studies in Poultry

illustrated volume on his work (1919), it is unnecessary for us to give
more than a brief account of our experiments, pointing out wherein
they agree with his^ and wherein they differ.

Experimental data.

Our F^ birds were bred from two Silver Sebright hens and a Gold
Pencilled Hamburgh cock. Of the eight ^i cockerels reared (which of
course were all Silvers), five were fully hen-feathered^ (cf Text-fig. 1),
two were cock-feathered (cf Text-fig. 2), and one was intermediate,
though with a pronounced tendency to the henny type. To the nature
of such intermediates we shall recur later. The Sebright hens were

Text-fig. 1. Henny F^ ^ (ex Sebright x Hamburgh) Text-tig. 2. Normal feathered Fi ^ (ex Sebright x Hamburgh)

not tested separately, but it is clear that one of them, at any rate, must
have been heterozygous. The indication that henny feathering behaves
as a dominant to normal cock feathering was confirmed by the results
obtained in 1912 and in subsequent years. During the period 1912 —
• 1919 we raised 463 male birds from birds heterozygous for henny
feathering mated with birds known not to carry the henny factor.
These 463 cocks fell into two distinct classes, viz. those normally
cock-feathered, and those which were either fully hen-feathered or
intermediate. For reasons that are given below (p. 42) the inter-
mediates are to be regarded as hen-feathered birds. As is shewn in

1 The parents and various progeny from this cross are figured in this Journal, Vol. iv.
1914, PI. 4.

R. C. PUNNBTT AND THE LATE MaJOR P. Gr. BaILEY 39

Table I (p. 55) henny and normal cocks appear in approximately equal
numbers, the total containing 229 henny and 234 normal. We should
mention that although we have frequently bred from normal cocks born
of henny parents, we have never bred a henny bird in this way. Our
experiments suggest therefore that the difference between a henny and
a normal bird is a difference of a single factor throughout.

It will be noticed that several of the earlier matings recorded in
Table I are between F^ birds. In 1912 we put up five pens of F^ birds,
since we wished to raise a considerable F., generation in connection with
the inheritance of weight. Two of these were headed by normal i^i cT cT >
and three by henny F^ </(/• Two of the latter turned out to have been
mated with F^%% {% 152 and $ 380 in Experiments 1 and 5) which
did not transmit the henny character to their sons, for when subse-
quently mated with normal cocks they gave only normal cock-feathered
sons. The remaining F^ henny cock {^ 203) was mated with F^ % 383,
and of the 19 (/•</ reared 16 were henny and 3 were cock-feathered.
This experiment is referred to in Table I as Exp. 43. $ 383 was
subsequently mated to a cock -feathered F^ ^ (398) and proved to be
heterozygous (Exp. 25). The mating between ^ 203 and % 383 was
a mating between two heterozygotes, and the proportions of henny and
normal </•(/ resulting was not far removed from the expected 3 : 1 ratio.

Of the birds bred in Exp. 43 three henny c/c/ and eight pullets
were tested in order to ascertain whether any of them were homozygous
for the henny factor. Two of the henny (/</ were mated with normal
% % (Exps. 10 and 11) and both proved to be heterozygous. The third,
mated with a sister ( % 668), produced six henny and two normal sons^
All three of the cocks tested proved to be heterozygous. Of the eight
pullets, two, mated to heterozygous cocks, gave both hennies and normals;
three, mated with normal cocks, gave both hennies and normals (Exps. 27,
28, 29) ; while the remaining three, mated to normal cocks, gave 9, 6,
and 11 normals respectively. None of the eight therefore could have
been homozygous. On ordinary expectation one of the three cocks and
two of the eight pullets should have been homozygous for the henny
factor. We failed however to find such a bird among the eleven that
were tested.

Later on we made another attempt to find a homozygous bird. A
henny ^ (24/15) from Exp. 31 was mated with a sister ( % 95/15), who
was subsequently proved to be heterozygous in Exp. 40. This mating is

1 This experiment is not included in the table as the hen used may have been hetero-
zygous.

40 Genetic Studies in Poultry

referred to in Table I as Exp. 44. All of the five cockerels reared turned
out to be henny, while all of the five pullets tested proved to be hetero-
zygous. One of these henny cockerels was mated with a sister ( $ 98/16)

c? 24/15 X ? 95/15 (cf. Exp. 40)
(henny) I

cJl/16 X 2 98/16 $210/16 9235/16 $294/16 $388/16 4^?^?

(henny) ] cf. Exp. 39) (cf. Exp. 34) (cf. Exp. 25) (cf. Exp. 37) (cf. Exp. 36) (henny)

(J 90/17 (J 321/17 $ 150/17 $ 320/17 ? 322/17 2 .? <J

(henny) (cf. Exp. 22) (cf. Exp. 41) (cf. Exp. 45) (cf. Exp. 42) (henny)

who was shewn to be heterozygous in Exp. 39. From this mating all of
the four males produced were henny. Two of these were subsequently
mated to normal hens. One of them {^^ 321) proved to be heterozygous.
The other (</ 90) produced two henny cocks from a Brown Leghorn hen
in 1918, but in 1919 he failed to get any chickens. Mated with a different
hen he has given some chicks in the present year, but as yet they are
too young to be sure whether they are all henny or not\ All of his
three sisters proved to be heterozygous. It is possible that (/ 1/16 was
homozygous, but as four of his five offspring tested were certainly hetero-
zygous and the remaining one so far doubtful, it is more likely that
he was heterozygous. We attempted to breed from him again in 1918,
but although he was apparently a vigorous bird he failed to fertilise
an egg. He was killed on April 26, 1918, and histological examination
shewed that his spermatogenesis was abnormal. The matter is being
investigated further as a good deal of sterility has occurred in this
inbred strain.

So far then the search for a homozygous henny bird has not been
successful. Of 15 birds produced by mating heterozygotes together,
and shewn to transmit henny feathering, 14 were definitely proved
to be heterozygous, the remaining one at present being doubtful.
Normally we should have expected five homozygotes among the 15.
This may of course be a chance result, but there is the possibility of
the homozygous bird being non-viable in the strains that have arisen
in the course of our experiments. Further work to decide this point is
now in progress.

1 [Note added November 9, 1920.] Of the four cockerels reared this year all have coine
henny. Up till now therefore all of the six sons of ^ 90 have proved henny, and it is not
unlikely that he may be homozygous. It is hoped to test him further next year.

R. C. PUNNETT AND THE LATE MaJOR P. G. BaILEY 41

Iiiter'7nediates.

We have already stated that birds of an intermediate type of
feathering, as well as purely hen-feathered birds, have arisen during
our experiments ; and indeed the intermediates have greatly out-
numbered the typically hen-feathered ones. Though we are not at
present in a position to offer a satisfactory explanation of the nature
of these intermediates, we may record certain observations which we
have made in connection with them.

(1) The intermediate condition shews a great deal of variation.
Some intermediates differ but slightly from a purely henny bird in
appearance. The majority of the saddle feathers are typically henny,
but scattered among them are a few in which the tip of the feather
exhibits the lack of barbules and the orange colour' characteristic of
the typical male feather. Where the general colour of the feather is
dark, the orange fringe at the tip is very conspicuous. Three feathers
from such a bird are shewn on PI. VII, fig. 3, of which two are inter-
mediate in type and the other a purely henny one.

In other cases the intermediate approximates far more closely to
the condition found in the normal cock. Sometimes indeed the approxi
mation is so close that such birds might easily be classified as normal
feathered unless the observer were on his guard". Careful search however
has always revealed the presence of a few henny or nearly henny feathers
among the young feathers coming through. Moreover, as will be ex-
plained below, any doubt as to the nature of the bird is always resolved
at the first moult. Where we have had any doubts as to the true nature
of the feathering, we have kept the bird until it was 18 — 20 months
old. A good example of this type of intermediate feathering is figured
on PI. VII, fig. 1. In their gold colour and general appearance, they might
pass for feathers from a typical male. Between these two extremes all
sorts of grades of intermediate feathers are to be found, such for example
as those figured on PI. VIII, figs. 1 and 3. From the almost henny feather
to that which resembles a normal cock feather, a practically continuous
series might be formed from a small number of intermediate birds. For
the feathers on the same bird are not all of precisely the same grade.

1 This of course when the bird is genetically a gold. The tip is white, or nearly so,
when it is genetically a silver.

- When such birds first made their appearance in a certain pen during the early days
of the work we recorded some as normal ^ ^ . However we discovered our mistake, and
the results from this pen are not included in the present account.

42 Genetic Studies in Poultry

(2) Nevertheless all of these grades of intermediates behave alike
in one respect. At the first moult they all become fully henny, or nearly
so. We have now kept some dozens of intermediate birds over their first
moult without meeting any exception to this rule. In some cases the
change to henny feathering is complete. As a typical example we may
take the case of ^ 105/18 (PI. VII, fig. 1). A full henny cock {^ 90/17 in
Exp. 44) was mated in 1918 to a Brown Leghorn hen. Only two male
chicks were reared, viz. j/ 104/18 and c^ 105/18. These two birds grew
up so closely alike that they were indistinguishable in appearance. They
were clear golds with traces of brown occurring in some of the quills and
in the feathers at the base of the tail. Both were so like normal males
in appearance that one of them, cT 104, was kept, and has now moulted
twice. In 1919 all of his new feathers were purely henny, as shewn
on PI. VII, fig. 2\ At this period however he was not full henny in
appearance because some of the old intermediate feathers were not shed
until the following year. By the second moult these worn feathers,
which had survived from the original plumage, were all shed, and the
bird is now fully henny. Noteworthy is the considerable development
of melanic pigment in the saddle feathers which has accompanied the
structural change.

Another good example of the change from intermediate to henny
feathering at the moult is illustrated on PI. X. Two brothers of inter-
mediate plumage were closely similar, both in colour and in the nature
of their feathering. One of them (Fig. 1) was killed at nine months old ;
the other was kept for a year longer, until he was over his first moult
(Fig. 2). This case differs from the preceding one in that the plumage
did not become fully henny. Some of the feathers shew traces of the
intermediate condition, but the change as a whole is sufficiently striking.
Genetically these two birds were silvers.

(3) Although intermediate birds always moult out henny, in so far
as our experience goes, they may sometimes revert to the intermediate
type at some later moult. In illustration of this we may take the history
of (^ 201/14, who was bred in Exp. 28. In his first plumage he was
intermediate, with a marked tendency towards normal cock feathering.
The sample of his feathers taken in 1914 is unfortunately lost, but they
were very like those of {/ 126/15 (from Exp. 15) shewn on PI. VIII, fig. 1.

1 The bird recorded by Darwin (A. and P., i. p. 258) "which, after assumiug its
perfect masculine plumage, became hen-feathered in the autumn of the following year "
was doubtless an intermediate with a pronounced tendency to normal feathering in its
first plumage.

R. C. PUNNETT AND THE LATE MaJOR P. G. BaILEY 43

In 1915 he moulted out full henny, as shewn by the feathers figured on
PI. VIII, fig. 2. The colour of the plumage was dark mossed all over.
There was no trace of any intermediate feathering, and he remained
a fully henny bird until his moult in 1918, when he put up a number
of intermediate feathers (cf PI. VIII, fig. 3). Scattered among these were
a number of feathers of the henny type, but the general appearance of
the bird at this stage was distinctly cock-like. At his moult in 1919
he again became purely henny (cf. PI. VIII, fig. 4), with the exception
of five worn intermediate feathers which doubtless belonged to the
previous set\ In the spring of 1920 he unfortunately went light, and
died at the age of six years. It should be added that we bred from
him in 1915 (Exp. 14), selecting him as an intermediate with a pro-
nounced tendency to normal feathering. He proved a good getter, as
again he did in 1918 and 1919 when tested for fertility. There are no
grounds in this case for supposing that the assumption of intermediate
feathering is connected with sterility.

We have so far no other case of the reversion to intermediate
feathering after a lapse of several years, but this is probably because,
owing to lack of space, it is only in exceptional cases that we have been
able to keep birds over a period of four or more years. In the three
other cases where we kept a henny cock until its fourth year, there was
no appearance of any intermediate feathering. In each of these cases
it is true that the bird started as a fully hen-feathered bird, and not as
an intermediate. Nevertheless a bird may start as an almost henny
bird and later revert to a more cock-like type, as is shewn by the
history of (^ 338/17 (ex Exp. 19). Nearly henny in his first plumage
he became fully so in 1918. In 1919 however he put up so large a
proportion of long intermediate feathers that his general appearance
was towards that of a normal cock, particularly when viewed at a
distance of 20 yards or so away. The tail sickles were also well de-
veloped, as is usual in the case of intermediates with a tendency to
normal feathering. In his present moult (1920) all of his new feathers
are so far henny, but many belonging to the older series are not yet
shed. We shall endeavour to keep this bird, and several others, as long
as possible in order to find out what the condition of the feathering is
in old age.

(4) With regard to the relation between intermediate birds and
those that are fully hen-feathered from the first, we have not yet

^ The history of this bird recalls that of the polecat hen-cock recorded by Darwin
{A. and P., i. p, 253).

44 Genetic Studies in Poultry

obtained satisfactory evidence. It is however clear that it is not due
to the factors that influence the length of the hackles in normal breeds.
Intermediates with a marked tendency to normal cock feathering can
be bred from short-feathered races such as A seels and Game Bantams,
as well as from breeds relatively long in the feather such as the Brown
Leghorn. Further, the various gi-ades of intermediates together with
full henny birds can be bred from the same mating, where all of them
are heterozygous for the henny factor. Whether the homozygous cock
can be an intermediate we cannot at present say, for, as we have already
stated above, we have not yet succeeded in identifying such a bird.

Colour of Intertnediates.

In our account we have restricted the term " intermediate " to birds
differing from the normal cock in the structure of those feathers in
which the sexes of normal breeds differ from one another. The structural
difference is due to the elongation of the rachis and distal barbs, and to
the loss of the barbules on the barbs. Proximally, of course, even in a
truly male saddle-hackle, barbules are present and are furnished with
barbicels.

When an intermediate of a pronounced cock-like type becomes henny,
a marked difference in the general colour of the bird results, as well as
in the form of the feathers. The principal factor in this change is the
deposition of melanic pigment in the feather. It is well exemplified by
the feathers shewn on PI. VII, figs. 1 and 2. The deposition of the melanic
pigment on the assumption of the henny plumage brings about a very
great difference in the general appearance, and there arises the in-
teresting question whether the henny cock assumes the colour as well
as the plumage structure of the hen that corresponds genetically to
him. That the colour correspondence as well as the structural one may
in some cases be very close is shewn by Morgan's experiments in the
castration of Sebright Bantams (cf Morgan, 1919, PI. I, figs. 3 and 4).
If we assume that the castrated bird represents the appearance that the
cock of this breed would have if the henny factor were not present in
him, then we must suppose that the " hennification " of such a male
would result in a cock which in colour, as well as in plumage structure,
was practically identical with the hen. There are many breeds in which
the cock differs markedly from the hen in colour in those feathers only
in which there is a structural difference in the sexes. This is naturally
true of self-coloured breeds such as blacks, buffs, and whites, as well as

R. C. PUNNBTT AND THE LATE MaJOR P. G. BaILEY 45

of blues. It is true also of many breeds where the feathers exhibit a
definite pattern such as most of the barred, pencilled ^ laced, and
spangled breeds. In the last case it is true whether the spangled
feather is predominantly light with a black tip, as in Hamburghs, or
predominantly black with a light tip, as in Anconas. It is true also of
Brown-breasted Game. On the other hand there are breeds in which
the colour differences between the sexes are not limited to the sexual
feathers. This is the case in Black-reds and in Silver-greys, where
we find a marked difference between the black breast of the cock and
the salmon-coloured one of the hen. The difference also occurs in the
abdominal and leg feathers, which are brown mossed in the hen'^ but
black in the cock. In Piles also we find the breast colour difference,
the salmon breast of the hen contrasting with the white one of the
cock. A sexual colour difference, apart from the sexual feathers, is found
also in certain of the laced and pencilled breeds, such as Indian Game,
Partridge Cochins, and Partridge Wyandottes. In the case of these
breeds, it is important to remember that the cock is either a black-red
or a duckwing.

We are inclined to think that there is an essential difference between
the breeds of the first group where the colour difference is restricted
to the sexual feathers, and those of the second where the sexes exhibit
a colour difference apart from the sexual feathers, and to suppose that
the sexual colour differences are present in the second group, but are
obscured by some dominant factor. The obscuring factor we consider
to be a different one in different cases. In whites it is the inhibitor for
colour that occurs in the dominant whites, such as Leghorns^ ; in blacks
and blues, it is due to the factor for black ; in brown breasted, there is
also some factor leading to increased melanism ; in buffs and light golds
there is probably a factor which inhibits melanic pigmentation.

In the absence of such obscuring factors, we have the condition
found in the second group with its marked sexual dimorphism in colour.

In the first group it is clear that " hennification " would lead to the
production of males which in colour, as well as in feather structure,
resemble the hens. Provided that the cock is fully henny, he will be
coloured like the hen. Now, as we have already pointed out, the hetero-

1 The amount of melanic pigment is throughout much reduced in the cocks of pencilled
breeds, but such as exists exhibits the same arrangement as in the hen.

2 In writing this we had in mind the Brown Leghorn and the Silver-grey. In wheaten
hens the Salmon tint is found also over the abdomen and thighs.

3 In recessive whites of course the obscuration is brought about through the absence of
some factor essential to pigment production.

46 Genetic Studies in Poultry

zygous bird may be fully henny, and among the fully henny heterozygous
birds that we have bred we have had some which in colour, as well as in
feather structure, closely matched the corresponding hen. The hetero-
zygous Silver-Pencilled Hamburgh, figured on PI. IX, offers a good
example of this. Again, among the fully henny F2 heterozygous birds
that we bred, ex Sebright x Hamburgh, we had examples of silver-laced
and of gold-pencilled birds which were henny in colour and pattern as
well as in feather. When we introduced the Brown Leghorn we obtained
some dark mossed cocks which corresponded exactly in colour to hens
similarly bred. The feathers of these birds were very like those of
^ 201/14, figured on PI. VIII, figs. 2 and 4. From their dark eyes and
the purplish tinge in the face we have little doubt that, as compared
with a Black-red, they contained some obscuring factor operating for
the increased production of melanic pigment. This is further borne
out by the down colour of such birds, which was dark blackish-brown,
a colour known to be dominant to brown stripe. The henny Silver
Wyandotte figured by Lamon and Slocum (p. 105) is evidently like
a hen in colour and pattern, and as these birds are uncommon it is
extremely likely that he was heterozygous. We are inclined therefore
to consider that in our first group the heterozygous bird, if fully henny
in feather structure, resembles the hen also in colour.

We doubt however whether this applies to the birds of our second
group, where sexual differences occur in colour apart from the sexual
feathers. In order to try and throw light on this point we attempted to
" hennify " the Brown Leghorn by repeated crosses of heterozygous cocks
with Brown Leghorn hens. The nature of the experiment is summarised
in the accompanying scheme :

n Leghorn 9 j

Brown Leghorn 9

Brown Leghorn J 1 ^ ^ 240/13, 309/13

(J 286/14
(henny)

[Exp. 16] 1- ^ 242/16

9 Brown Leghorn ) (henny)

[Exp. 19] f ^ 338/17

9 Brown Leghorn I (henny)

In this series of matings our procedure was to choose fi:-om our
available birds the cock that was nearest to a black-red, and mate each
time with a Brown Leghorn. (^ 286 x Brown Leghorn % gave chicks
with two types of down, viz. dark black-brown and brown stripe in
almost equal numbers. Of the brown-striped </ 242 resembled a Brown
Leghorn in juvenile plumage, subsequently becoming a henny. The

R. C. PUNNETT AND THE LATE MaJOR P. G. BaILEY 47

general colour however was much redder than that of a Brown Leg-
horn $ , while he had a black breast and the characteristic chestnut
wing-bow of the Brown Leghorn cock. Mated with a Brown Leghorn
hen (Exp. 19) he gave 17 chicks, all brown-striped. They proved deli-
cate, however, and only five {^ ^ ^ and 1 ?) were reared to maturity.
Of the 4 (/(/, two were exactly like normal Brown Leghorns; one was
like a Brown Leghorn, except that the breast feathers were mixed
salmon and black instead of full black; and the last, a henny bird
(</ 338), was very like the father. The solitary hen was indistinguish-
able in appearance from a Brown Leghorn. She proved to carry the
henny factor (Exp. 38), and was subsequently mated with ^ 338 in
the hope of establishing a homozygous henny Brown Leghorn strain.
Unfortunately the union proved sterile. But although both ^ 242 and
^ 338 were, apart from the henny factor, almost certainly genetically
similar to the Brown Leghorn, and although, at some period, both were
fully henny birds, nevertheless their colour was very different from that
of the Brown Leghorn hen. The most noticeable differences were (a) the
retention of the chestnut-brown wing-bow, (6) the presence of black or
black and salmon feathers on the breast, and (c) the presence in the
saddle of warm chestnut-brown feathers, as well as of others shewing
some beetling, interspersed among the brown mossed hen-like feathers.

We are therefore inclined to think that in the breeds belonging to
our second group, where sexual colour differences exist apart from the
sex feathers, the heterozygous bird may be fully henny in plumage
structure, but not in colour. Whether the plumage would be fully
hen-like in colour in a homozygous henny cock is a point we have been
unable hitherto to decide. But there is evidence for supposing this to
be the case. In the picture of light red hennies given in Cassell's Illus-
trated Book of Poultry (1876), on Plate 25, the colour of the two sexes
is identical, the cock having a deep salmon breast and saddle feathers
coloured exactly like the hen. We are also indebted to Mr Herbert
Atkinson, the well-known authority on Game Fowl, for the information
that the true breeding pile henny cock is coloured like the hen.

On the whole therefore the available evidence suggests that the
henny factor may bring about a complete change to the henny condition
in the heterozygous bird both in the structure of the feathers and in
the colour of the sexual feathers, but, where the sexes differ in colour in
other than the sexual feathers, the bird must be homozygous before the
colour change is complete.

48 Genetic Studies in Poultry

The hypothesis of complementary factors.

In the account of his experiments on the nature of henny cocks,
Morgan suggests that the Sebright differs from the normal (in this
case, Black-red Game Bantams) in two factors. The presence of both of
these factors is assumed to be necessary for the development of henny
feathering. This interpretation rests almost entirely upon the pro-
portions of henny and normal cocks appearing in the F2 generation,
and Morgan points out that the experimental numbers, viz. 31 henny
to 28 normal, are far removed from a 3 : 1 ratio but approximate closely
to a 9 : 7 ratio. A few birds were bred by crossing F^ with game, and" of
these two are recorded as henny and seven as cock-feathered. This again
is close to expectation on the hypothesis of two complementary factors.

We venture to suggest a possible interpretation of Morgan's results
which would bring them into harmony with our own, viz. that some of
the birds recorded by him as cock-feathered were in reality intermediate
birds of the type that shews a marked approximation to normal cock
feathering. If a dozen of his 28 F^^ ^ , recorded as normals, were of this
type, we should have a close approximation to a 3:1 ratio. We have
already pointed out how easily one may classify these cock-like inter-
mediates as of normal plumage, unless one is on one's guard. The
distinction was forced upon us by the change to henny plumage that
occurred at the moult of such birds. Morgan makes no remark about
changes in the character of the feathering at moult, and one is led to
suppose that his birds were generally destroyed before this stage was
reached. We consider that our view is indirectly supported by the state-
ment that Morgan makes with regard to intermediates. He recognises
such birds among his Fg cTc/, but states that "the line between inter-
mediate and cock feathering is sharp, all the intermediates belonging
distinctly to the hen-feathered group, but the line between the two
subdivisions of hen-feathered birds is not sharp, and occasionally a bird
is found that is difficult to place " (p. 15). Unfortunately we are given
no illustrations of the feathers of these intermediate birds, but it is
evident from the account that they approximated to the full henny
condition, and perhaps resembled such birds as the one whose feathers
we have figured on PI. VII, fig. 3.

We may add that we have examined our own data carefully on
the supposition that two complementary factors might be concerned
with henny feathering, but find no grounds for supposing such to be
the case. In our experiments, which involve a far greater number of

R. C. PUNNETT AND THE LATE MaJOR P. G. BaILEY 49

matings and of birds than Morgan's, the hypothesis of a single factor is
so clearly adequate that we have ventured on the above criticism in an
attempt to reconcile his results with our own.

Castration Experiments.

Among the most interesting features of Morgan's work was the
demonstration that castration of the henny cock led to the assumption
of normal male plumage at the next moult. Thanks to the co-operation of
Dr F. H. A. Marshall we are able to confirm Morgan's results. As an
example we may give the history of (/ 204/15 (ex Exp. 14). In plumage
structure this bird was nearly full henny, a few of the feathers being
slightly intermediate, of the grade shewn on PI. VII, fig. 3. In general
colour he was a warm chestnut, due to the presence of a number of
feathers of this colour mingled with the typical hen-like brown mossed
feathers. He was very like the henny cock figured on PI. XI, fig. 1, and
we have little doubt that if he had been allowed to moult in the normal
way he would have become a full henny bird with a preponderance of
hen-like brown mossed feathers on the back and saddle. In June 1916
the testes were removed by Dr Marshall. On Oct. 1 following he is
recorded as throwing up normal male feathers. Subsequently he became
feathered like a Brown Leghorn cock. He was kept over another moult,
and eventually killed for a skin in Nov. 1917. He was, as shewn on
PL XI, fig. 2, very like a Brown Leghorn cock both in plumage structure
and colour. We may add that although he would flirt with hens intro-
duced into his pen, he made no attempt to tread them. After death an
autopsy made by Dr Marshall revealed no traces of testes.

The birds placed at Dr Marshall's disposal for castration were all
adult or nearly so. For we had to make use of material from families
containing both hennies and normals, and until a bird is about three
parts grown it is often difficult to distinguish the intermediate fi-om the
normal, and to be sure that it belongs to the henny class. In the first
batch of experiments the mortality was high, and Dr Marshall suggested
that this might be avoided by removing one testis at a time, allowing
the bird to recover in between. Accordingly three brothers (cTc/ 78/15,
104/15, and 276/15 from Exp. 16) were operated on between Oct. 28
and Nov. 7, 1916. In each case the right testis was removed. All of
these birds were intermediates in their first plumage, but had already
assumed the full henny type at the date of the operation, being at that
date more than 18 months old. In each case an excellent recovery was

Journ. of Gen. xi 4

50 Genetic Studies in Poultry

made, and when examined in Feb. 1917 it was found that two of them,
104 and 276, were putting up some feathers of the cock-like inter-
mediate type. Further, these feathers were ahnost entirely confined to
the right side of the body, i.e. the side on which the operation took
place. The third bird, (/* 78, remained henny. The left testis in each
iDird rernained intact throughout. Of the two birds which put up inter-
mediate feathers, one ((/ 276) was killed in July 1917 in order to
preserve the skin as a record. The other was kept over another moult,
and, having become fully henny again, was killed on March 3, 1918.
The explanation of this peculiar phenomenon is obscure. It is clear
that a local lesion may influence the character of the feathering on
the side on which it occurs, although the hormone must be supposed
to be circulating equally on both sides. It seemed conceivable that
injury to the sympathetic nervous system rnay have been concerned,
for it is so closely in contact with the genital gland that it is impossible
to remove the latter without nervous damage. We think however that,
in the light of some subsequent experiments by Dr Marshall, another
explanation is possible.

Dr Marshall removed the right testis in three birds hatched in 1919,
the operation being performed in November. Two of these birds (t/ 337
and (^ 400) were almost fully henny in their first plumage, and remained
so after their moult in 1920. In neither case did they put up any
intermediate feathers on the side on which the operation took place.
The third bird ((/ 346) was intermediate in the character of the
feathering. In March 1920 he shewed a well-marked patch of full
henny laced feathers on the fore part of the right saddle, i.e. in the
position in which most of the intermediate feathers appeared in
cf 104 and ^^ 276. The relatively sharp lacing and paler ground of
these henny feathers made the patch shew up conspicuously against
the more warmly tinted intermediate feathers. Later on in the year
this bird moulted out practically full henny, the feathers on both sides
being similar to those found on the patch that appeared earlier on the
right side. Whatever the reason, the operation seems to have caused
an earlier appearance on the right side of feathers which would normally
have developed at the next moult. We know that henny birds may
subsequently develop an intermediate type of feathering (p. 42). It
seems just possible that cT 276 may have been such a bird, and that
the operation led to a premature development of intermediate feathers
because he would have developed such feathers at his next moult had
he not been killed. On the other hand, (/* 104 was allowed to live and

R. C. PUNNETT AND THE LATE MaJOR P. G. BaILEY 51

moulted out henny. In his case however the development of inter-
mediate feathers was much less marked, and the tendency to develop
them may have been overcome by the time the moulting season arrived.
The whole matter remains obscure. It is clear that much more experi-
mental work must be undertaken before we can hope to understand it.

General Discussion.

Granted then that hen-feathering behaves in heredity as though
it 'were dependent upon a single factor ; further, that its manifestation
depends upon the presence of the gonad, we may ask whether we can
associate the factor with any particular tissue. This Morgan claims to
have done, regarding the so-called luteal cells of the ovary as the
seat of a hormone which brings about the development of feathers
belonging to the hen type. He states that these luteal cells are also
found in the testis of henny, though not in those of normal cocks. If
this is so, it is plausible to identify the henny hormone with a specific
tissue. The identification is supported by Goodale's work on the Brown
Leghorn hen, as well as by the work of Pezard, castration in either case
leading to the assumption of the normal cock plumage. It seems clear
that henny feathering, in either sex, is due to a hormone produced in
the gonad : and that something necessary for the production of this
hormone can be represented by a factor in gametogenesis. So far as we
can see, there is no reason for regarding a hormone that leads to hen-
feathering in the female as different from one leading to hen-feathering
in the male. Each produces a similar effect, and the effect of castration
is in each case similar. We are led therefore to infer that the factor
transmitted is identical in the two cases. If this is so, it is clear that
the precise mode of transmission in the two cases is different.

In normal breeds we must suppose the transmission to be sex-linked.
The female we may regard as heterozygous for the henny factor as well
as for sex. To her daughters she transmits both the factor for female-
ness and that for hen-feathering : to her sons, neither. In other words
there is complete linkage between these two flxctors in the gameto-
genesis of the normal hen ; and if we had no knowledge of henny cocks,
we might have considered the two characters as due to one and the
same factor. There is however a point of difference between this and
the other sex-linked cases hitherto described in poultry. In all of these
the heterozygous hen transmits the recessive character to her daughters,
the dominant to her sons. We may recall the barring of the Plymouth
Rock, silver plumage as opposed to gold, the inhibitor for the pigmen-

4—2

52 Genetic Studies in Poultry

tation of the silky fowl, as well as that for shank pigmentation, etc. In the
present case however it is the dominant character that is transmitted
to her daughters by the heterozygous hen, and the recessive to her sons.
When we come to the purely henny breeds we might at first sight
suppose the factor to be carried by both sex-chromosomes, in the male
as well as in female. We are, however, at once met with the difficulty
that the heterozygous hen, produced by mating a hen of a normal breed
with a henny cock, transmits the character to all of her daughters but
to only half of her sons. Such a result challenges comparison with the
cases of sex-linked heredity in man where the female carrier, mated with
a normal, transmits to half of her sons. The resemblance however is but
superficial : in the human case the sexually homozygous female transmits
to half of the heterozygous sex, while in poultry it is the sexually hetero-
zygous female that transmits the peculiarity to half of the homozygous sex.
We feel that the simplest way of summarizing the facts is to suppose
that in all poultry the factor associated with the sex-linked mechanism
comes into play ; whereas in families in which henny cocks appear, an
independent, though apparently identical factor also comes into operation,
and that in heredity this factor behaves in the usual way, uncomplicated
by sex linkage. The hen of a pure henny strain is, on this view, homo-
zygous for A, which can be transmitted to either sex. At the same
time she is also heterozygous for the factor A' which she transmits only
to her daughters. Again, a hen that arises from mating a pure henny
cock with a normal hen is heterozygous for both A and A'. The latter
she transmits to her daughters only ; the former to half of her sons and
half of her daughters. Half of her sons will be henny. But since all of
her daughters receive A' through the sex-linked mechanism, those which
receive A will not be visibly distinct from those which do not. The
former, however, differ from the latter in transmitting henny feathering
to half of their sons. A and A' we regard as producing the same effect,
but difi'ering in the mechanism of their transmission. Translated into
terms of chromosomes, this amounts to saying that A' is only found
in the sex chromosome peculiar to the female S whereas A may occur in
some pair of chromosomes which is equally represented in both sexes.
Whether the sex-linked connection preceded the ordinary one in evo-
lutionary history, or the reverse ; whether the double connection may
be regarded as more primitive than either single one ; and how the
dislocation, if there be one, was brought about, are questions to which
we can offer no answer.

^ Assuming, of course, that there is such a thing.

R. C. PlTNNETT AND THE LATE MaJOR P. G. BaILEY 53

We may however recall that certain human defects present analogous
features in their heredity. Night-blindness may in some families be
sex-linked (cf. Newman, 1913), while in others it behaves as an ordinary
dominant (cf Nettleship, 1907). And from the data collected by Nettle-
ship (1909) on Leber's disease, it would appear that both forms of
transmission may occur here also. It seems not unlikely that some
of the complications found in such pedigrees may be due to the dual
mechanism involved in the transmission of the same factor. There is
however a remarkable feature about night-blindness, in that it behaves
as a dominant when transmitted equally to either sex ; whereas, when
associated with the sex-linked mechanism, its behaviour is that of a
recessive. Apart from this peculiarity, and having regard only to the
sex-linked cases, there is a further point of difference between the
human being and the fowl. From analogy with Drosophila we are led
to infer that the sex-chromosome peculiar to the heterozygous sex
cannot carry a sex-linked factor. But in the poultry case it is difficult
to see how we can avoid supposing that the factor for henny plumage
in normal breeds is carried by the sex-chromosome peculiar to the
heterozygous sex, in this case the female. Though the human cases
resemble the poultry case in shewing both sex-linked and ordinary
inheritance, there are clearly difficulties in the way of instituting a
closer parallel between them.

In suggesting the view of henny-feathering outlined above, we have
not overlooked the possibility of an interpretation on the lines so ably
developed by Goldschmidt for the intersexes in his Lymantria crosses.
Such an interpretation, however, implies a different result in the reci-
procal crosses between normal and henny breeds ; and this, according to
Morgan, does not obtain. Moreover, the only type of intersex produced
would be the henny cock, in which comb, wattles, spurs, and sexual
behaviour are all as in normal males. It seems unlikely that this single
type^ can represent both male and female intersexes. Were we to sup-
pose that certain henny cocks were genetically females, i.e. in so far as
the sex-chromosomes are concerned, we should look for a considerable
excess of females in their progeny from normal hens. Our results from
a number of such matings, however, shew that the sexes are produced
in nearly equal numbers, i.e. 324 ^ ^^ and 345 $ ? (cf Table I). We
incline to agree with Goldschmidt when, in his discussion of henny
feathering, he says "Wir mtissen gestehen, dass im Augenblick die

1 Single, because a given cock may exhibit the extremes at diSerent times in his life
history.

54 Genetic Studies in Poultry

Schwierigkeiten, die diese Tatsachen eincr einfachen Losung cntgegen-
setzen, noch nicht bchoben sind, und dass noch manche Experimente
notig sind."

Note on the sex-projiortions and ratio of lienny to normal cocks.

Bearing in mind the irregular sex -proportions that frequently occur
where intersexes are found in animals, we thought it worth while to
examine our figures from this point of view. In attempting to ascertain
whether families with a marked preponderance of either sex shewed a
preponderance among the cocks of either normal or henny feathered,
we have divided our families into three gi'oups, viz. (a) those with
marked excess of $ $, (b) those with marked excess of (/</, and (c) those
in which the proportions of the sexes approach equality (cf Table II,
p. 56). Our standard of "marked excess" must naturally be an arbitrary
one, and in practice we have classified families in which three-fifths or
more of the total belong to a given sex as shewing a marked excess of
that sex. Thus a family with 20 (/</ and 11 $ $ is regarded as shewing
a marked excess of cTc/" J a family with 7 cfc/ ^^'^^ 12 $ $ as shewing a
marked excess of $ $ ; while a family with 12 </£/• and 9 $ $ is classified
as one shewing approximate equality. The classification is of course
quite rough, and we do not suggest that any special significance should
be attached to the resultant groups. In classifying the families, we have
kept those in which the henny character came from the mother distinct
from those in which it was transmitted by the father. We should add
that the sexes were in most cases only recorded in the live birds as they
gi'ew up. In relatively few cases was the sex determined by dissection
of a chick that had died early. The results of our grouping are given in
Table II, p. 56. It will be noticed that, where the henny character is
transmitted by the father, the proportion of normals and hennies among
the sons is nearly equal in each of the three groups. This is also true
for cases where it is transmitted by the mother, in the group where the
sexes are approximately equal, and in that in which there is an excess
of females. But where there is an excess of males in the offspring, the
normal sons are nearly twice as numerous as the hennies. Moreover
this excess of normals is found in all of the seven families concerned
except one (Exp. 41), in which the numbers are very small. The figures
as they stand suggest something more than chance distribution, and
further experiments have been planned to test the point.

R. C. PUNNETT AND THE LATE MaJOR P. Gr. BaILEY 55

TABLE I. Ileterozygote x Recessive.

A.

Cock hcnnij.

Details of MatiiiR

No. of Exp.

Progeny

_ 1

Sexes

Male-

recorded

No. of
Exp.

in which
male was bred

Henny
cocks

1

No. of Pen

Male Female

cocks

Males

Females

1

21/12 + 13 + 14

396/11 (l\) X 152/11 (h\)

—

14

8

33

52

2

21/13 + 14

,, X Br. Legh.

—

10

10

21

15

3

23/13

203/11 (Fi) X 376/11 (Fj)

—

7

6

13

11

4

J,

,, X Br. Legh.

—

2

3

5

4

5

24/12 + 13

397/11 (Fi)x 380/11 (Fi)

—

9

17

39

33

6

9)

,, x Br. Legh.

—

5

7

16

8

7

7/12

Fi x 2 9 $

—

5

6

12

11

8

10/13

1.39/12 (F.,)x Br. Legh.

—

8

9

19

22

9

12/14, 4/15

x3$9

—

12

8

21

28

10

16/13

28/12 X Br. Legh.

431

8

11

19

31

]1

27/13, 24/14

190/12 x 807/12

431

3

4

9

9

12

12/13

552/12 X 964/12

23

7

8

23

22

13

4/12

266/11 X 827/11

7

7

2

12

9

14

lO/lo

201/14 X 2 $ $

29

3

5

13

11

15

11/15, 9/i(;

165/14 X 2 Br. Legh.

28

—

4

4

10

1(5

17/15, 19/16

286/14 X 2 $ $ Br. Legh. (a)

4

6

10

26

17

23/15

492/14 X 2 9 0

32

3

4

7

3

18

8/16 + 17

70/14 X 3 ? $

2

4

4

10

10

19

6/17

242/16 X Br. Legh.

16

3

1

5

1

20

18/17, 23/18 + :

19 268/16 X 479/16

18

9

10

20

11

21

11/18

338/17 X Br. Legh.

19

2

—

8

9

22

19/18

321/17x3 9 9

442

1

3

5

9

Totals

126

136

324

345

B

. Hen henmj.

Details of Mating

No. of Exp.

Progeny

Sexes

■Vr» r»f

Male-
leaLiierecl

Henny
cocks

J recorded

1^ O. UI

Exp.

No. of Pen

Female Male female was bred

cocks

Males

Females

23

22/12 + 13

151/11 (Fi)x 379/11 (Fi)

—

10

15

29

36

24

4/14

,, xBr. Legh.

—

11

8

24

12

25

25/13

383/11 (F,) X 398/11 (Fj)

—

6

3

15

10

26

14/12

146/11x187/11

W)

6

8

14

14

27

26/13

658/12 X Br. Legh.

431

2

1

13

11

28

4/14

75/12 X

431

4

1

7

12

29

19/14

998/12 X

431

6

5

11

12

30

5/15

i Silver ,„,,.„
t Hamburgh ^'^^^/l^

—

1

1

2

4

31

8/15

110/14 X 191/14

(7)

5

5

11

12

32

26/14 + 15

^oi/i-^MS'lS

3

14

5

20

13

33

21/15

501/14x407/14

33

3

4

9

7

34

10/17

210/16 X 271/16

442

3

5

8

7

35

19/17

235/16 X 134/15

442

7

2

9

7

36

20/17

388/16 X 44/16

442

5

5

11

9

37

23/17

294/16 X 131/16

442

4

7

11

16

38

10/18

353/17 X Aseel

19

4

2

9

5

39

17/18

98/16 X 44/16

442

1

3

4

11

40

20/18

95/15x271/16

81

9

5

17

11

41

22/18

150/17 X 119/17

(S)

2

2

6

—

42

10/19

322/17 X R. I. B.

<S)

4

2

12

7

45

-21/18,-11/19 320/17 xj^Xrcup

(5)

1

4

6

6

108

93

248

222

From A above

126

136

Totals

234

229

1 See p. 39. * See p. 40.

(a) Grandson from Exp. 43 (of. p. 39).

(/3) Granddaughter of Fj henny ^ .

(y) Granddaughter from Exp. 43 (cf. p. 39).

(5) Granddaughter from Exp. 44 (cf. p. 40).

56

Genetic Studies in Poultry

TABLE II.

A. From heiiny S •

No. of Normal llenny

Exp. Males Females Males Males

Families with
excess of

??

14

11

4
6
3

B. From henny $ .

No. of Normal Henny

Exp. Males Females Males Males

28 7 12 4 1

30 2 4 1 1

37 11 16 4 7

39 4 11 1 3

Totals

Families with
excess of

71 128

/ 6
17
19
20

16
7
5

20

3

1

11

27

32

7

4

1

10

24
25
32
38
40
41
42

24

24
15
20

9
17

6
12

43

12
10
13
5
11

10

11

6
14
4
9
2
4

12

Totals

48 23 20 22

/ 2

21

15

10

10

3

13

11

7

6

4

5

4

2

3

5

39

33

9

17

Families with

7

12

11

5

6

proportions of

sexes nearly

equal

8

9

11

12

19

21

9

23

22

28

9

22

8

12

3

7

9

8
4
8

13

12

9

7

2

14

13

11

3

5

18

10

10

4

4

. 21

8

9

2

—

Totals

205

194

79

82

23
26
27
29
31
33
34
35
36
37
45

103

29

14

13

11

11

9

8

9

11

11

6

58

36

14

11

12

12

7

7

7

9

16
6

50

10
6
2
6
5
3
3
7
5
4
1

27

15
8
1
5
5
4
5
2
5
7
4

132 137

52

61

EXPLANATION OF PLATES.

PLATE VII.

Fig. 1. Saddle feathers from an intermediate cock with a marked tendency to normal
plumage. The feathers were gold throughout, and were removed when the bird was
28 weeks old.

Fig. 2. Saddle feathers from brother of above bird taken after the moult. Before the
moult the feathers of this bird were of the type shewn in Fig. 1. (Of. p. 42.)

Fig. 3. Feathers from saddle of slightly intermediate bird taken before the moult. The
fringes of the two "tipped" feathers are gold in colour.

Fig. 4. Two henny feathers from saddle of above bird taken after the moult. All of the
feathers were of this type.

R. C. PUNNETT AND THE LATE MaJOR P. G. BaILEY 57

PLATE VIII.

Fig. 1. Saddle feathers of an intermediate with a marked tendency to normal plumage.
Feathers taken before the moult.

Fig. 2. Henny saddle feathers of an intermediate taken after the moult. Before the
moult the feathers were similar to those shewn in Fig. 1.

Fig. 3. Saddle feathers of same bird taken after the fourth moult. They are of inter-
mediate type and fringed with gold.

Fig. 4. Henuy saddle feathers of same bird taken after the tifth moult. (Cf. p. -12.)

PLATE IX.

Skins of two cocks bred from Silver Hamburgh 5 x Gold Hamburgh ^ (Exp. 30). On the
left (1) a henny and on the right (2) a normal feathered bird. Both were kept over
the moult before being killed.

PLATE X.

Skins of two intermediate feathered birds, from same mating. No. 1 was killed at
9 months while No. 2 was killed a year later. Before the moult No. 2 closely
resembled No. 1 in the nature of its feathering. (Cf. p. 42.)

PLATE XI.

1. Skin of henny (J . General colour warm brown.

2. Skin of castrated henny (J 204/15, in general colour closely resembling a Brown

Leghorn. Before castration he resembled No. 1. (Cf. p. 49.)

REFERENCES.

Darwin, C. The Variation of Animals and Plants nnder Domestication. Loudon,

1868.
GoLDSCHMiDT, R. " Untei'sucliuugeu iibei' lutersexualitiit." Zeit. iad. Ahst. Vererb.

XXIII. 1920.

Mechanismus und Physiologie der Geschlechtsbestimmung. Berlin, 1920.

Good ALE, H. D. " Gonadectomy in relation to the Secondary Sexual Characters

of some Domestic Birds." Carnegie Inst. Wash. Publ. 243, 1916.
Lamon, H. M., and Slocum, R. R. The Mating and Breeding of Poidtry. New York

and London, 1920.
Morgan, T. H. " The Genetic and the Operative Evidence relating to Secondary

Sexual Characters." Carnegie List. Wash. Publ. 285, 1919.

"The Endocrine Secretion of Hen Feathered Fowls." Emlocrinology, iv. 1920.

Nettleship, E. "A History of Congenital Stationary Night-blindness in nine con-
secutive generations." Ophthal. Soc. Trans. Vol. xxvii. 1907.

" Some hereditary diseases of the eye." (Bowman Lecture.) Ophthal. Soc.

Trans. Vol. xxix. 1909.

Newman, H. H. " Five Generations of Congenital Stationary Night-blindness in an

American Family." Journal of Genetics, Vol. iii. 1913.
P^ZARD, A. " Le conditionnement physiologique des Caractferes sexuels secondaires

chez les Oiseaux." Bull. Biol. Fr. Belg., lii. 1918.
Punnett, R. C, and Bailey, P. G. " On Inheritance of Weight in Poultry." Journal

of Genetics, Vol. iv. 1914.
Wright, L. The Illustrated Bool: of Poultry. London (1876).

JOURNAL OF GENETICS, VOL XI. NO. 1

PLATE VII

Fig. 1. cjl05/18 (12. 11.18)

Fig. 2. (J 101/18(1.8.20)

Fig. 3. (^104/15(21. 4. 16)

Fig. 4. .^104/15(21.10.16)

JOURNAL OF GENETICS, VOL XI. NO. 1

PLATE VII

Fig. 1. ^ 126/15 (14. 11. 13)

Fig. 2. ^ 201/14 (8. 11. 15)

Fig. 3. (J 201/14 (19. 9. 18)

Fig. 4. ^ 201/14 (8. 4. 20)

JOURNAL OF GENETICS, VOL. XL NO. 1

PLATE IX

JOURNAL OF GENETICS, VOL XI. NO. 1

PLATE X

- /

.1^ f^

Ef

^

JOURNAL OF GENETICS, VOL. XL NO. 1

PLATE XI

A MENDELIAN EXPERIMENT WITH ABERDEEN-
ANGUS AND WEST HIGHLAND CATTLE.

By J. A. S. WATSON, B.Sc, M.C.

(With Plate XII.)

While Mendclian analysis has been very successfully applied to the
study of the heredity of the smaller animals and of plants, very little
progress has so far been made in the investigation of the mode of in-
heritance of the characters of the larger and more important domesticated
animals. This fact is of course sufficiently explained by the obvious
difficulties — the large financial outlay and the extended period of time
necessary to obtain experimental results. It would appear, however,
that there is no alternative to the method of tedious breeding experi-
ments if the many and important problems connected with animal
breeding are to be solved. A certain amount of progress has indeed
been made by the study of records contained in herd- and stud-books,
but it is certain that the results of such investigations lack the scientific
conclusiveness of those of carefully planned and critical experiments.
Thus such an apparently simple question as that of the inheritance of
the red, white and roan colours in Shorthorn cattle, on which practically
unlimited statistical evidence is available, has been investigated by
several workers, and three quite distinct Mendelian interpretations
have been proposed — by Wilson (1), Laughlin (2), and Wentworth (3),
respectively — no one of which gives a complete and satisfactory expla-
nation of the facts.

The present experiment was commenced in 1910; the object laid
down was to determine the mode of inheritance, in crosses between
Aberdeen-Angus and West Highland Cattle, of the horned and polled
conditions, of colour differences, of hair characters and of differences in
conformation. The two breeds may be briefly described. The Aberdeen-
Angus is black in colour, with the exception that a small amount of
white is permissible on the underline behind the umbilicus, and on the
end of the tail. It is hornless, with a relatively short, smooth coat, and
is of the early-maturing beef type — i.e. short of leg, blocky, wide, heavily
fleshed and small boned. With regard to the purity, in the Mendelian

60 Mendelia7i Experiment ivith Cattle

sense of the colour and of the hornless condition, it is to be noted that
red calves are occasionally born from pure bred black parents even in
the most carefully selected herds. White markings, beyond those per-
mitted by the breed standard also occur at times, more particularly
white feet. Horns, or more generally what are called "scurs" — small,
firm or loose horny excrescences — were once comparatively common but
have now been all but eliminated by selection.

The general character of West Highland cattle is well known. They
carry large spreading horns, a long and shaggy coat, and are of beef type,
short of leg and blocky but distinctly narrower in build and less heavily
fleshed than the Aberdeen- Angus. They are also characteristically later
in reaching maturity and less readily fattened. The colours are many —
red, black, yellow, dun and brindled — accurate classification being some-
what difficult in a percentage of cases. In the present experiment cattle
of only two colours were included — viz. red and dun.

Four different pedigreed Angus bulls were employed in the experi-
ment, and were mated to seven different pure bred West Highland
cows. Apart from the latter, two cows, which happened to be available,
were included, the result of a cross between the Chartley (a white horned
breed, with black " points ") and the West Highland. One female of the
Angus X Chartley-Highland cross was retained for breeding. The in-
heritance of the individual characters may now be considered.

(1) Coat and Conformation.

Unfortunately no data worth presenting were obtained with regard
to the inheritance of these characters. Attempts were made during the
earlier part of the experiment to reduce the descriptions of the coat to
actual measurements of the length and diameter of the hairs. The main
difficulty, apart from that of sampling, was that the seasonal differences
varied greatly as between different animals — e.g. an animal which had
a relatively heavy winter coat might have a relatively light summer
coat. Again the hair development is influenced by causes other than
heredity, notably by the "condition" or degree of fatness, and pregnancy.
The value of these measurements was therefore doubtful, but frequent
and regular examination might still have yielded something of interest.
The author, however, was able to see the cattle only at rare and irregular
intervals during the period of the war, and this line of investigation was
of necessity given up. The general impression made by the cattle was
that the first hybrid generation was intermediate with regard to coat.

J. A. S. Watson

61

The individuals were similar in coat character, with one exception
( $ No. 8), which was distinctly shorter haired. In F^ there was certainly
some form of segregation, individuals being obtained which approached
somewhat the Highland type and others which were almost or perhaps
quite as smooth as the average Angus (cf PL XII, fig. 2). The distribu-
tion, however, suggested a ratio considerably more complex than the
simple monohybrid one of 1 Rough : 2 Intermediate : 1 Smooth. With
regard to conformation, certain of the F.^ suggested rather strongly, in
one or other point, one of the parental types; but no statement beyond
this would be justifiable.

(2) Horns and Colour.
Below follows a descriptive list of all the animals bred during the
course of the experiment,

{a) Horns.

The existing true breeding types of cattle are either (a) horned in
both sexes, or (h) polled in both. The third possible condition, which is
found for example in certain breeds of sheep, in which the males bear
horns while the females are hornless, is not known among living cattle,
though Major (4) has described this condition in skulls from the tertiary
deposits in Italy.

F^ Generation

Sire

Dam

J

1

Date

Kef.

Ref.

Breed and

Ref.

Breed and

No.

Sex

No.

description

No.

Description

born

Horns

Colour

Remarks

Fil

O

AA 1

Angus-Black-
Polled

H 1

W. Highland
Dark Red,
Horned

Dec. 1911

Short
horns

Black

F,2

?

AA.3

' '

H, C3

Highland x
Chartley
Reddish Dun,
Horned

Feb. 1912

Polled

Dun

F,S

o

AA3

'*

H, C4

Highland x
Chartley
Reddish Dun,
Horned

Apr. 1912

Polled

Red

Sold at 3
months old

F,4.

c?

AAl

»)

H2

Highland Dark May, 1912

?

Black ?

Still born

Red, Horned

Fi5

?

AA2

"

H5

Highland
Yellow Dun,
Horned

Jan. 1912

Polled

Dun

Fi6

?

AA2

H6

Red Horned

Dec. 1911

Polled

Black

1^1 7

$

AA2

H7

Red Horned

Jan. 1912

Polled

Black

Fi8

?

AA4

) )

H9

?* Horned

June, 1911?

PoUed

Black

F,d

9

AA4

H9

? * Horned

Feb. 1912?

Polled

Black

* Fi No. 8 and No. 9 were purchased as yearlings. Their descent from a pedigreed
Angus bull and pure bred Highland cows was unquestionable, but their particular dams
could not be identified with certainty. The latter were probably red.

62

Mendelian Experiment with Cattle

ii^o Generation.

Ref. No.

Sire

Dam

Calved

Sex

Horns

Colour

^2 1

F, 1

l'\ 2

1914

Female

Polled

Black

Fo 6

1915

Female

Horned

Dun

^2 11

1916

Female

Polled

Dun

Fa 17

1917

Female

Polled

Silver Dun

^2 23

,,

»»

1918

Female

Polled

Red

^2 2

Fil

Fi 5

1914

Female

Polled

Black

Fa 7

1915

Female

Polled

Dun

F2I2

,,

,

1916

Male (east.)

Polled

Red

Fa 18

)>

>,

1917

Female

Polled

Dun

F2 3

Fjl

F, 6

1914

Female

Polled

Black

Fa 13

„

1916

Male (cast.)

Horned

Black

F, 19

J,

1917

Female

Horned

Red

Fa 24

)>

j>

1918

Female

Polled

Red

i*2 4

Fa .8
Fa 14
Fa 20
Fa 25

F,l

F, 7

1914

liilo
1910
1917
1918

Male (cast.)

Male (cast.)
Female
Female
Female

Hardseurs Black
under hair

Horned Black

Horned Black

Polled Red

Polled Black

Fa 5
Fa 9
Fa 15
Fa 21
F222

F, 1

Fi 8

1914
1915
1916
1917
1918

Female
Male (cast.)
Male (cast.)
Female
Male

Polled

Horned

Horned

Polled

Polled

Black
Black
Black
Black
Black

Fa 10
Fa 16

F, 1

Fi 9

1915
1916

Female
Female

Polled
Polled

Black
Black

Note. There is some doubt about the respective dams of Fa 13 and 14 and 19 and 20.
It is possible that the dams are reversed in either case or both, i.e. 13 and 19 may be from
F] 7 and 14 and 20 from Fj 6.

In the present experiment the F^ generation females were all " clean
polled," i.e. without any trace of horn development. F^ No. 3, a bull calf,
was sold for slaughter at three months of age, at which time no horn
development had occurred. The bull ^1 No. 1 (PI. XII, fig. 1) developed
peculiar short stout horns, which grew at an abnormally slow rate. At
18 months they were about 4 ins. long. At 6 years they measured 10 ins.
There can be no doubt that this animal, like the other F^ crosses, was
heterozygous for the polled character. We should therefore conclude
that the polled condition is completely dominant in the female, while
in the hybrid male the horn development is inhibited, but at least not
always completely suppressed. These results are in accord with those of
Gowen (5) who obtained the following in first hybrids between Aberdeen-
Angus and various horned breeds.

Horned

Hard scurs

Loose scurs

Polled

Male

J. A. S. Watson 63

On the other hand, Lloyd Jones and Evvard (6), in crossing the
Galloway (polled) and Shorthorn obtained the following in 78 i^, :

70 Clean Polled,
6 Scurred,
2 Horned.

Unfortunately the sexes of the eight exceptional animals are not given,
but the two horned animals are attributed by the authors, with much
reasonableness, to the probably impure condition (with regard to horn-
lessness) of certain " grade " Galloway cows which were included in the
experiment ; and they found " no evidence that sex is in any way con-
nected with the inheritance of these characters." Against this it may be
mentioned that among crosses between Red Polls and Ayrshires at
present being bred in Dumfriesshire, a large proportion of the F^ males
bear horns. It appears certain that the degree of dominance of the polled
character in the male varies according to the particular breeds employed,
and varies too as between different individuals of the same cross.

In the present experiment the distribution of horns in Fo was as
follows :

Polled Hard"scurs" Normal horns
Females 15 — 3

Males (castrated)

Totals

Including the male with
18 polled, 7 horned, which ag

2 1 4

17

scurs " as polled, we obtain the numbers
>Tees very closely with the simple Men-
delian ratio of SZ) : IR. As the numbers are small, it is probably worth
while to combine them with those obtained by Lloyd-Jones and Evvard
in the Fo of the Galloway x Shorthorn cross above referred to, thus :

Female Male Totals

Present Experiment
Lloyd Jones and Evvard...

Tolled

15

9

Horned
3
4

Polled
3
7

Horned
4

1

Polled
18
16

Horned

7
5

Totals

Expectation (SD : IR)

24
23-25

7
7-75

10
11-25

5

3-7.5

34
34-5

12
11-5

The numbers are again in very close agreement with the hypothesis,
originally advanced by Bateson and Saunders (7) and Spillman (8 and
8a), that the homed and polled conditions form a simple pair of Men-
delian characters.

To revert to the question of dominance in the male, Wood (9) and
Arkell (10) have found that in crosses between horned and hornless

64 Mendelian Experiment with Cattle

breeds of sheep, the heterozygous males are horned, while the hetero-
zygous females are hornless. The horn development in the hybrid male
has been shown to be dependent on a hormone secreted by the testis,
and horn growth is stopped (again in the hybrid) by castration. The
analogous experiment has not been performed with cattle. It is to be
noted that the F, males were all castrated.

(h) Colour.

The ^1 generation, with the sole exception of the red calf F^ 3, were
either dun or black. With regard firstly to the behaviour of black and
red, we should conclude that black is dominant, red recessive. Matings
of the heterozygous ^i blacks inter se {F^ females 6, 7, 8 and 9) produced
16 ^2 of which 13 were black, 3 red, against an expectation of 12 and 4.
These results may be regarded as in complete accord with the accepted
view, originally advanced by Spillman(ll) that red behaves as a simple
recessive to black. The red calf F^ 3 calls for further explanation, but a
highly probable explanation is at hand, namely that its Aberdeen-Angus
sire J. J. 3 was heterozygous for black. The appearance of the recessive
reds in pure bred Angus herds, already referred to, renders such an
explanation inherently probable. Moreover the bull in question was in
use by a local farmer for crossing with heifers of mixed colours, and with
them he begot a large proportion of red and red-roan calves. In this
connection it may be said that the sires A A 2 and A A 4c left no red
calves in the herds where they were used on cattle of various colours.
AAl cannot be regarded as having been tested.

The dun x black matings constitute a more complex problem. The
hypothesis thus far proposed must first be considered. Wilson (12 and
13) has proposed a series of Multiple allelomorphs or " polygamous
factors," any one of which behaves as a Mendelian alternative to any
other. In so far as the present experiment is concerned, three factors
would be concerned, viz. B (black), R (red), and L (Light Dun). The
following are the colours allotted by Wilson to the various possible factor
combinations :

BB Black (homozygous)

BR Black (heterozygous)

BL Dun

RR Red

RL Yellow

LL Light Dun

J. A. S. Watson 65

According to this the dun F^ females 2 and 5 would be BL and the
Fi male of course BR. The chances of the various combinations wf»uld be

/ BB Black (homozygous)
I BR Black (heterozygous)

1 BL Dun

/ RL Yellow

We should therefore expect a ratio of

2 Black : 1 Dun : 1 Yellow.
The obtained results :

2 Black : 5 Dun : 2 Red,

not only suggest a different ratio, but give a colour which is not pro-
vided for by the scheme. As has already been pointed out by Babcock
and Clausen (14), Lloyd Jones and Evvard in the experiment, already
alluded to, obtained 6 reds out of 26 Fo in crosses of White Shorthorn
and Galloway, which again are inexplicable on Wilson's theory. The
latter, in its present form, must therefore be regarded as inadequate.

Wright (15), on the contrary, has proposed a system of ordinary unit
factors, only two of which would be concerned in the present case, viz. :

E, black, its absence e giving red, and

D, a dominant pigment dilution factor, in whose presence black is
modified to dun, and red to yellow. The nine possible factor combinations
give the following respective colours :

dclEE Black, homozygous

dclEe Black, heterozygous

ddee Red

DdEE Dun (homozygous for black factor)

DdEe Dun (heterozygous for black factor)

Ddee Yellow

DDEE Cream dun (homozygous for black factor)

DDee Cream (light dun)

On this hypothesis the male F^ 1 would be ddEe, and both dun
females, F^ 2 and F-^ 5, DdEe\ The possible combinations and the
probable frequency would then be :

1 No. 5 might conceivably have been DdEE, although the chances are slight, black
being a comparatively rare colour among Highland cattle. The fact that she produced a
red calf, however, shows definitely that she was heterozygous for the black factor.
Journ. of Gen. xi **

66 Mendelian Experiment with Cattle

1 DdEE (dun), 2 BdEe (dun) = 3 dun
1 Bdee (yellow) = 1 yellow

1 ddEE (black), 2 ddEe (black) = 3 black
1 ddee (red) = 1 red.

The results obtained, 5 dun, 2 black, 2 red, while not agreeing closely
with expectation, contain nothing that is definitely opposed to Wright's
hypothesis.

It is obvious, however, that, so far as concerns the present experiment,
a third explanation is possible, which has the merit of greater simplicity,
viz. that the factor D is not a colour dilution factor, modifying both
black and red, but an independent factor for dun colour, epistatic to E
(black), and producing dun whenever present. This hypothesis would
give an expected ratio of 4 dun, 3 black, 1 red, which is comparatively
near to the ratio obtained. The numbers are, however, obviously too
small to furnish any definite proof of such an hypothesis.

Conclusions.

(1) The polled and horned conditions form a simple Mendelian pair.
The polled condition is completely dominant in the female, while in the
heterozygous male horn development is inhibited but not always sup-
pressed.

(2) Black is dominant to red, and the colours behave as a simple
Mendelian pair.

(3) The hypothesis of multiple allelomorphs for colour, proposed by
Wilson, is not in agreement with the results obtained.

(4) Dun is dominant to black, but whether as a simple epistatic, or
whether produced by a dilution factor capable of modifying colours
other than black, does not appear from this experiment.

The author desires to express his thanks to the Moray Fund for
Research, Edinburgh University, and to the Board of Agriculture for
Scotland, for the necessary funds; to Lord Forteviot of Dupplin for
providing facilities for the woi'k, and to his lordship's agent, Mr J. J.
Simpson ; to Mr Wm. Bruce, B.Sc, Edinburgh and East of Scotland
College of Agriculture, who supervised the experiment during the greater
part of the period 1914-1918 ; and to Prof J. Cossar Ewart of Edinburgh
University, and Sir R. B. Greig of the Scottish Board, for assistance
and advice.

JOURNAL OF GENETICS, VOL. XL NO. 1

PLATE XII

Fi". 1. l\ Bull at 2 years 'J uiontlis old.

Fig. 2. F^ Yearling heifers showing extreme types of coat. Of similar age, and from the same lot.

Fig. 3. Dun i-'i Cow with Black Fo Calf.

J. A. S. Watson 67

EXPLANATION OF PLATE XII.

Fig. 1. Fi (J (No. 1) at 2 years 9 months old.

Fig. 2. Two F2 yearling heifers shewing extreme types of coat.

Fig. 3. Dun Fx cow with black F^ calf.

LITERATURE CITED.

1. Wilson, James. " Mendelian Characters among Shorthorn Cattle." Set. Proc.

Roy. Dub. Soc. II. (1908.)

2. Laughlin, H. H. "The Inheritance of Colonr in Shorthorn Cattle." Amer.

Naturalist, Vol. v. 45. No. 540. (1911.)

3. Wentworth, E. N. "Colour in Shorthorn Cattle." American Breedem^

Magazine., Vol. v. No. 4. (1913.)

4. Major, C. F. J. " On the Mammalian Fauna of the Val D'Arno." Qaart.

Jour. Geol. Soc. Lond. Vol. v. No. 41. (1885.)

5. GowEN, J. W. " Studies in Inheritance, of certain characters of Cro.sses between

Dairy and Beef Breeds of Cattle." Jour. Agric. Research, Washington, xv. 1.
(1918.)

6. Lloyd Jones, 0. and Evvard, J. M. " Inheritance of Colour and Horns in

blue-grey cattle." Iowa Sta. Research BidL 30, Ames, Iowa, 1916.

7. Bateson and Saunders. " Experimental Studies in the Physiology of Here-

dity." Royal Society (London). Repts. Evolution Committee, No. 1, footnote,
p. 160. (1901.)

8. Spillman, W. J. " Mendel's Law in Relation to Animal Breeding." Repts.

Amer. Breeders^ Association. (1905.)
8a. Spillman, W. J. "A Mendelian Character in Cattle." Science, N.S. 23,
No. 588, p. 549. (1906.)

9. Wood, T. B. " Note on the Inheritance of Horns and Face Colour in Sheep."

Journal Agric. Science, Vol. i. pt. 3. (1905.)

10. Arkell, T. R. "Some Data on the Inheritance of Horns in Sheep." Hamps.

Agr. Exft. Sta. Bidl. No. 160. (1912.)

11. Spillman, W. J. Science, N.S. 25, No. 634. (1907.)

12. Wilson, J. "The Colours of Highland Cattle." Sci. Proc. Roy. Dublin Soc.

Vol. XII. (N.S.). (1909.)

13. Wilson, J. A Manual of Mendelism. London. (1916.)

14. Babcock and Clausen. Genetics in Relation to Agriculture. (1918.)

15. Wright, Sewall. " Color Inheritance in Mammals." (VI. Cattle.) Journal

of Heredity, Vol. viii. 11. (1917.)

5—2

NOTE ON THE EVOLUTION OF THE DOUBLE
STOCK (MATTHIOLA INC AN A).

By EDITH R. SAUNDERS,

Fellow of Newnham College, Cambridge.

(With Three text-figures.)

The question whether the fully double Stock, as we know it to-day,
arose in the first instance by a single considerable mutation producing
the form with which every flower lover is familiar, must have arisen in
the minds of many horticulturists since the plant became a favourite in
our gardens. So far as I am aware no statement exists as to when, where
or by whom the double form was first observed. The earliest reference
known to me, as I have stated elsewhere\ is that by Dodoens^ in 1568.
In this account as in the later illustration by de I'Obel and Pena* (1576)
it is a full double which is depicted, a double, that is, destitute of any
semblance of either stamens or carpels. In view of this testimony and
of the further facts (1) that all double Stock strains now in cultivation
are of this fully double type, (2) that we know that in the case of certain
other genera having different grades of doubles (Wallflower, Lobelia) the
appearance of the full double preceded — in the former case probably by
some centuries — that of the semi-double, (3) that in a full double of such
recent origin as Arahis albida we have no knowledge of any intermediate
stage in the doubling process, we may safely accept the evidence for the
mutation having been accomplished in the Stock by a single step as
amounting to proof. Nevertheless we meet with some records in the
literature dealing with Matthiola which might be construed as conflicting
with this view, and which in any case call for some further explanation.
I have been unable until now to come to any definite conclusion in regard
to this counter-evidence, but certain specimens noticed of late in my own
cultures appear to me to provide the solution. I have therefore thought

1 "The Double Stock, its History and Behaviour," Journal Roy. Hort. Soc. Vol. xl.
Part III. 1915.

•^ Florum et coronarianivi

* Stirpium adversaria nova.

70 Evolution of the Double Stock

it worth while to put these observations on record and so clear the
ground of any doubt which these earlier references, as they stand, might
throw on the validity of the conclusion expressed above. Taking the
references in question in their chronological order we have :

(1) Sowerby's illustration of Cheiranthus incanus in the first and
second editions of English Botany^. The specimen figured was obtained
in 1808 by lowering a boy over the cliffs near Hastings. It is on record
that Hooker was present on the occasion as well as Borrer (who with
Dawson Turner discovered the plant in this locality in 1806), and it may
be inferred from the account that only the one specimen was taken. The
illustration represents a simple raceme on which only four flowers are
as yet open, the remainder being still in the bud stage. The first (lowest)
of these is shown viewed from the back with only calyx and corolla
visible. Both appear to be normal. In the succeeding (second) flower
there are nine petals of about normal size and an additional very small
one in the centre where only three out of the four long stamens appear.
The third flower is shown with five petals and the anthers of the four
long stamens. In the fourth flower we have, so far as can be judged,
again a normal single. Now it is to be noted that in the third edition
this illustration is replaced by a drawing of a plant obtained from the
Isle of Wight. [The plant had been lost apparently from the Hastings
locality as far back as 1835^.] I have been unable to trace in botanical
writings any allusion to the reason for this substitution and although
many of Borrer's specimens are preserved at Kew, I cannot find that the
1808 Stock is among them. It may be presumed, I think, that it is not
now in existence. The probable explanation of why it was discarded will
appear presently.

(2) A statement in Hopkirk's Flora Anomala^ (1817), a work brought
to my .notice by Dr Agnes Arber. The passage runs as follows : " Of
multiplicate polypetalous flowers we have many examples, as Gistus
helianthemum, Anemone nemorosa, Sanguinaria Canadensis, Tormentilla
officinalis, Hibiscus rnutahilis, Tulipa sylvestris, Cheiranthus Cheiri, Hes-
peris matronalis, Matthiola incana, and many others. In these instances,
the stamens and pistil not being injured, the seeds may be produced as
in single flowers." [Multiplicate is here used to indicate flowers in which
the number of petals is augmented without affecting the other parts.]

1 Vol. XXVII. pi. 1935 in order of issue (Vol. xi. pi. 993 iu copies bound in order of
systematic relationship).

2 See Watson, The New Botanisfs Guide, p. 51.
^ P. 114.

E. R. Saunders 71

111 a footnote Hopkirk adds the remark — " The semi-double variety of
Hibiscus mutabilis, frequently produces seeds, and these almost always,
in their turn, produce semi-double and double flowers, which the seed
from single flowers seldom do, and the same thing may be observed in
Stocks, Wallflowers, etc." A pronouncement in such general terms as
that contained in the concluding sentence of the footnote we may, I
think, without hesitation, disregard. It carries with it a suggestion of
being merely the usual repetition of untested tradition, not a confirmation
from first-hand observation. I doubt whether Hopkirk's earlier state-
ment is likely to be based on any better authority. Even if well-founded
it imports no more, probably, than that other observers have met with
the same type of abnormality as that which has occurred in my cultures
(see later), the appearance of which has made it possible to clear up the
position.

(3) The statement in De Candolle's description of Matthiola annua
(Si/st. II. p. 165, 1821) that the flowers are in colour similar to those of
M. incana and always single o?* rarely semi-double. (The italics are
mine.)

(4) The statement by Phillips {Flora Historica, Vol. ii. p. 29, 1824)
that there is frequently a straggling anther to be found in the double
blossoms [of the Stock].

(5) The illustration of Matthiola incana in Flowering Plants and
Ferns of Great Britain, by Anne Pratt, which is very similar to Sowerby's
drawing. At the time of writing this work Anne Pratt was living at
Dover, at no great distance, therefore, from the spot where Sowerby's
specimen had been obtained. But, as stated above, M. incana was believed
to have disappeared from this locality long before this date'. We gather
from Bromfield- that in the Isle of Wight, another well-known station,
the plant, as at Hastings, was by no means easy of access. It was
known to occur also on the cliffs of Ramsgate and Broadstairs^ and all
three localities are cited by Anne Pratt. From a passing remark^ one
infers that recourse was not had by the authoress to a garden form as
a model, and it is therefore a question of some interest whether she
procured a new wild specimen for her plate or whether we may take
it that her illustration is a free rendering either of Borrer's original
plant or of Sowerby's drawing of it. A comparison of the two plates

1 1854 probably, but the first edition is undated.

^ Flora Vectensis, 1856.

3 Cowell, Floral Guide to East Kent, 1839.

^ Loc. cit. Vol. I. p. 135.

72 Evolution of the Double Stock

leaves no doubt in my mind that the illustrations both of M. incana
and of M. sinuata in Anne Pratt's work are adapted from Sowerby. One
has only to note the number and position of the open flowers and of the
unopened buds in the two cases to become convinced on this point. So
far then as illustration (as opposed to description) of anything approaching
a semi-double Stock goes we need only take account of Sowerby's original
plate. In this case we have the categorical statement that the drawing
was made from the actual specimen obtained with so much difficulty.
It represents, as will be clear from the description given above, a type
of individual apparently of an intermediate grade between a normal
single and a typical double. Nevertheless, from what follows, it will
be seen that it must be regarded as a genuine single, and that the
partially double flowers result from a morphological modification different
from that of ordinary doubling. It is well known that an additional
fifth petal may occasionally make its appearance in one or two flowers
on a stock plant otherwise single. We find this abnormality mentioned
by HilP, Phillips^, Chate* and others. It is of rare occurrence and
appears to have no hereditary significance. We cannot tell on which
individual we shall find, among the hundreds of normal flowers pro-
duced in the course of the season, one or two with Cg instead of C^.
We can however predict with some certainty the position in which,
if present, these flowers will occur. My experience is that when this
abnormal condition exists it is to be found in the lowest flowers on
the axes, and among those that open first in the season. That is to
say the abnormality is associated with the region and the period of
greatest vigour. Now it is also among flowers occupying this position
that a more pronounced deviation from the normal is occasionally to be
met with, and I feel no doubt that it is this more extreme type of mon-
strosity which chanced to be exhibited in Sowerby's specimen, and which,
occurring now and then, as it evidently does, gave rise to statements
such as those quoted above. We may safely conclude that it was the
realisation that the original specimen was an aberrant form which led
to the substitution of a new plate of M. incana in the third edition of
English Botany. The modification in this case arises from a more or less
complete " twinning " of a normal single flower. Figs. 1 and 2 represent
two instances of imperfect twinning occurring, in each case, in the first
{lowest) fiower of the inflorescence.

1 Eden, p. 567, 1757.

'^ Loc. cit.

^ Cult. prat, des Girqflies, pp. 63, 64.

E. K Saunders

73

In plant A (Fig. 1) the " flower " showed 5 sepals, 6 petals, 10 stamens
(only those extending above the stigmas are shown) and a gynoccium
with 4 distinct stigma lobes terminating what appeared to be 4 carpels

Fig. 1. The lowest "Hower " on the main axis of plant A viewed from
above, showing imperfect "twinning"' (slightly enlarged).

showing slight torsion and incomplete fusion along one suture so that
the ovules were exposed. In plant B (Fig. 2) the calyx was formed of
6 sepals arranged in two groups of three. As in the previous case there
were 6 petals and 10 stamens of which eight only are here represented,

2c

Fig. 2. A similar case from another plant B. a "flower" viewed as in Fig. 1, b calyx seen
from below, c monstrous gynoecium (considerably enlarged).

74

Evolution of the Double Stock

thu other two being hidden. The monstrous gynoecium resulting from
the twin ovaries appeared to have 5 stigma lobes. In Fig. 3 taken
from a flower in the same position on a lateral axis of a third plant (0) the

twin nature of the gynoecium is seen at a
very much later stage. In the case of plant
A, the only one examined for this point,
the pedicel was somewhat thicker than
usual and showed, on cross section, an ellipti-
cal instead of the normal ring arrangement
of the fibro-vascular bundles, characteristic
of the true double as well as the single.
The stouter form of the pedicel can be de-
tected in Sowerby's drawing (second flower)
and though slight, the increase is no doubt
faithfully portrayed. The fact that the twin
flowers appear to develope from a common
basis accounts for the incomplete duplica-
tion of the parts on the adaxial sides. We
get no indication from Sowerby's figure
(owing to the view presented) of the com-
position of the calyx and gynoecium of the
second flower, but if my interpretation is
correct we have here a case of twinning
carried to a further stage, combined, ap-
parently, with petalody of at least one
member of the androecium. We may include in the same category no
doubt the case observed by Master's^ and described under the head of
synanthy^ Here duplication of calyx, corolla and androecium was all but
complete, only the two short stamens at the point where separation
was imperfect being suppressed. We may then I think conclude, not-
withstanding the statements of various writers which might be inter-
preted to the contrary, that there is no case on record of a genuine semi-
double Stock ; and that the records in question owe their origin to the
occasional occurrence of more or less perfect " twinning " of the lowest
flowers on the earliest flowering axes.

I wish, in conclusion, to express my grateful thanks to Miss D. F. M.
Pertz who kindly made the drawings from my plants.

1 Vegetable Teratology, p. 38, 1869.

2 I have intentionally refrained from using this term in the present instance in order to
avoid the implication that in the normal condition there are two distinct flowers which
become occasionally more or less fused together.

Fig. 3. Nearly mature fruit from
the lowest "flower " on the first
lateral branch of a third plant C
(much foreshortened). A dis-
tinct furrow occurs on both sides
in the dividing plane.

THE ANALYSIS OF THE RESULTS OF PROFESSOR
JOHANNES SCHMIDT'S DIALLEL CROSSINGS
WITH TROUT.

By H. L. TRACHTENBERG, BA. (Cantab.), A.I.A.

Actuarial Assistant in the Statistical Bepartment of the Medical
Research Council.

In the Journal of Genetics for December, 1919, Professor Johannes
Schmidt described an experiment on trout ^ The basis of this experiment
was his method of diallel crossings in which each female is crossed with
each male, and the character measured was the number of vertebrae.

In analysing his results, he sought for a " simple rule connecting the
number of vertebrae in the offspring with that in the parents." He first
distinguished between "the realized purely personal value of a given indi-
vidual trout — this value would have been a different one if the individual
in question were developed in different environments — and the generative
value of the same individual, and that is the value which it imparts to
its offspring." He then assumed that " the average for a number of
offspring-individuals closely coincides with the average of the generative
values of the parents," and proceeded to inquire whether this supposition
did agree with the values arrived at in the experiment. He obtained,
however, insufficient independent equations for a solution of the problem,
and had recourse to an arbitrary assumption to give him the necessary
additional equation. This arbitrary assumption is unnecessary. I will
demonstrate this, and indicate my alternative process.

On the basis of his original assumptions Professor Schmidt writes
down the twelve equations connecting the offspring averages with the
means of the generative values of the parents :

-^—=61-14, ^^-^— = 61-3o, -^- = 60-b5, etc.,

1 " Eacial Studies in Fishes. III. Diallel Crossiugs with Trout (.S'a^Hto TruttaL.)." By
Johs. Schmidt, D.Sc.,. Director of the Carlsberg Physiological Laboratory, Copenhagen.
Journal of Genetics, Vol. ix. No. 1, pp. 61 — 67.

7H Schnildt'^ Dlallel Crosshiys with Trout

where x, y, z, a, h, c, d are now understood to refer to the generative
values of the parents. These are written in the form of equations but
he indicates of course that exact equality is not expected, differences
between the two sides being allowed in the degree permitted by the
variation of the measurements in the samples. He therefore combines
them into seven equations, which are in fact the seven equations yielded
by applying the method of least squares. Unfortunately these seven
equations are not independent. They are equivalent to but six indepen-
dent equations and thus could be satisfied in an indefinite number of ways.
To obtain another equation Professor Schmidt assumed that the gene-
rative value of one of the parents {y) coincided with its personal value
(60). This is the arbitrary assumption referred to above. I surmount
the difficulty as follows. Seven equations which are all independent and
therefore do not require recourse to any arbitrary assumption can be
obtained if in writing down the original relationships, account be taken
not only of the relation of the offspring average measurements to the
average parental generative values, but of the parental individual
measurements to the parental generative values. In other words we
have simply to set out to find such values of x, y, z, a, b, c, d as will
satisfy the twelve offspring equations

*^ = «-14, 2^" = 61-35,etc.

and the seven parental equations

x = 59, y = 60, etc.,
such differences being allowed as are permitted by the degree of varia-
tion of the measurements. Since the sample is 50 in the case of the
offspring-averages, and but unity in the case of the parent measure-
ments, a greater divergence is permitted between the two sides of the
parent equations.

The equations duly weighted by VSO, prepared for the application
of the method of least squares are :

X = 59

2/

=

60

z

=

59

a

=

61

b

=

59

c

=

57

d

—

58

H. L. Trachtenberg 77

V50 -^ = V5() 61-14
V50 '^^ = V50 59-06
V50 ^-^ = VSO 58-29
V50 ^^ = VSO 59-03
V50^ = V50 61-35
V50 ^-|^ = V50 59-22
V50 y~^ = VSO 58-59
V50 ^-i^=V50 59-28
V50 ^-~ = V50 60-65
V50 ^^ = V50 58-48
V50 ^^ = V50 57-90

V50^^ = \/50 58-55.

The seven equations for solution yielded by these are :
51a; +12-5(a + 6 + c + fZ)=:5997
51y + 12-5 (ft + 6 + c + d) = 6021
51^ + 12-5(a + 6 + c + cZ) = 5948-5
38-5a+12-5(a; + 2/ + 2) =4639-5
S8-5b + 12-5 (x + y + z) =4478
S8-5c + 12-5(x + y + z) =4426-5
SS5d+12-5(x + y + z) =4479-5.

These are all independent and supply definite values of
X, y, Zy a, h, c, d.

78 Schmidfs Diallel Crossings ivith Trout

The fo]l(nving Table shows the theoretical figures against the actual.

Parents Offspring

Mean of

Mean of

Personal

Generative

personal values

Generative values

Value

Value

of group of 50

of Parents

59

5978

61-14

61-11

60

60-25

59-06

59-01

59

58-83

58-29

58-34

61

62-44

59-03

59 03

59

58-24

61-35

61-34

57

56-90

59-22

59-25

58

58-28

58-59

58-58

—

59-28

59-27

60-65

60-63

—

58-48

58-54

—

57-90

57-87

—

■

58-55

58-55

It will be noticed that like Professor Schmidt's results the average
values for the offspring are much closer to the means of the generative
values of the parents, than are the personal values of the parents to
their generative values. The most extreme individual difference is 61
personal value against 62*44 generative value. Though this is less
marked than Professor Schmidt's contrast between a personal value of
61 and a generative value of 62'72, it is still capable of bearing his
conclusion as to the marked difference that can occur between the
personal value and the generative value of an individual.

NOTES ON BREEDING FOR INCREASE OF MILK
IN DAIRY CATTLE.

By ELIZABETH ROBERTSON.

(With Eight Pedigree Charts.)

The following paper deals with the breeding of Dairy Cattle. The
subject is one of great complexity, and the methods suggested are, I am
aware, open to statistical criticism. Much more work is required before
the principles urged can be considered as fully established, but the
results are sufficiently striking to justify this statement in the hojje that
other breeders may be induced to try to improve their stock along similar
lines. The breed used is the Kerry, one of the remnants of the Celtic
cattle that are supposed in pre-Roman times to have covered the Conti-
nent of Europe and to have been gradually replaced by breeds from
elsewhere brought by the Romans. They are longer and narrower in the
skull and face as well as smaller than the Long Horns, Shorthorns and
Herefords. For size they stand midway between the Jersey and the
Ayrshire and obviously have far more affinities with the former than
with the latter. When used for cross breeding they are strongly pre-
potent. Being comparatively few in numbers they are apt to be inbred,
and being for the most part in uneducated hands the inbreeding has
been casual.

The contents of the paper may be briefly summarized as follows :

1. Inbreeding to a male relationship tends to increase both the
quantity and the quality of the milk produced.

2. Inbreeding to a female relationship tends to decrease both the
quantity and the quality of the milk produced, especially the quantity.

In Table I, the three tables of male inbreeding include 26 cases
of which five were failures in respect of quantity. One of these cows
(No. 109) failed to retain the milk in the udder ("ran out") but not
taking this into account she is included as though she had been a normal
cow. The results are (discarding the decimals) 80 /^ of cases in which
the cows showed an increase of milk when compared with the record of

80 Breeding f 07' Increase of Milk in Dairy Cattle

the dam and 20°/„ which showed a decrease in the yield of the cow when
compared with her dam.

In Table II are given the results : (a) of mixed inbreeding to both
male and female relationships, the male being the nearer and therefore,
presumably, the stronger. It will be noted that though there are only
10 cases the percentages showing increase of milk and decrease of milk
are respectively 80 "/^ and 20°/^, (6) mixed inbreeding with the female
relationship the nearer, these are all decreases.

Table III gives out- or chance-breeding experiments and its results.
There are 31 cases of these — only four were increases, i.e. 87 °/^ showed
a decrease in milk yield and 13°/^ an increase.

The pedigrees attached explain what is meant by related breeding.
It may be described as " the mating of a bull with a cow who is so
related to him that their first common ancestor is a bull " or in other
words, one or more bulls must be repeated in both the sire and dam's
pedigrees. These, then, become " links," i.e. linking bull or bulls and
make a malely related breeding.

A femalely related breeding is described in the same terms substi-
tuting cow for bull.

Full relationships are made when the same bull and cow give rise to
the animals " through" whom the link is carried on. It is interesting to
note that breeding to full relationships appears to produce a maintenance
of the dam's record in her female calves.

In Pedigree IV Gort Sheen </ is mated to Mona % in both pedigree
of sire and dam and the full brother and sister Vaddy Sheen cT and
Glenelly $ ultimately make a " full " first cousin once removed relation
between their descendants Burntollet II (/ and Glenelly II $ .

Pedigree V illustrates a " nearly full " relationship.

H, B, and 8 is an abbreviation for half brother and sister.
Avunc. is an abbreviation for Avuncular relationship, i.e. Uncle

and niece or Aunt and nephew.
1st once signifies first cousin once removed and so on.

The third lactation is taken as the standard year and most cows do
reach their full development then.

In Table IV sundry tables are given :

(1) The annual average of butterfat for the whole herd showing the
gradual rise as the herd became more and more " bred."

(2) A table showing the results of classifying the heifer by the rise
of the yield in milk in comparison with their dams' records in quantity
and quality.

E. Robertson 81

The difficulties of breeding for milk arc best illustrated by the fact
that no one has made any real study of the question and by the nature
of the " points " that are supposed to indicate a good milker.

Beef cattle have been wonderfully developed within the last century
but, if anything, the milk yields of cattle have declined ; they have
certainly not improved.

It does not seem to have struck anyone that while the eye, and the
" touch " or " feel " of the skin, were quite good guides as to the flesh
and fat that an animal developed, there was a better standard for dairy
cattle in the milk yield. But in order to utilise the yield as a standard
to which to breed, it is necessary to weigh the whole yield of every
cow and her heifers for some years. This means no little trouble and
attention. Moreover, it means steady trouble twice a day regularly.

The milk record kept for these experiments has been maintained
without a single break since the beginning of 1904. Not one single
milking has escaped record.

The butterfat estimations have been as regular since 1905. At first
tests were made for every milking, but as this was found to be too
expensive both in money and time, a sample is taken at every milking
for each cow and the combined sample is tested regularly every Friday
afternoon ; the Gerbers method being used.

This ensures as correct a record as is possible of each cow's perform-
ance in the year.

Nevertheless, the best kept record suffers from certain imperfections
that must be understood and, if possible, corrected. The amount of
milk given by any cow follows two curves, the first a seasonal one, due
to the time of year at which she calves, and the second a physiological
one due to the date of service. The same cow calving in September
will have a different seasonal curve from what she would have Avere
she calved in April. And again the curve of yield will be different
Avhen she is put in calf in the second month after dropping a calf to
what it is when put in calf in the third or any subsequent month.
How far these two curves contradict each other has not been fully
worked out. I have tried to do so, but once a cow has started work in
any given month, one does all one can to keep her calving dates in that
month so that real control experiments are rare. So far, I find that
they vary a good deal while following a general law more or less
closely.

A curve beginning in any of the late autumn months, say October,
never rises as high as one that starts in, say March, but it keeps at a

Journ. of Gen. xi 6

82 Breeding for Increase of Milk in Dairy Cattle

steadier level so that I believe that in the end the cows calving between
October and May ultimately yield just about the normal amount. On
the other hand, cows calving in July, August and September start their
physiological rise on the seasonal fall and I believe that their yield is
lessened. I am aware that the Irish Department of Agriculture states
the exact reverse but they do not give the evidence upon which they
base their view.

Temperature has a marked effect on yield. I have come to the con-
clusion that the optimum lies between 50" — 60° Fahrenheit. Wind
direction and intensity also affects yield as does sunshine: both are to be
avoided, though light is most desirable. The general character of the
season, whether wet or dry (the first is to be preferred), the nature of
the fodder or roots used, the skill of the milkers, the temper of the
byreman, all leave their mark on the milk record, and the effect of the
most transitory illness is at once visible in the amount registered.

One source of inaccuracy can be corrected more or less adequately.
A cow running twelve months neat between calves gives a normal
lactation, but if this period is shortened or lengthened the lactation
ceases to be quite normal. Here, however, we must remember the
following facts. Some cows will remain in milk for two years between
calves, some will go out of milk at the end of ten months or thereby
even if yeld.

Some cows if run dry will give more milk than they would have,
had they been put in calf, and some will give less. After much study
I have come to the conclusion that the proper correction is to take the
total lactation figure and divide it by the total of days between calving
dates and multiply the quotient by 365, This decreases the figure of
yield for the long lactations and increases it for the short ones. All the
figures of yield in these tables have been thus obtained. The butterfat
percentages are averages for the whole lactation.

This correction which may appear arbitrary to workers unfamiliar
with data afforded by the consideration of large numbers of records is,
I am convinced, a fair representation of the facts and it should be noted
that while it reduces the maximum range of difference in yield, it has
never occasioned the placing of good milkers in a low category nor the
reverse.

The curve of butterfat exactly reverses the milk yield curve, falling
as long as the milk curve rises and rising when it falls. I have only one
exception to this rule, a case where the butterfat curve, if drawn, would
make an almost straight line rising slightly in the second half of the

E. Robertson 83

lactation period. She appears in Table II as No. 105 with 5*9% of
butterfat.

In considering the facts disclosed in these tables, it has to be
remembered that all our beef breeds are inbred. But as the qualities
required are as often seen in the female as in the male, both sides are
bred to. I append on p. 90 the pedigree (VII) of the first noteworthy
shorthorn bull as a sample of breeding for beef It causes me no
surprise that highly bred shorthorns are notoriously deficient in milk.
Mr Taylor produced a well-known herd of " milking shorthorns " and
his pedigrees show inbreeding to male relationships.

The Jersey herd book is full of pedigrees inbred to the male though
there is a tendency to use full relationships.

In conclusion, I think that the results here set out afford prima
facie evidence of the direction of future breeding for an increase in milk
production.

G— 2

84 Breeding for Increase of Milk in Dairy Cattle

9^ Q

S; S c8

^ § ^

S 03 c3

. — . a =»-■

CO "^ «J

^ cl 5b

to 23 ^<

~ '.00

o '^
o a

S5 a oPh

.22 2 ^

iX!

SPS5 at!
^ $ S'-'

fTS > 0) (U

£ »^ S
S _ '^ to

<X) ,o <u

"q. 6C CO Ph

a ? sa

I

«^

^

(J

pq

K5

.^.Q'-l

te I

n

o o

/45

^ ^-s

iH S -S -"S fi

J2

pq

pq

oi
•gfi

/S

1.

O'

II OWIiMlNOOii tH(Mift'M

II r-l "H i-l 11 (N i-H (M

+ +

6j tH

+ + + + I + + + + ^ + I I

M 5 ^

" o

8 M

w a

?D t- Cq O 05 05

IN (N CO IN 0-1 CO

p CO OS p CO p
Tt^ CO CO ""^ CO ■^

O O U5 o o o
O lO 00 o »c o

U5 U5 '^ lO CO lO

^- CO

' CO 00

CO OI OI p o t>
CO ^ ^ "^ "^ "^

IN o o o o CO

CD O O 00 00 ^
-rH o CD O »0 CO

■* IN p >p p p

■*■*-* TjH CO -tH

O CO o o o o

CO CO CD O -^ CO
CO CO CO -"^ 00 CO

p t^ p 05 >rt IN

t- CO o t~ o o

o i-H >ra lo o o

lO CO CD CD lO lO

iC OS CO to 1-1 Ol (N ■<1< ■* IN IC CO

C-CDI>00COCD O t~ 00 00 05 i-H

tH ,-1

!z; lo

O p

o o

00 o
«5 OS

a.s

-fl CO
O c8

;> OP>

>>

I + + + + + +

+ + + + +

CO o c~

p »p p 00 00 p p

CO ■<*( CO CO CO CO eo

o o iH o o >o w

O O IN lO O 00 00
lO 1* CO CO CD Tft ■>*<

p p p 00 p ^- CO

CO IC -^ CO "* Tt< ■*

o o o o Tfi '^H a.
CO t- O O ira CO o

CO Ol O I— CD lO SD

p p p

T)H 4n 4*1

H o -S
«i s a
<B a rj

a t^ r2

r CO a

T3 rQ "^

"73 "^ '^

eg a,

o IB a
o > o

O 3

3 8
2-q

CO

m'^ Sag
K fl > o <a

fl ^ S -a
'^ a ^

+ + +

■* ■* oi

+ +^5 §+ +

00 CO !*
tH 00 II

Oi »p -JH
IN CO -^

O O O
lO 00 lO
lO <N «0

CO 05 U5
CO «5 *«

o ^ o

ira r-l lO

to »o o

--^ IT»-

03 P _i

a o3 ^5

'« « -w

13 M S-i

03 O O

00 o

E. Robertson-

85

^l^ c8

80:

^3

' -2 .B<

"cs S

-5 a

a a 2

O «i5 O

Oc-t <o

'w . ^ 00

5- e3 e8 ^

a c j5 t.

2! ►*

r^ r^ ^ OJ

.2 ea.2.2

pa

•<
^

i

^

r
•-i

'^

is -
> S^

o o

Ha

i.S.2 » e8

■— I f O)

«.

s

oS-^

tia

rr

g

o

pel

^a

+ + + + + + + +

jo ^ a „

> _i_ .-^ 03 " 1^ — ip ZO

»

+ + + 4 ■ + +

+ + I

i-H O "* '■'^ lO o
■-I rH i-H rf -# rH

ff5 0>C«5000505

TjiO-*-#0OOO
1-1 >0 QO 00 03 lO >o o

■^ ^* CO '^ lO "^ ■^ ^

OioOoOOOOOO
0«000«5>0|>.IOC-01

o o S

^ ^ I

a

5 o -g

-. --; a>

O

pq

> > )4

'^

^

8

'S

+

+

+

+

1

+

CO

OS

ilo

00

>c

>ra

o

CO

1-i

CI

CO

1-1

1-1

1

1
ip

1
'P

1

1

CO

1

C5

o

t-

ifi

t-

«

ao

CO

1—1

i-i

i-\

r~l

C5

o

cp

-it

"*

V

CO

-^

CO

■^

■*

■^

o

o

o

O

o

'M

o

w

CO

CO

Ui

1-1

t-

?o

o

>o

>o

CO

co

CO

00

I-I

CO

1

CO

«>

«o

00

•<*

1

>p

05

05

fH

o

CO

TtH

U5

■^

>o

Tf<

■*

00

o

1-1

CO

o

«o

■^

o

«5

"*

CO

OJ

'*<

o

O

-^

»o

-*

CO

\a

CO

o

1-1

o

o

O

r-<

00

1

1—1

1-1

1-1

r-t

r^

%

o

o

^

^

*

T(t

o

'

z

o
o

OS

:

=

r

o

00

1

fl

.

o

I— 1

TS

<—l

m

n3

•K

M

M

<D

<o

c3

•*

^

•»

q

W

S

0)

fl

>

o

:;

:;

;

^

<

c3

86 Breeding for Increase of Milk in Dairy Cattle

TABLE III.

Out or " Chance " Breeding. The Bulls are Pedigree Animals, the Cows are not.

Record of Bull s
Dam

Number

of
Heifer

Heifer's Record

Number

in
register

Record of Heifer's
Dam

Increase or decrease iu Heifer's

record as compared with that

of her Dam

Siring Bulls

Gallons

Butterfat

in

register

Gallons

1-
Butterfat

of Heifer's
Dam

Gallons

Butterfat

Gallons

°L + or -

Butterfat

7o + or -

Gort Desmond . . .

Not known

12

333

4-2%

1

676

4-3%

50%-

2-3°/o-

J) 5>

)>

»j

13

280

3-5

3

596

3-5

53 -

= with her
Dam

Vaddy Erne

900

4-0°/o

26

580

4-0

2

718

4-2

19 -

5 -

> > 9 9 " '

—

28

500

3-9

9

620

4 1

19 -

5 -

Vaddy Erne III...

580

4-0

66

700

3-9

34

600

3-7

16 +

4-4 +

Gort Gallant

Not

known

40

550

4-4

10

460

4-0

20 +

10 +

,,

,,

41

437

4-2

14

530

3-9

17 -

7-6 +

J ) 5 ? • ■ •

,,

43

550

4-4

4

900

4-0

31 -

10 +

J,

45

511

4-4

16

750

4-0

31 -

10 +

5 J ) t ■ • •

jj

46

644

3-7

9

620

4 1

4 +

12 -

99 99 • ' •

)>

38

500

3-9

2

718

4-2

30 -

7 -

Dromyrourke . . .

14

530

3-9

8

545

3-8

2-3 -

2-6 +

Gort Sheen III ...

j^

39

600

3-8

9

620

4-1

3 -

7 -

,,

J,

54*

500

4-0

9

620

4-1

20 -

2-5 -

1) ,1

,j

67

600

4-0

9

620

4-1

3 -

2-5 -

J) >.

J,

89

640

4-7

9

620

4-1

3 +

14 +

)i

jj

30

500

3-9

2

718

4-2

30 -

7-6 -

) 5»

jj

52

485

3-9

2

718

4-2

32 -

7 -

).

,,

31

500

3-5

14

530

3-9

5-6 -

11 -

\

) )

53

500

3-3

8

543

3-8

7-9 -

12 -

99 99 • ' •

51

550

3-8

11

585

4-5

5-8 -

18 -

'

,j

37

500

3-8

15

788

4-0

36 -

5 -

,,

60

500

3-9

22

750

4-1

40 -

5 -

>) ) J • • •

, J

36

650

3-3

18

800

4-0

23 -

26 -

Vaddy Sheen . . .

585

4-5

48

580

4-0

15

788

4-0

36 -

= with her
Dam

M >1

—

—

49

600

4-2

22

750

41

20 -

- 2 +

Vaddy Moile

900

4-0

56

490

4-0

23

570

3-9

10 -

3 +

99 99 * • •

—

—

57

550

4-0

11

585

4-5

5-8 -

12-5-

5 ) ») ■ • •

—

—

59

456

3-9

20

600

4-1

25 -

5 +

Vaddy BurntoUet

900

4-0

93

488

3-7

16

750

4-0

35 -

7 -

) J M

900

4-0

76

570

3-9

34

600

3-7

5 -

5 +

Averages

—

—

—

528

3-9

—

655

4-03

19-4 -

3-2-

* See Pedigree VI.

TABLE IV.

(1) Annual Avei'aye of Butterfat
for the whole herd.

1905

4-i7o

1906

3-9

1907

4-05

1908

4-04

1909

3-97

1910

3-97

1911

3-94

1912

3-99

1913

4-0

1914

4-04

1915

3-97

1916

4-33

1917

4-94

1918

4-85

1919

4-85

(2) Percetttage increases of sticcess-
fdly bred Heifers comjjared with
their Davis, expressed in averages
for the whole herd.

Yield of Dam

Increase per cent,
in Heifer

Under 300

Between 300 and 400
400 and 500
500 and 600

Over 600

77% +
47 +
16 +
14-8 +
7-6 +

Yield of Dam in

Butterfat

Increase per cent
in Heifer

3 -3-5 °/o
3-5—4

4 —4-5
4-5-5

13-2% +

23 +

8-6 +

8-47 +

E. Robertson

87

Pedigree No. I to illustrate Half-hrother and sister (= HB and S) mating.
Also a "normal" mating, i.e. one passing through a bull and a cow sired
hy the same hull.

Gort SheCii III

Vaddy Sheakin

Vaddy Glenelly

II

(heifer)

oi

I
I

Dew Beauty

05

Oi

Vaddy Gleiielly

'i

Gort Sheen HI

Mona

Vaddy Cusher^

fir I

(heifer) -^

Warrington II

Walton
Madden

Madden

Walton Fame

Vaddy Awe

Kilmorna
Duke 22nd

/

C. L Minnie

Vaddy Gusher
11

Duke 22nd

"I— cil

r

Vaddy Gusher

Pedigree No. II to illustrate avuncidar mating and also shoiving the link
passing through two cows daughters of the same bull.

88 Breeding for Increase of Milk in Dairy Cattle

Pedigree III illustrating first cousin tnating and a second cousin once removed,
therefore showing a case of multiple inbreeding to the male.

Walton .
Madden

Vaddy Bredagh

IV

(heifer)

Warrington
III

V.Belfast

Vaddy Bredagh
III

Burntollet II

Vaddy Bredagh
II

Gort Sheen III

0. Buttercup

&I S^ '0. Ma

rro^'"^

Dromyrourke I t-I'^^-oi'^9-

idenhair >,

?;>°:

bu'

S I s
Burntollet

Vaddy Sheen k

Gort Sheen III
N Mona

Z

Pride

■aj-

I "Si Dew Daniel
C. L. Daisy 8 _

5 I w 1 1 Cahir Beauty

Gort Sheen III

Gort Sheen

Ellengrane

Vaddy Bredagh

Gort Sheen III

Dew Johanna

Dromyrourke
Cahir Biddy

Vaddy

Burntollet

IX

(bull)

.a. Burntollet II

to

I §

*- s

U Co

§2

^ Glenelly II

Burntollet

Vaddy Sheen

Gort Sheen III

.4_

/Mona/
/ ^ /

^

Pride

U-^

I

I I
■t—r

C. L. Daisy

Dew Daniel '
/

u

"Dromyrourke

■7~7

/ l/ /

Cahir Beauty H — f-

/ ■' '■

Sheskin

1^ ^'/| ^
Gort Sheen III U^

'/I

7^-

^1

Dew Beauty ,[■

M-

Glenelly

Gort Sheen III

/
J!

Mona

Pedigree IV illustrating a "full relationship " tvith bull repeated in the

cow's pedigree.

E. Robertson

89

Pedigree V illustrathtg a " nearly fulV relationship that seems to be of imich
the same value as a quite '•''full " relationship.

wenl);iliii J II

BurnU.llet IX

Burntollet II

Burntollet I

XT \\ c^\. 1 ^"'Tl Rhei^ti III

1 Mona T

Pride 1 ' '

t ' '

C. L. Daisy

'" ■ i ' /

1 1 /

1 1 /

Glenclly II

She.sk in

^ \ 1 /^

^'T^' 1 1^

Glenelly

1 K?

()

1 , /-■

1 n^

Qwenbahn II

Walton
Madden

Madden

,5 1 1/

XI f

31

.5 1 /I

Walton Fame

^1 / 1

' / '

' / 1

' 1 '

Owenkilloe II

Vaddy Moile

1 *^ 1

Pride 1

V. Owenkilloe

a:i„-_ (

Eden beg II
No. 137

Warrington V

o

s

Walton
Madden

Madden

W. Fame

Owenkilloe II

Moile

Goi L hlieeii 111 |

Pndc 1

Owenkilloe

(joit aiieeli 111 1

■71

S
o

Edeiibeg ~
No. 109 1

GO

V. Shannon °3'

t5

Gort Sheen III

\

Morna XIII

ri . r.1 1

N

V. Eden 03
No. 83 ^

V, Sheskin

r ;^-i TIT 1 Gort Sheen

>f

T

V. Eden

No. 54

^ 1 Gort Sheen

G Fortune 1

No " I

Pedigree VI an unsuccessful strain so far. iVb. 109 ran her milk out so that no real
record could he kept. No. 137 Aas calved recently; she is more promising.

6—5

90 Breeding f 07' Increase of Milk m Dairy Cattle

Pedigree VII. Shorthorn Pedigree quoted from Wilson' s'''' Evolution of British
Cattle." There are six generations and only three hidls : Foljambe, Dalton
Duke and the Sire of Lady Maynard, and three cows : Lady Maynard and
the dams of Foljambe and Dalton Duke.

Favourite

\
\
\

Lord
Bolingbroke

Foljambe

^1

Young
Strawberry

Dalton Duke

Lady Maynard

Phoenix

I

Foljambe

/

/

/

/4

/N?

/^'

Lady Maynard

\

Young
Phoenix

\
\
\

^ .

Favourite

Lord
Bolingbroke

Foljambe
1

Young
Strawberry

Dalton Duke

Phoenix

/
/

J.

I

Foljambe

Ladv Maynard
'—=^

/

Lady Maynard

/

Phoenix

/

Foljambe

Lady Maynard

KOOT-CUTTINGS AND CHIMAERAS 11.
By W. BATESON, M.A., F.R.S.
(With Plates XIII and XIV.)

Bouvardia.

In a former paper^ I recorded the production of the red-flowered
variety Hogarth from the roots of Bouvardia Bridesmaid which has
pinkish white flowers. This behaviour is perfectly consistent. At various
times we have raised from Bridesmaid 29 root-cuttings (besides others
not counted), which in every case bore Hogarth flowers. Very rarely a
streak or flake of pinkish white has occurred on these flowers such as
I have seen on other varieties of Bouvardia (e.g. Cleveland and Lemoinei)
but they are exceptional. Perhaps one or two such flakes are seen in
a season among our collection, and they do not come with any special
frequency on the derived Hogarths.

From Hogarth ex Bridesmaid's roots 64 root-cuttings have been raised,
all Hogarth except one luhicli was Hogarth colour but single- jioiuered,
whereas Hogarth and Bridesmaid are fully double. This single had
anthers which dehisced containing much pollen, in microscopical appear-
ance all bad. Singles arising as root-cuttings from double Bouvardias
have often been recorded in literature, especially from Alfred Neuner,
but this is the only instance observed here. A. Neuner has only given
us three root-cuttings which have flowered as yet, both double, exactly
like the parent plant.

From Hogarth ex Hogarth ex Bridesmaid 63 normal Hogarth have
been raised as root-cuttings and from these again have come 13 normals.

From the single Hogarth {ex H. ex B.) we have raised 6, all singles
like the immediate parent.

The following kinds have given root-cuttings which on flowering were
exactly like the parent:

A. Neuner, white double (3).
elegans, scarlet single (15).

1 Journ. of Genetics, 1916, Vol. vi. p. 75.

92 Root-Ciitt'mgs and Chimaeras II

Humboldtii, white single, glabrous (23).
jasrainiflora, whitish pink (1).

King of Scarlets, scarlet single (12); and from these again (7).
Lemoinei, red double (25).
Pres. Garfield, pinkish white double (2).
Priory Beauty, rose single (3).
The Bride, white single (2).
Vreelandii, white single (10).
leiantha, a species, scarlet single (9).

and from a seedling, scarlet single, produced by the late Mr Allard
from Cleveland x leiantha (2),

Vulcan, a scarlet single, short-styled, gave 27 root-cuttings like itself
and one rose-pink single^.

President Cleveland, scarlet single, long styled, gave 23 true to type,
and 8 peculiar, having flowers scarlet marbled with white. These flowers
were rather smaller and looked unhealthy, but for several seasons they
have kept the same character, and the plants seemed in normal health.
The margins of the petals are scarlet, and the marbling consists of whitish
lines irregularly distributed over the petals, coalescing towards the central
and basal parts. In the streaks the substance of the petals is somewhat
thinned, a feature which gives them a depauperated appearance. Our
stock of Cleveland came from two commercial sources, and the two lots
were not kept distinct. It is therefore doubtful whether both kinds of
root-cuttings came from one plant in this case.

From roots of true Cleveland ex Cleveland 21 were raised all true;
and from them again 7 all true.

From roots of marbled root-cuttings of Cleveland we had 16 marbled,
11 with slight and irregular traces of the marbling, and one not marbled,
being true Cleveland.

A variegated form of Humboldtii, having a green core and a pale
yellowish green cortex, gave 2 root-cuttings green, as might be expected.
These have not yet flowered.

In style-length the root-cuttings have always agreed with the parent
plant, shorts coming from shorts and longs from longs.

Pelargoniums (Fancy).

Fancy and especially "Regal" (i.e. semi-doubles with petals crumpled)

Pelargoniums have been said to give root-cuttings with flowers differing

1 Though this "sport "came alone there is no reason to doubt that its origin was
correctly recorded. We had no similar variety.

W. Batbson 93

from those of the parent plant. This has been observed in three
cases.

Escot (PI. XIII, fig. 1) has flowers white with a large purplish red
blotch on each petal. It is characteristic of the variety that the petals,
especially the two dorsals, roll back more or less. A plant gave off a natural
"sucker" (from a root) which bore flowers as shown in PI. XIII, fig. 2. They
are larger than type (70 mm. across the largest flower against 62 nmi. in
the parent); and flat, showing no disposition to roll back. The peripheral
areas of the petals are pinkish, not white, and the blotches much redder
than those of type, a feature insufliciently i-endered in the coloured
figure as reproduced. A root-cutting has since been raised from Escot,
and its flowers agree with those of this sucker. The rolling-back of the
petals in the parent is perhaps due to strain produced by the greater
size of the flower proper to the included "core." The root-cuttings are
somewhat taller than the type.

Mrs Gordon has flowers white and pink as shown in PL XIII, fig. 3,
with guide-marks of the dorsal petals only lightly represented. Three
root-cuttings are all alike (PL XIII, fig. 4), have much more colour,
a full pink, on all petals, and in addition deep crimson guide-marks.
These root-cuttings are very like and probably identical with the variety
called "Cardiff." Both type- and root-form^ may have more colour than
appears in the figures, especially in newly opened flowers, but the
relative amounts of colour are correctly represented. In Mrs Gordon the
guide-marks are only distinguishable as " ghosts."

Pearl is a white semi-double Regal (PL XIII, fig. 5) having small and
evasive purple patches in the area of the guide-marks. Sir W. Lawrence
kindly gave me a plant of Pearl having a large branch with flowers
heavily marked with red (PL XIII, fig. 6) much as in the varieties kno^vn
as Mme Thibaut and Emmanual Lias (? synonymous). Three root-cuttings
raised from Pearl all have exclusively flowers of this coloration. Doubling
in parent and the root form is similar in degree. Pearl itself has been
grown here on a fairly large scale. At various times three flowers have
been produced by it with a patch of red as shown in PL XIII, fig. 7. These
patches must be regarded as indicating a break through of the under-
lying tissue, like the green patches so often seen on the leaves of some
variegated chimaeras composed of a white cortex overlying a green core.

The three examples mentioned are the only cases in fancy Pelar-
goniums of root-cuttings differing from parent plants up to the present.
Numerous varieties have been tried, but at first we were not very
successful in raising such plants. The technical difficulties have now

94

Root-Cuttings and Chimaeras II

been largely overcome, and we have a large series which will flower in
due course. Plants identical with the parent have been raised from
Lady Doreen Long (1); Queen Alexandra (1); Kingston Beauty (several);
and from Touchstone, a scented-leaved Pelargonium (12).

Pelargonium (Zonal).

Though no root-cutting has yet been raised from it the properties of
a salmon, fringed zonal raised by Messrs Jarman and by them called
"Golden Flame" (PL XIV, fig. 1) should be considered here. The leaves
of this variety are shiny, stiff and crumpled or buckled. Its flowers have
laciniated petals and are devoid of functional female organs, the pistil
being reduced and the ovary aborted. The anthers are normal, containing
abundant and good pollen. The plant is therefore a male. Upon these
plants not rarely a shoot arises which has leaves flat, with a dull matt
surface like that of most leaves of zonals. On such shoots the flowers
are normal hermaphrodites, and the petals are entire. Occasionally leaves
otherwise of the characteristically crumpled kind have areas, small or
large, of the sport variety.

Miss Cayley has supplied the following notes on the differences in
structure between the two kinds of leaves:

Macroscopic Characteristics of Leaves.

Type (shiny)

Sport (dull)

1.

Surface, upper

Shiny

Dull

ditto lower

Dull

Dull

2.

Size of leaf

Smaller

Larger

3.

Shape

Somewhat ivy-leaved :
crumpled

Typical zonal leaved : flat

4.

Lobes of leaf

More acutely pointed

Bounded and more obtuse

5.

Substance of leaf

Stiff

Soft

6.

Hairs, upper surface ...

Numerous, but fewer than

More numerous than on the

on the " sport"

"type"

Somewhat erect

Markedly bent towards the

Thinner than the sport,

periphery of the leaf

but slightly longer

Somewhat coarser

7.

Hairs, lower surface ...

Somewhat erect

Not so markedly bent, much
the same as on the " type "

8.

Petiole

Stiff

Less stiff

Shiny

Dull

Somewhat shorter

Longer

Hairs less coarse but erect

Hairs coarser but erect

Microscopic Characteristics of Leaves.

Type

Sport

1.

Cuticle

4 — 5 IX thick

2 — 4 /i thick

2.

Surface of upper Epi-

Flat. In section the out-

Lrregular : in section outline

dermis

line is straight except
at bases of hairs

wavy

3.

Surface of lower Epi-

No clear difference : somewhat wavy in outline

dermis

W. Bateson 95

T A B LE — continued.

Type Sport

4. Epidermal cells: upper Small: tcith both upper li&Tgev.nothickeniiifjofloioer

and lower cell-ioalls cell-ioalls

thickened

5. ditto lower No clear difference

6. Palisade cells Elongated in vertical Shorter, wider, less crowded

plane : compressed to- and less compressed : more

gether, and regular in irregular in shape
shape

7. Layers of palisade cells Variable. 1 — 3 layers Less variable : mostly one

layer, occasionally more
than one

8. Spongy parenchyma... No clear difference

9. Bases of hairs ... Raised More raised than in type

When shiny and dull areas exist in the same leaf all stages of tran-
sition from the typical form to the other occur very irregularly. The
change from thick cuticle to thin cuticle can be fairly abrupt, but the
change in the shape of the palisade cells is more gradual and very
irregular. A few "sport" palisade cells can occur in areas of "type" tissue
and vice versa, and the differences are not so clearly defined in the in-
ternal tissues as might be expected from the external appearance of
the leaf The fixation of the transitional parts is difficult, and no good
microtome preparations have so far been obtained. The two types of
tissues react differently with the same fixative; the "sport" tissue is
penetrated more rapidly than the "type" tissue, and hence is fixed sooner
than the "type."

The shiny appearance of the short leaf is possibly due to :

(1) Thickness of cuticle.

(2) Flat surface of the epidermis.

(3) Fewer hairs which are more erect and somewhat finer as com-
pared with the bent hairs of the sport.

The darker green appearance of the same leaves is probably due to:

(1) Longer, more densely packed palisade cells.

(2) Extra layers of palisade which occur somewhat irregularly.

D. M. Cayley.

The inclusion of cells belonging to either form within the tissues of
the other is especially remarkable.

96 Root-Cuttings and Chimaeras II

Spiraea ulmaria.

A variegated form has stems, petioles, and central parts of the leaf-
lets devoid of chlorophyll (PI. XIV, fig. 2), and of a whitish yellow colour.
Ordinary green plants are perfectly fertile on both male and female sides,
but this variegated plant is quite sterile, forming no seeds or pollen,
A few ill-formed carpels have been found on it, but the seeds they con-
tained were aborted and did not germinate. The condition is closely
reminiscent of the zonal Pelargonium, "Freak of Nature," mentioned in
Jour. Gen. 1919, viii. p. 97, note, which has green borders to leaves and
stipules and is totally sterile on both sides. The extraordinary feature
of that plant is that the green, white, and green-over- white shoots which
Freak of Nature often produces are perfectly fertile. The variegated
Spiraea has not hitherto produced any shoots other than those described.
From its roots it readily gives rise to adventitious buds, and all leaves
borne by them are albino, quite destitute of chlorophyll, like the stalks
and petioles.

In my previous article on root-cuttings I spoke of the dissimilar forms
which arise as being in all probability included as "cores" within a
cortex of the ostensible type. The whole plant is thus regarded as a
periclinal chimaera, one variety enclosing another, and this enclosed form
may be expected to come out whenever the plant makes an adventitious
bud by endogenous growth. Though this view is presumably correct in
most cases the distribution seen in the variegated Spiraea and Freak of
Nature show that other possibilities must be remembered. For in these
plants the white tissue is not covered in, but extends through the whole
of the internodes, and doubtless the root also. The growing point alone
carries up with it the power of making green tissue. In such plants as
the Bouvardia or Pelargoniums which give dissimilar root-cuttings the
two kinds of tissue are not recognizably distinct in the plant until they
flower, and though perhaps unlikely, it is not impossible that the kind
which arises by adventitious buds may really provide the whole of the
root and perhaps the internodal regions. Many herbaceous variegated
plants arranged periclinally are liable to give shoots composed entirely
of either their external or their internal constituents. Such shoots with
special frequency arise near the base of the plant, i.e. just above the level
at which the stem was divided in propagation. Though their mode of
origin is not always easy to decide, it must be supposed that they are
generally produced by adventitious buds. These evidently are not always

JOURNAL OF GENETICS, VOL XI. NO.

PLATE XIII

JOURNAL OF GENETICS, VOL. XL NO. 1

PLATE XIV

Fk'. 1.

Fig. 2.

Fig. 3.

W. Bateson 97

endogenous but may be originated by a periclinal division in the cortical
layer.

In such an example as the crumpled zonal here described, or indeed
in any periclinal chimaera with patches of its core coming out on the
surface of leaves, the jjrocess by which this change is brought about is
very difficult to imagine and I do not know how it may be represented.
The growing point must contain both elements, but the emergence of that
which is normally enclosed seems at present to be purely fortuitous.

As regards reversals of the layers, such as I described in Jour. Gen.
VIII. 1919, p. 94, it is worth observing that though we have now fairly
numerous cases of white-over-green turning to green-ovcr-white, a change
which in some plants happens frequently, we have not hitherto seen a
single instance of the contrary. Of Euonymus, some zonals, an ivy-leaved
Pelargonium, and Arahis we have several large and well-grown plants
of the green-over-white kinds, but though occasionally the white has
come to the surface in a small area, no reversal has been found on such
plants. Possibly we may regard white-over-green as an arrangement
mechanically less stable than green-over-white.

Several attempts have been made to breed the root-cuttings with
their parent plants, an experiment offering attractive possibilities, but
we have hitherto been unsuccessful.

EXPLANATION OF PLATES.

PLATE XIII.

Fig. 1. Pelargonium Escot.

Fig. 2. Boot Cutting of Escot.

Fig. 3. Pelargonium Mrs Gordon.

Fig. 4. Boot Cutting of Mrs Gordon.

Fig. 5. Pelargonium Pearl.

Fig. 6. Root Cutting of Pearl.

Fig. 7. Flower with red blotch, a form occasionally seen on Pearl.
This plate is from drawings by Mr C. H. Osterstock. In the photograph the colours
of Escot and the two forms of Mrs Gordon are approximately correct. In Escofs root-
cutting the red should be distinctly brighter, without any bluish tinge, and the same applies
to the red colour in figures 6 and 7.

PLATE XIV.

Fig. 1. Zonal Pelargonium Golden Flame. On the right is seen the foliage of the type,
which is buckled and shiny. On the left the sport, with leaves flat and dull. The two
right-hand inflorescences are laciniated. The two left-hand inflorescences bear mixed
flowers, some entire, some laciniated. The fruits formed in two entire flowers are
■visible. These mixed inflorescences belonged to the area of transition.

Fig. 2. Spiraea ulmaria: leaf of variegated form.

Fig. 3. ditto leaf of normal green form.

CAMBRIDGE UNIVERSITY PRESS

Devonian Floras. A Study of the Origin of Cormophyta.
By E. A. Newell Arber, M.A., Sc.D. With a Preface by
D. H. Scott, F.R.S. Demy 8vo. With a frontispiece and 47
text-figures. 17s 6d net.

Water Plants. A Study of Aquatic Angiosperms. By
Agnes Arber, D.Sc, F.L.S., Fellow of Newnham College, Cam-
bridge. Royal 8vo. With a frontispiece and 171 text-figures.
31s 6d net.

The Origin of Man and of His Superstitions. By

Carveth Read, M.A., Lecturer on Comparative Psychology
(formerly Grote Professor of Philosophy) in University College,
London. Demy 8vo. IBs net.

The Influence of Man on Animal Life in Scotland.

By J. Ritchie, M.A., D.Sc, F.R.S.E. Royal 8vo. With 91 illustra-
tions, including many original photographs and drawings, and
8 maps. 28s net.

The Concept of Nature. Tarner Lectures delivered in
Trinity College, Cambridge, November, 1919. By A. N. White-
head, Sc.D., F.R.S. Demy 8vo. 14s net.

An Enquiry concerning the Principles of Natural
Knowledge. By A. N. Whitehead, Sc.D.. F.E.S.
Demy 8vo. 14s net.

An Introduction to the Study of Cytology. By L.

Doncaster, Sc.D., F.R.S., late Derby Professor of Zoology in the
University of Liverpool. Demy 8vo. With 24 plates and 31
text-figures. 21s net.

The Determination of Sex. By L. Doncaster, Sc.D.

Demy 8vo. With 23 plates. 10s net.

Mendel's Principles of Heredity. By W. Bateson,

M.A., F.R.S., V.M.H., Director of the John Innes Horticultural
Institution. Third impression with additions. Royal 8vo. With
3 portraits, 6 coloured plates and 38 figures. 15s net.

The Methods and Scope of Genetics. An Inaugural

Lecture delivered 23 October 1908 by W. Bateson, M.A., F.R.S.

Crown 8vo. 2s net.

Growth in Length: Embryological Essays. By

Richard Assheton, M.A., Sc.D., F.R.S., Trinity College, Cam-
bridge. Demy 8vo. With 42 illustrations, 2s 6d net.

Cambridge University Press

Fetter Lane, London, E.G. 4 : C. F. Clay, Manager

CONTENTS

All Rights reserved

PAGE

B. MiYAZAWA. Studies of Inheritance in the Japanese Convolvulus.

Part II. (With Plate I (coloured)) 1

A. F. Blakeslee. a Graft-Infectious Disease of Datura resembling a

Vegetative Mutation. (With Plates II— VI) 17

R. C. PuNNETT and the late Major P. G. Bailey. Genetic Studies in

Poultfiy. III. Hen-feathered Cocks. (With two Text-figures and

Plates VII— XI) 37

J. A. S. Watson. A Mendelian Experiment with Aberdeen- Angus and

West Highland Cattle. (With Plate XII) 59

Edith R. Saunders. Note on the Evolution of the Double Stock

(Matthiola incana). (With Three text-figures) .... 69
H. L. Trachtenberg. The Analysis of the Results of Professor Johannes

Schmidt's Diallel Crossings with Trout ...... 75

Elizabeth Robertson. Notes on Breeding for Increase of Milk in

Dairy Cattle. (With Eight Pedigree Charts) .... 79
W. Bateson. Root-Cuttings and Chimaeras. II. (With Plates

XIII (coloured) and XIV) . , 91

The Journal of Genetics is a periodical for the publication of records of original
research in Heredity, Variation and allied subjects. The Journal will also, from time to
time, contain articles summarising the existing state of knowledge in the various branches
of Genetics, but reviews and abstracts of work published elsewhere will not, as a rule, be
included. It will be issued in parts as material accumulates, and a volume will consist of
three such parts.

Papers for publication may be sent either to Mr Bateson, Manor House, Merton,
S.W. 19, or to Professor Punnett, Whittingehame Lodge, Cambridge. Other com-
munications should be addressed to the University Press, Cambridge.

Papers forwarded to the Editors for publication are understood to be oft'ered to the
Journal of Genetics alone, unless the contrary is stated.

Contributors will receive 50 separate copies of their papers free. A further 50 may be
had at cost price : these should be ordered when the final proof is returned.

The Subscription price for a volume is 42s. net (post free) payable in advance ; separate
parts 15s. net each. Subscriptions may be sent to any bookseller, or to Mr C. F. Clay,
Manager, Cambridge University Press, Fetter Lane, London, E.C. 4.

Volumes I — X (1910 — 1920) are now ready. The price of each volume, in fom- parts,
stiff paper covers is 42s. net. Single parts can be supplied at 125. Qd. net each. Quotations
can be given for bound copies of these volumes, as well as for supplying binding cases and
for binding subscribers' sets.

The Cambridge University Press has appointed the L'niversit^^ of Chicago Press agents
tor the sale of the Journal of Genetics in the United States of America, and has authorised
them to charge the following prices: — Annual subscription, $8.50 net ; single copies,
$3.00 net each.

SPECIAL NOTICE. To effect economies in binding and postage a volume will, in
future, be issued in 3 parts instead of in 4. The size of the volume will however remain
asjsefore.

The complete set of volumes I— X will be supplied at £1Q. 16s. Od. net,
until July 1, 1921. After that date the price will be £2\. Os. Od.

CAMBRIDGE: PRINTED BY J. B. PEACE, M.A., AT THE UNIVERSITY PRESS

Vol. II, No. 2

September, 192 1

JOURNAL OF GENETICS

EDITED BY

W. BATESON, M.A., F.R.S.

DIRECTOR. OF THE JOHN INNES HORTICULTURAL INSTITUTION
AND

R. C. PUNNETT, MA., F.R.S.

ARTHUR BALFOUR PROFESSOR OF GENETICS IN THE UNIVERSITY OF CAMBRIDGE

CAMBRIDGE UNIVERSITY PRESS

C. F. CLAY, Manager
LONDON : FETTER LANE, EC. +
also

H. K. LEWIS & CO., LTD., 136, GOWER STREET, LONDON, W.C. 1

WHELDON & WESLEY, LIMITED, 28, ESSEX STREET, LONDON, W.C. 2

CHICAGO : THE UNIVERSITY OF CHICAGO PRESS

(Agent for the United States and Canada)

BOMBAY, CALCUTTA, MADRAS : MACMILLAN & CO., LTD.

TOKYO : THE MARUZEN-KABUSHIKI-KAISHA

Price Fifteen Shillings net

Issued d October^ 1 921]

THE UNIVERSITY OF CHICAGO PRESS

The Origin of the Earth. (The University of Chicago
Science Series.) By Thomas C. Chamberlin, Head of the Depart-
ment of Geology in the University of Chicago, xii -f- 272 pages.
12mo. Cloth, lis 6d net.
This book, by one of the leading geologists of the world, sets forth the disclosures
that led to the rejection, one after another, of the older views of the origin of our
planet, the futile attempts then made to amend these or to build others upon the
same foundations, the final rejection of all these, and the constniction of a radically
new view based on a new dynamic foundation. The later chapters of the book treat
of the early stages of the earth and of the way in which its leading processes took
their start from their cosmogonic antecedents, these being held to be essential
factors in the genesis of the planet.

Heredity and Eugenics. By John M. Coulter, William

E. Castle, Edward M. East, William L. Tovi^er, and Charles

B. Davenport. 312 pages. 8vo. Cloth. 19s 6d net.

Heredity and Eugenics is a book intended to present the recent developments of

knowledge in reference to evolution, heredity, and eugenics, and related subjects.

Five of the leading investigators in this field have collaborated in its production,

and the result is a work which will appeal to the general public, because it presents

the latest conclusions in a popular and extremely interesting manner.

Senescence and Rejuvenescence. By Charles Manning

Child, Associate Professor of Zoology in the University of Chicago,
xii + 482 pages. Svo. Cloth. 26s net.

The author of this volume, after some fifteen years of experimental investigation
of the nature and origin of the organic individual, has established certain facts which
afibrd a more adequate foundation for the general consideration and interpretation of
the age changes in the organic world than we have hitherto possessed.

Certain experimental methods have made it possible not only to follow the
physiological age changes in some of the lower animals, but to learn something of
their nature. The most important residt of this investigation is the demonstration
of the occuri'ence of rejuvenescence quite independently of sexual reproduction.
The book differs from most previous studies of senescence in that it attempts to
show that in the organic world in general rejuvenescence is just as fundamental and
important a process as senescence.

Individuality in Organisms. (The University of Chicago

Science Series.) By Charles Manning Child, x + 212 pages.

Small 12mo. Cloth. 10s net.

Professor Child's work is an attempt to state, and to present evidence in support

of, a conception of the nature of organic individuality which the author has developed

as a I'esult of fifteen years of investigation of the processes of reproduction and

development in the lower animals. In these forms organic individuality appears in

relatively simple terms, and it is here that we must look for the key to the problem

of individuality in the higher animals and man. The book appeals not only to the

physiologist and to the botanist, but also to the neurologist, to the psychologist, and

even to the sociologist.

The Biology of Twins. (The University of Chicago Science

Series.) By Horatio H. Newman, Associate Professor of Zoology

in the University of Chicago, x + 186 pages. Small 12mo. Cloth.

10s net.

In this volume many questions relative to twinning have been answered partly

through a study of data on human twins themselves and partly through a much

more detailed study of what has proven to be the most remarkable type of twinning

known — that in the armadillos of the genus Dasypus.

Sold in the British Empire (except Canada) by the
Cambridge University Press, Fetter Lane, London, E.C. 4

Volume XI SEPTEMBER, 1921 No. 2

LIBRARY

NEW Y04<lt

•OTANICAL

aAKUe^

NOTES ON THE GENETIC BEHAVIOUR OF CERTAIN
CHARACTERS IN THE POLECAT, FERRET, AND
IN POLECAT-FERRET HYBRIDS.

By FRANCES PITT.

(With One Text-figure and Plates XV and XVI.)
CONTENTS.

PAOB

I. Introduction 99

II. Description of the Polecat and the Ferret 101

(a) The Polecat 101

(b) The Ferret 104

III. Variation in the Polecat and in the Ferret 105

(a) The Polecat 105

(6) The Ferret 107

IV. Hybrids between the Polecat and the Ferret .... 109

(a) Material used in Experiments 109

(b) The Hybrids 109

(c) Notes on the Development of the Young Hybrids . . 110

(d) Second Litter 112

(e) Reciprocal Cross 112

(/) Fa Generation 112

(g) Fi X the Feri-et 112

{h) Fi X the Polecat 113

(i) Weakening of Pigmentation in Hybrid x Ferret Crosses . 113

V. Recapitulation and Conclusion 114

I. Introduction.

These notes deal with the genetic behaviour of colour and certain
cranial characters in the ferret, Martes furo (Linnaeus), in the polecat
Mustela putorius (Linnaeus), and in the hybrids obtained by crossing the
two species. Unfortunately, owing to the difficulty of keeping a number
of flesh-eating animals, which difficulty was much greater during the
War, and since then to the loss of all my ferrets from distemper, my
observations and results are very incomplete. They are only published

Journ. of Gen. xi ^

100 Polecat, Ferret, and Polecat-Ferret Hybrids

o o

A U

•B - -^ ■S B ^

^ 2 o ^

;z; H ^ 2i

^ 2

fe o - ■
CO *^

^ e3 !^ -^ ^ cS

ts

.5>J

y^pvduq '
2vfiqM-}soj

■^ o

H H EH

S ao ^ o

H H H

>!, E

a ^

&

m

^

-a

a>

a;

Cl)

<v

H

H

w o o 9 o I p I

c;, 4x CO 00 oc I 12' '

rH .-I i-H --I --I '-'

CO

O

00

(M

CO

IN

1—1

1—1

05

1—1

MCI-IROO}

9
o

o

o

o
i-i

O

o

(M

9
o

O

I-I

9

1—1

9

iH

00

aiqipuBH

O

'P

CO

o

CO

o

CO
CO

9

9

9
o

9

9

co

9

^-

co

CO

<N

00

SMOj-iiioo^ puiqaq
q?d9p [■BI'BlBd

o
6

o

o

9
6

9
o

00

9

CO

9

00

9

00
«b

as'BO-ui'Bjq
JO q'jdaa

OOO OOOco p'^09

'^

^^

a

!-i

a

o

CK

o

a

0)

a

X

HN

00

iH

CO

iH

9 p 9 9

rH 1-1 CO O O O 00

o o o o

9 0900

,00 so O -i 1^5

0000

^ 2

2 S o

M).2 &

^■»-^

CO o ■**

H k1

0990 9 9

CO oa -^ 03 "5 do

<M (M iM rH <M <N

0000

00 cnCOO OOOoi t-OOtr-t^ 00 IC-^CDlO * 00 t-
1-1 1-li-lrH (M<M<NrH .-(i-IiHtH i-l i-|i-li-li-| rH ,_| ^

a .2
a> a

'O o

a ja

s a

^ :=! a

5 a
a a

03 05 g -

9 9 ■« ^

«o 10 &I «

saniu'BO jaAO

>o oo>o 1099-^ 09995

C- •4)<C0C0 t— t— «Ot~ lO'P'^CO

p lo p ip 9

CO (N CO CO "*

s

qipBaaq
I'B'jiqioaaiui

q'jp'Bajq
;pTO^s'Eiv[

q^p'Bajq
oi')Brao2iz

X'ESBqoi^pno^

000 ooooo 9'?'"*<?'

000 >co9oo 0^900

(N005 >oa)f-i. Mi52'-'

COCOiM COCOCOto cocoeoco

1000 90'^Tf 09*^9

irtcocq oegoA< oSg^i'o

cococo ■<i<T)<-*i^ CO " "> en

000 o o o 00

S ^

9 9 9 p

10 "* U5 ■*

0000

O 05 O 05 "5 «0
CO (M CO <M CO IM

0000

0000

-* to -^

I ^T^ CH-OfCH- ^ '^ '^ fO CM-C>fOC>f ^T) CHOfOCH- 't) ^

. '•'•'

Oh

05 Oi

>o

10

CM

rH

f^

'~'

r^

a a

Ph

05
CO

d

'0
0

CO

S §

0
0

n -

CO

a

1— (

h-l

s

c3 e3

t>

Ul ^

S

>

•p,

0 00

K^

h-1 ' :

X

< <

(N OJ

CO

0 'C

."S

.

.ti

m m

iz; ^ ;z; ;z; ;z; iz;-

^ ^ iz; iz;

^S^zsmpq !z;mpp

^-s

, QJ " -

0

a

tH

S

s

•c

1^

m

m

3

0)

6=5

<u

^

a

H

£:

a
a
a

c8

a
0

"a

a

1

rt

1)

-2

g-

a

H

S

p=4 fe

F. Pitt 101

because I am not aware of any work done on these species, and the notes
may draw the attention of other workers to an interesting subject for
research.

II. Description of the Polecat and the Ferret.

(a) The Polecat, M. putorius, resembles the domestic ferret in build
and general appearance. The males generally exceed 400 mm. in length
(head and body), the females being smaller and averaging but 350 mm.
Its most obvious superficial characteristic is its deep blackish brown
colouring that has almost a plum tint, and which is much darker than
the hue of the deepest of so-called "polecat" ferrets. The colour deepens
into black on the underparts and extremities. The coat consists of two
kinds of fur, a thick drab or yellowish wool next the skin, and a longer
covering of coarse shining black hairs, which repel rain and damp. The
exact hue of the animal differs according to the season and the propor-
tions in which the two kinds of fur are present. In the winter, when the
woollen under-fur is thickest and longest, the polecat appears lighter
than after the summer moult. It often appears larger than its true size
owing to its habit, when frightened or annoyed, of erecting its fur and
fluffing out its thick tail. At the same time it makes a hissing noise,
and if this attempt at intimidation fails ejects from its anal glands a
most evil-smelling fluid. How disgusting an odour it is only those who
have smelt it can appreciate ! The polecat is highly nervous in disposi-
tion and I have never succeeded in taming one caught adult. In shape
the polecat is not so light or elegant as the ferret, being somewhat more
heavily built. Its head is decidedly blunter in both sexes, and, when
viewed from above, forms roughly an equilateral triangle. The muzzle,
tips of the ears, and a small patch of fur over each eye, are greyish white.
The latter patches sometimes join to form a pale band across the face, but
this is never so pronounced as in the dark form of the ferret (see PI. XV,
figs. 2 and 3). The skull is strongly built, but not more so than in the
ferret, from which, however, it differs in several particulars, namely in the
more flattened triangular bullae, the greater breadth of the post-orbital
region, which in this species averages 18 mm. in the male, and the larger
size of the carnassial teeth. Furthermore, in such polecat skulls as I have
been able to examine, the ridges extending from the post-orbital processes
to the sagittal crest form a much longer and more acute angle than they
do in ferret skulls (see Diagram I, figs. 1 and 2). But I must add that
Miller (3) figures a polecat skull with ridges forming as blunt an angle as
that of any ferret's skull. The polecat ranges through Europe eastward,

7—2

102 Polecat, Ferret, and Polecat -Ferret Hybrids

4 5

Diagram I.

1. Ferret skull, adult ^.

2. Polecat skull, adult ^J.

3. Skull of an F^^ individual, adult ^ No. 10.

4. Skull of No. 45, immature ^, the produce of an i^j x Polecat mating.

5. Skull of No. 19, adult $ , the produce of an Fi x Ferret mating.

F. Pitt

103

and from the Mediterranean north to Scandinavia, with practically no
variation, Miller describing but one doubtful subspecies from Spain, the
distinguishing character of which being its more golden under-fur. In
Britain the polecat is now rare, and is practically confined to central
Wales. Any specimens found outside that area should be viewed with
the gravest suspicion, generally proving to be merely escaped ferrets of
a dark colour. The species still exists in some numbers in Cardiganshire,
but how long it will be able to hold its own in the face of steady perse-
cution is another matter.

Measurements of Polecats and Ferrets.

Number of Specimen Sex H. and B.

Typical dark form of the Polecat :

XVIII coll. F. P.
LIII
♦Coll. Brit. Mus. N. H.

350 mm.
455 mm.
403 mm.
415 mm.
360 mm.

Tl.

127 mm.
165 mm.
1(53 mm.
190 mm.
140 mm.

Ear

20 mm.
25 mm.
27 mm.
30 mm.
23 mm.

H. F.

Remarks

56 mm. Cardiganshire
65 mm. ,,

61 mm. ,,

62 mm. ,,
53 mm. ,,

Erythristic variety of the Polecat :

LIV coll. F. P. ... ^ 450 mm. 150 mm. 21 mm. 58 mm.

(the above measurements are not absolutely reliable, being taken
from a specimen that had been set up)

Albino Ferret :

15

<?

358 mm.

126

mm.

18

mm.

50

mm.

—

32

3

375 mm.

137

mm.

20

mm.

54

mm.

—

33

?

312 mm.

112

mm.

16

mm.

54

mm.

—

Dark Ferret :

29

?

371 mm.

118

mm.

21

mm.

51

mm.

—

19

?

312 mm.

117

mm.

20

mm.

49

mm.

—

Erythristic Ferret :

20

$

340 mm.

130

mm.

21

mm.

53

mm.

—

28

?

350 mm.

117

mm.

20

mm.

49

mm.

—

Hybrid Ferret-Polecat :

10

(?

370 mm.

131

mm.

22

mm.

59

mm.

—

Hybrid x Polecat produce

45

c?

237 mm.

68

mm.

20

mm.

42

mm.

Died young

46

$

340 mm.

110

mm.

20

mm.

47

mm.

—

The Roman numerals refer to the numbers of specimens in my collection, and the
Arabic to the reference figures used in my experiments.

All the measurements given above were taken after death, it being found impossible to
measure a live animal satisfactorily, for which reason many of those bred were never
measured at all.

* From Miller's Catalogue of the Mammals of Western Europe,

104 Polecat, Ferret, and Polecat- Ferret Hybrids

(6) Description of the Feri'et. It has been long assumed that M. furo
is a domesticated form of the polecat, which it so much resembles in
general build and character, the fact that the majority of ferrets are pink-
eyed cream -coated albinos being considered confirmatory rather than
contradictory of the idea, especially as pigmented individuals resemble
polecats in their colouring and facial markings. That is to say they
resemble them in their general type of marking, but not exactly, as all
"fitchet^" ferrets I have seen have been much lighter in hue than the
palest of polecats, showing, too, broader facial markings, so that there is.
a marked reduction of pigment compared with the wild animal. Not
only are the white facial markings so much increased, but the woolly
underfur, which in the polecat is invariably a buffy-drab colour, is whitish
in the dark ferret. The "fitchet" ferret bears in outward appearance the
same relation to the polecat that a half-printed photograph bears to the
fully printed one from the same negative. The head of the ferret is
narrower than that of the polecat, appearing when seen from above as an
isosceles triangle, whereas the face of the polecat makes an equilateral
triangle ^ As regards size the average ferret is a trifle smaller than the
average polecat, it being a large male which exceeds 380 mm. head and
body measurement. The appended table shows the respective measure-
ments. In cranial characters we come to some important differences be-
tween the two forms, which differences have made Miller(3) express the
opinion that, " the ierret, Martes furo Linnaeus, though usually assumed
to be a domesticated variety of Mustela putorius, appears to be more
nearly related to the Asiatic M. eversmanni Lesson." The differing points
in the ferret's skull are : the more narrow constricted post-orbital region,
averaging in the male only 13 mm. in breadth, against the 18 mm. of the
male polecat, which gives the skull a "small-waisted" appearance ; the more
inflated and less triangular auditory bullae ; and the smaller carnassial
teeth. The only skull of M. eversmanni that I have been able to examine
showed a marked likeness to the ferret's in all the above characters, but
every ferret's skull I have seen has differed in one respect from both the
single specimen of M. eversmanni, and from every polecat's that I have met
with, this character being the shortness of the ridges that proceed from
the post-orbital processes towards the brain-case, and which converge at
a much blunter angle than they do in the polecat. The length in seven

1 " Fitchet" is a term commonly employed by keepers, rabbit-catchers, etc., to denote
the dark or so-called " polecat " ferrets. I have avoided herein the use of the latter word
for fear of misunderstanding.

2 Mr A. H. Cocks first drew attention to this in the Zoologist, Vol. xv. 1891, p. 344.

F. Pitt 105

ferret skulls averaged 14 ram., against 21 mm. in three polecats. In the
M. eversmanni skull that I examined the ridges were as long as those of
the European polecat and noticeably different from those in M. faro.
This seems to indicate a character in the ferret that is not possessed by
either of the other species. In such intangible peculiarities as tempera-
ment and disposition the ferret is very different from M. putorins, as is
shown by the ease with which it is tamed even after being neglected
while young. An adult-caught polecat is quite untameable, and even
half-bred ones require constant handling from their earliest youth to
make them docile. It takes a very serious fright to make a placid easy-
going ferret emit the vile defensive odour, but the hybrids never hesitate
to make use of it. In disease-resistance too the ferret differs from the
polecat, being less susceptible than the wild animal to some of the
diseases met with in captivity. For instance the majority of polecats
exhibited at the Zoological Gardens, Regent's Park, succumb more or less
speedily to pneumonia, and I have found the hybrids nearly as delicate.
In the case of epidemic diseases, such as one of the complaints that get
classed under the heading of "distemper," polecat-bred ferrets have in my
experience been always the first to take it, and with more fatal results
than those of pure ferret descent. For instance I recently lost six ferrets
in the following order: first No. 19, a dark female the offspring of a
hybrid and an albino ; secondly another dark female. No. 29, whose sire
carried polecat blood ; then a fawn ferret, No. 28, full sister to the pre-
ceding ; and lastly three white ones. The latter were not only the last
to take the disease, but did not succumb nearly so quickly.

From the foregoing it will be seen that there are quite a number of
points of difference between the polecat and ferret, both structural, pig-
mental, physiological, and temperamental, so that the grounds for doubt
as to whether the one is derived from the other are quite substantial,
and the materials are such as to afford a most interesting subject for
genetical investigation.

III. Variation in the Polecat and in the Ferret.

(a) Variation in the Polecat. It is noteworthy when considering the
relationship of the polecat and ferret, that, as far as I can ascertain, no
albino polecat has been recorded. Nevertheless it is probable such sports
do occur, for unless the specimen passed into the hands of an expert and
was examined as to structural details, it would almost certainly be
dubbed a feral ferret. The probability is enhanced when we find that

106 Polecat, Ferret, and Polecat-Ferret Hybrids

erythristic mutations are known. In this variety the black pigment is
entirely suppressed, leaving the animal a beautiful reddish colour.
Erythrism is usually regarded as partial albinism, and in many species is
accompanied by absence of pigment from the eyes, but in the " red "
polecat the eyes are dark. The following is a good description of this
mutation. " The colour may be described roughly as : under fur light
buff, the longer hairy coat being a reddish-brown. When seen in sun-
light the latter shows a kind of faint purple 'haze,' it is difficult to
describe it otherwise." The writer, Mr F. S. Wright (4), adds that in all
the specimens examined by him the facial markings were normal, " but
owing to the lighter ground colour of the animal the contrast was
naturally not so great as in the common dark polecat"^. I can confirm
this from my own observation.

Light brown polecats are only known from Cardiganshire, and chiefly
from the district around Aberystwyth, the greater number being recorded
from the neighbourhoods of Borth and Tregaron bogs. The first specimens
were made known in 1902-03(2), since when more than a dozen have
been killed. From this it seems that the form has originated compara-
tively recently in the locality mentioned, and that the mutation (for we
may safely assume it to be due to the loss of the black pigment factor)
far from tending to disappear is even on the increase. As all the red
polecats that have so far been recorded have been met with on the
Costal Plateau of Cardiganshire, an area which extends from the River
Dovey to Tregaron— about eighteen miles apart — it is evident that the
area, wherein we must seek the point of origination, is a comparatively
restricted one. As the first specimens were killed on Tregaron Bog it is
possible that this is the locality. As would be expected in the case of a
variation originating from the loss of a factor the reddish-brown colouring
gives us an example of simple Mendelian inheritance. The case is some-
what interesting however, as it is not very often that we get evidence of
the action of Mendel's law in wild Nature. Even in this case the infor-
mation is but scanty. In July, 1915, two young polecats of the same
litter, the one normal in colour, and the other red, were killed on Borth
Bog, " It is noteworthy that in the two offspring the colours show no
intermingling whatever " (4). Unfortunately there is no evidence to
show whether the light brown type is dominant or recessive, but judging
by a similar fawn-coloured variety of the ferret, presently to be described,

1 Mr H. E. Forrest (2) has stated that the facial markings are sometimes suppressed in
this variety, but he now tells me {in litt. Feb. 16, 1920) that he is confident it was an error
due to the lessened conspicuousness of these markings on a light brown ground.

F. Pitt 107

and by analogy with the behaviour of similar pigmentary characters in
other species it is probably the latter. If we are right in supposing that the
red coloration is due to the absence of the factor for black, it becomes
at once apparent that it would be a most difficult task to decide how and
when the factor first fell out. The tendency to the red coloration might
lie latent for many generations until a suitable mating brought together
gametes from both of which the factor was missing. It must be remem-
bered that the dark offspring of a red x dark mating would necessarily
be heterozygotes, of the constitution DBdB (I am herein using the
letter " D " to indicate the dark black-brown colour, " B " the light red-
brown, "db" the absence of all pigment or albinism), and if such heterozy-
gotes are fairly widely distributed it would account for the sporadic way
in which the red-brown individuals have turned up during the last
seventeen or eighteen years since the first specimens were obtained. It
is evident that in this variety of the polecat we have not really anything
new, for the factors for it have undoubtedly been latent in the animal's
constitution since it assumed its present type of coat and colouring.

(b) Variation in the Ferret. We may confidently assume that in the
pigmented and so-called " polecat " ferret (known among keepers and
rabbit-catchers as the " fitchet ") we have a near approach to the
ancestral type, from which the well-known white form is almost certainly
an albinistic mutation. Erythristic ferrets analogous to the reddish-
brown polecats are also met with, but, as the " fitchet " ferret is paler
than the dark brown-black wild polecat, so is the fawn ferret paler than
the " red " polecat, being much more " washed-out " looking. It is
evident that we here meet with similar if not identical pigmentary
characters to those found in the polecat, therefore the same letters may
be used to denote the factors, D for dark black-brown or full pigmenta-
tion, B for light red-brown, and so on.

A detailed account of the behaviour of the black-brown pigmentary
character will be given later on under the heading Ferret-Polecat
Hybrids, and I will now confine myself to the erythristic variety. The
exact tint of the fur is best described by the word " sandy." The colour
is confined to the long outer hairs, the woolly undercoat being white.
As in the red-brown polecat the absence of black pigment is correlated
with size and vigour. Most of the sandy ferrets I have seen have been
exceptionally fine animals. They are also noticeably quicker tempered
than dark or white ones. It is my experience that they require a good
deal more handling to keep them quiet and docile, and even then cannot
be trusted. One sandy ferret I had was called " Ginger " on account of

108 Polecat, Ferret, and Polecat-Ferret Hybrids

her liability to snap. Other persons have expressed the same opinion to
me. In connection with this it must be remembered that red-headed
people are notoriously sharp tempered, and chestnut horses are often very
" hot." Erythristic ferrets are not very common, but I have met with the
variation in four distinct strains. The individuals that belonged to me
originated from a mating between "a white female and a very large dark
male... they had some dark, some white, and some sandy young ones^"
As there were several sandy young ones we may be certain that one
parent, presumably the dark, was heterozygous for the red character,
and that for the origin of the mutation we must go further generations
back, because if the factor had fallen out at this particular mating only

Pedigree of Light Brown Ferrets.

1915

9 X «^

1916 "Some dark" "Some sandy" "Some white"

9 X Gf

1917

T~i '

• o 9x«r 9 >< ef

20 J 21

r

1918 f

1

1

1 1

9

29

30

r

28

31 32

33

1

1

© 0

1^
©

© ©

T"
©

I
o

1919

9

34

9

35

9

36

0

o

Dark colouring.
Light brown.
Albino.

The numbers refer to animals bred by me, and serve to identify the same individuals

in different pedigrees.

one of the family could have borne the new character. When the
erythristic ferrets obtained in this litter were bred from, the owner tells
me, a litter of four " reds " and one dark resulted, which, provided no
error occurred, shows that the factor for full pigmentation may be borne
in a latent state by the red individuals. However the dark colouring is

^ For this and other information I am indebted to Mr Wm. Milner, of Much Wenlock,
who most kindly gave me some of his erythristic ferrets.

o

F. Pitt 109

usually dominant to its absence, as will presently be shown when the
evidence concerning the hybrids between the ferret and polecat is given.
It is probable that in fawn and in " fitchet " ferrets we meet with dilu-
tion factors that complicate matters, but the numbers so far bred are too
small for the facts to be elucidated. The various matings of which I have
records in which fawn-coloured ferrets have been involved are tabulated
below. It will be seen that in a cross of red x dark the white class was
larger and the red smaller than expected.

Tabulated results obtained from matings of erythristic ferrets :

Fawn X fawn

Dark
1
0

Fawn
: 4

: 8

Albino
: 0
: 2

Totals

1

: 12

: 2

Fawn X dark
Fawn X white . . .

2
0

1
1

: 3
: 2

IV. Ferret-Polecat Hybrids.

(a) As the characters in which the polecat and the ferret differ have
already been described, it is unnecessary to go over them again ; I will
only remind the reader that they consist of certain cranial peculiarities
and of coat colour. There is also considerable difference in temperament,
the domestic animal being placid and easy-going, the wild one very much
the reverse.

(b) In 1912 I was fortunate in obtaining a very fine male polecat
from Cardiganshire, also a female, but as she never bred she is only
mentioned here because she served as a standard with which to compare
the female hybrids that were subsequently reared. The male, No. 5 in the
accompanying pedigree (p. Ill), was mated with an albino ferret of known
pure white ancestry for several generations. The resulting litter num-
bered five, two males and three females, all dark in colour, and showing
complete dominance of the wild type in the F^ generation. Perhaps the
words "complete dominance" need qualification, for, notwithstanding the
hybrids were much darker than any so-called polecat ferret, they did
show slight traces of the ferret side of their ancestry, inasmuch they had
whitish under-fur instead of the buffy-grey wool of the true polecat.
Nevertheless they were very dark in colour, the black-brown of the long
hairs having the slight purplish sheen so characteristic of the polecat.
The polecat temperament too was fully apparent, for, despite the fact
they were made great pets of, and were constantly handled, they were

110 Polecat, Ferret, and Polecat-Ferret Hybrids

far more nervous than ordinary ferrets. They were easily frightened, and
when startled would emit the horrible odour from the anal glands that
constitutes one of the defences of the polecat. I have only known a
ferret do this under stress of great fear and excitement — e.g. when
pounced upon by a terrier in mistake for a rat ! They also hissed at the
least thing. They were far quicker and had more vitality than ordinary
ferrets, romping and dancing together in the most delightful manner.
When used for rabbiting they proved almost too quick, for it was quite
difficult to pick up the nimble females as they darted in and out of the
holes. They were deadly workers, killing their rabbits in a few seconds
by biting them over the eyes, but never behind the ears after the manner
of a stoat. A male (No. 10) was a good ratter, and I have seen him chase
a rat that had bolted in the open. He was a great pet but had some
undesirable traits in his character, his genius for escaping from all
descriptions of cages being extraordinary. He would bite and tear his
way through wire netting in a very short time, an ordinary wooden cage
never kept him at home for long, and the only place where he was really
safe was a loose-box with a strong well-fitting door. In his old age he
climbed over a four foot pig-sty wall, made his way into a fowl-house, and
therein slew six cockerells. When he died at four years old his teeth
were the worn stumps of a very old animal. It then also transpired
that though so polecat-like in appearance and behaviour his skull was
that of a typical ferret. When placed side by side with the skull of a
ferret of pure ferret ancestry it resembled it in all respects, being dis-
tinguishable at a glance from that of a polecat. Here we have an
interesting case of the colour and temperament of one parent, and the
cranial characters of the other, being dominant in the Fx gen^ation.

(c) Notes on the Development of the Young Hybrids. Before going on
to another litter of hybrids it may be worth while to give details of the
growth and development of the first family. They were born on June 26th,
1912, and on July 10th two crawled out of the nest, being then covered
with milk-white hair^ so short that their greyish-blue skin showed
through it. Their mother dragged them back into the nest at once. At
17 days old they had grown a great deal, were darker, and had traces of
the typical polecat facial markings. At 24 days old they were much
browner, but with yet a mane of "skim-milk" coloured hair down the
neck. The largest males' eyes had begun to open, and they were all

^ Young ferrets, whether dark or albinos, and likewise young polecats, have a white baby
coat. They are born naked, but quickly grow a scanty covering of milk-white hair, which
as described above then gives place to the dark fur.

F. Pitt

111

112 Polecat, Ferret, and Polecat -Ferret Hyhrids

trying to eat rabbit meat. The 25th day found two able to see a little,
and all making vigorous efforts to eat. By the 36th day they were quite
dark in colour, with white tips to their ears, white muzzles, but the
light patches on the sides of the face as yet shadowy and indistinct. At
seven weeks old they were perfect little polecats as far as outward
appearance went, and it was not until they had gained their winter
pelage that some slight trace of the ferret side of their pedigree could
be discerned.

{d) A Second Litter of Hybrids. In 1916 the male polecat No. 5
was mated with another white female ferret (No. 34). She had by
him a fariiily of eight, all dark, but being a bad mother only reared
three. These were in all respects exactly like the first hybrids, being
quite as polecatlike in appearance, and even more nervous, for they did
not get so much handling and petting.

(e) Reciprocal Cross. The reciprocal cross could not be made, as the
female polecat which I had obtained after so much trouble never bred ;
but as albinism is invariably recessive to full pigmentation, we may be
fairly confident that the result would have been the same.

(/) F2 Generation. Owing to one mishap and another no F2, litter
was reared to maturity. The F^ females that I kept to breed from caught
distemper and died, but a male and female that I had given to Mr Riley
Fortune did well until just before their young were born. The female
then escaped from her cage, found a hole under the greenhouse floor,
and there made a nest, in which her family was born. Mr Fortune kindly
sent me full particulars of the progress of the litter, but misfortune was
again in store, and the mother getting away was lost before her young
ones could see. When deserted they were a whitish colour. As the
young of all ferrets, whatever their ultimate colour is to be, have a white
baby coat, this is no indication as to what they would have been like
when adult. But one thing was proved, namely that the hybrids are
perfectly fertile inter se.

(g) Fi Generation x the Ferret. A mating between a F^ male and a
white ferret (Nos. 9 and 10) gave four albinos to two dark young ones.
These latter were not so dark as their sire, and approached the ordinary
fitchet ferret type. Another mating of an F^ male with an albino female
(Nos. 38 and 39) gave four whites to three darks. The total for the two
litters was eight to five, whereas expectation was equality, but the
numbers are too small for one to be able to arrive at any conclusions.
When an Fj female, No. 44, was mated with a male bred firom the

F. Pitt 113

^1 X albino ferret cross (No. 21), which was polecat coloured, and there-
fore presumably heterozygous in constitution, three dark and one white
young resulted (Nos. 47 to 50)\ When two second-cross ferrets,
Nos. 19 and 21 (i.e. the produce of the F^ generation x albino ferrets),
were bred together they had five dark young ones to one white. A mating
of interest was that which has already been referred to between Nos. 20
and 21. The former was an erythristic female, the latter a dark male,
the produce of a hybrid x a white ferret. The alliance between Nos. 20
and 21 gave rise to two dark, one red, and three white ferrets. As re-
gards cranial characters, all skulls that I have been able to examine of
these hybrid x ferret crosses have been indistinguishable in every par-
ticular from those of pure-bred ferrets. Full details and measurements
of these skulls will be found in the table on p. 100.

{h) Fi Generation x the Polecat. In 1916 I was fortunate in obtaining
another fine male polecat from Cardiganshire. He was mated with
No. 37, an ^i female by my first polecat. Her litter only numbered two.
These two youngsters appeared true polecats in every respect, having
the broad faces, dark coats, drab under-fur, and highly nervous disposi-
tions of the wild animal. I subsequently found that in cranial characters
too they were polecats, for their skulls were typical of 31. putorius. The
male died young, but the female lived to be eighteen months old, when
she too unfortunately succumbed to pneumonia without having bred.
This susceptibility to pneumonia is a characteristic of the wild animal,
most captive polecats sooner or later dying of it. Pure-bred ferrets
appear to be much more resistant to the disease. In temperament and
disposition this young female was quite the wild animal, she was
intensely nervous, bit whenever she got the opportunity, and in her
excitement was always ready to emit the horrible polecat smell ; in short
there was no trace of the ferret about her.

(i) Weakening of Pigmentation in Hybrid x Ferret Crosses. Before
glancing back over the evidence that has been presented concerning
polecat X ferret hybrids, it may be as well to draw attention to the
weakening of pigmentation that occurs when the polecat-ferret hybrids
are bred back to the ferret, the three-fourths ferret offspring being much
paler than the first cross. These individuals are very like the ordinary
" fitchet " ferret, but still further dilution takes place when another back
cross is made. After this the pigmented individuals are very " washy "

1 For information concerning this litter I am indebted to Mr Owen of Oswestry into
whose possession the parents had passed.

114 Polecat, Ferret, and Polecat-Ferret Hybrids

in appearance and exactly like the average dark ferret. Whether there
is a true dilution factor involved appears doubtful, the lighter appear-
ance of the animal being due to a smaller quantity of dark hairs in its
coat, not to reduction of the pigment in the hair, as in "dilute" mice,
rats, and rabbits. Were true dilute ferrets to occur I should expect them
to be a pearl-grey colour. In "fitchet" ferrets the lighter colouring, lighter
that is by comparison with polecats and their hybrids, is due to the
white or pale cream under- fur, to the cream basal portion of the longer
hairs, and to the general increase of cream coloured fur.

V. Recapitulation and Conclusion.

Attention has been drawn in this paper to the interesting material
for genetic study that is to be found in the polecat and ferret, the two
forms have been described, together with their points of difference and
resemblance, and it has been shown that the characters in which they
differ are such diversified ones as coat colour, cranial peculiarities, and
temperament. The doubt as to whether the ferret is a descendant of the
European or Asiatic polecats has been alluded to, and evidence has been
given that it at any rate will cross freely with M. putorius, the hybrids
being fertile inter se, and with either parent. The F^ generation shows
complete, or very nearly complete, dominance of the polecat type as
regards outward appearance, but what evidence is to hand indicates that
in cranial characters the ferret is dominant. When the hybrids were
bred back to the ferret the polecat coloration and temperament were
soon lost. Likewise when the hybrids were bred back to the polecat,
animals that were apparently pure polecats resulted. An interesting
result of the back crosses with ferrets (albinos) was the gradual weakening
of the colour in the pigmented offspring, due not to dilution of the pig-
ment in the hairs, but to a reduction in the amount of dark fur. Atten-
tion has also been drawn in this article to the erythristic varieties of
the polecat and ferret, and particulars have been given of the occurrence
of " red " polecats in the Aberystwyth district, together with some
evidence that this mutation has appeared comparatively recently, and is
inherited according to Mendel's law. In the case of the ferret it is shown
that erythrism is certainly dependant on a Mendelian factor, being
dominant to albinism and recessive to the black-brown coloration.
Both in the ferret and polecat erythrism seems to be correlated with
increased size, and certainly in the ferret is usually accompanied by a
quick temper and general increase of vitality. To the writer the most

JOURNAL OF GENETICS, VOL. XL NO. 2

PLATE XV

Fig. 1. Fi Ferret-Polecats at eight weeks old.

Fig. 2. Polecat {^) from Cardiganshire.

Fig. 3. Dark coloured (J ferret.
Ill the above phofog rajahs are from life.

JOURNAL OF GENETICS, VOL XL NO. 2

PLATE XVI

1. Head of Polecat ^ .

2. Head of Ferret ^.

3. Head of Fj ^.

F. Pitt 115

important points that have been brought forward appear to be the indi-
cation of the Mendelian inheritance of a structural character (type of
skull), and the evidence concerning a variation due to the loss of a factor
appearing and maintaining itself in a state of Nature — i.e. the ery thristic
polecats of Cardiganshire.

WORKS CONSULTED.

1. Bateson, W. 1909. MendeVs Principles of Heredity.

2. Forrest, H. E. 1907. Vertehrata Fauna of North Wales.

3. Miller, G. S. 1912. Catalogue of the Mammals of Western Europe.

4. Wright, F. S. 1916. " The Mammalian Fauna of North Cardiganshire," a paper

in the Zoologist^ September 1916.

5. PuNNETT, R. C. 1911. Mendelisni.

Journ. of Gen. xi

SOME EXPEEIMENTS ON THE ORIGIN OF NEW

FORMS IN THE GENUS HIERACIUM

SUB-GENUS ARCHIE RACIUM.

By C. H. OSTENFELD.
Landbohjjskolens Botaniske Institut, Copenhagen.

(With Plates XVII and XVIII.)

In earlier papers (1906, 1910, 1912) I have shown that within the
subgenus Pilosella of the genus Hieracium new forms arise by means
of hybridization associated with apogamy. That is to say : hybridization
produces the new forms (hybrids) and apogamy keeps them unaltered
from generation to generation.

In the other big subgenus Archieracium of the same genus the
matter seems different. Most of the forms are purely apogamic and only
very few are sexual. Amongst the latter is H. unibellatum, and it is
interesting to know from Mendel's letters to Nageli, published by C
Correns (1905), that Mendel had succeeded, by using this species as
father, in producing two hybrids, the only ones produced experimentally
within Archieracium, at least as far as I am aware. As regards another
sexual Archieracium, H. virga aurea, I have observed that the offspring
from an individual which stood unprotected in my cultures showed
distinctly the influence of hybridization as it was very heterogeneous in
its appearance.

With the exception of these few cases genuine hybrids amongst
Archieracium are not known, as we cannot rely upon the records of hybrids
given in floristic and systematic literature. Mendel complains that it is
" sehr schwiei'ig die Selbstbefruchtimg aufzuhalten " (according to our
present knowledge we have to replace the word "Selbstbefruchtimg" by
"apogamy"). All my own experiments have hitherto been fruitless.

Meanwhile, as we find abundant polymorphism within this subgenus,
we must search for other explanations for it. The many forms — micro-
species we may call them — are apogamic and constant, the offspring
being like the parent.

It seems quite natural to suppose that these microspecies are not
very old, nor all of the same age. Species of a wide geographical range

8—2

118 Origin of New Fomis in Hieracium

are probably old, species of a restricted range may be old or may have
newly arisen. In the former case they are to be considered as dying-out
species and have usually a rather isolated position from a systematical
point of vieAv ; the Archieracium microspecies can scarcely be included
in this category. They must most naturally be considered as newly
arisen species of restricted range.

A Swedish botanist, G. Samuelsson, has in a paper (1910) given the
geographical range of some Hieracia from the middle of the Scandinavian
peninsula. He shows that some of them have a very restricted area of
occurrence, mostly with a centre where they are most common and from
which they radiate becoming less and less common towards the periphery.
These microspecies are usually closely related to others which have a
somewhat wider distribution, and Samuelsson considers the former as
derived from the latter.

I think his supposition is right, and starting from this idea I tried
to get new forms in my cultures by making them as extensive as
possible. I thought that there would be a chance that between a large
number of uniform individuals single diverging individuals might appear.
This is the same method as that used by H. de Vries in his first Oenothera
Lamarckiana experiments.

Therefore I sowed as many seeds as possible of several rather widely
distributed forms which were tried before by culture after agamization^

Each form was constant and produced uniform offspring and apoga-
mically. Each of these new cultures came from seeds of one individual
and only from agamized heads. In the four first experiments I got from
100 to 300 plants ; the seeds were sown in 1910 and the plants flowered
in 1911. In three of the experiments all the plants were uniform and
like their parent, but in the fourth experiment, a H. rigidum Fr. {H.
tridentatum) which originally came from a wood near Svendborg, Denmark,
I got 154 plants, 153 of which were normal, while one was different in
several respects. Some flowerheads of this individual were agamized in
1911, the seeds sown in 1912 and the new generation flowered in 1913.
All the plants of this generation were uniform and like their parent,
thus different from the original H. rigidum. I have later repeated the
agamization and sown the seed twice (1912-13 and 1915-16) and ob-
tained the same result.

The new form — I call it H. rigidum Beta — was thus at once constant

' By the terms " agamization " and " agamize " I mean the process of I'emoving the
anthers and the stigmas by means of a cut with a razor before the opening of the flowers
of a flower head (the prefix a and the word yd/xos, marriage).

C. H. OSTENFELD 119

and produced a uniform offspring. It differs from the original form,
H. rigidum Alpha, in the following characters (compare the plates):

1. It is lower and smaller in all its parts and the branches are more
divaricate.

2. The involucral leaves (bracts) are more appressed than in Alpha,
where their tips are more or less recurved ; further they are of a lighter
colour (fewer dark hairs).

3. Before the flowerhead opens the surfice is nearly flat, while in
Alpha it is distinctly concave, the corollas of the outer unopened flowers
reaching higher up than those of the inner ones.

4. In the fully open flowerhead of the AlpJia-form the corollas of the
outermost flowers are much longer than those of the innoA- ones, while
in Beta the difference is not so marked and the diameter of the head
consequently somewhat less.

I have shortly mentioned (1912, 1919) this result in some papers,
but as I had to deal with only one single case I felt it necessary to tiy
to get more material showing the same result. The appearance of the
deviating individual might be referable to some experimental error. The
probability for such an error was, as far as I could judge, only very slight,
as I did not know any form which was like the new one and as all the
seeds sown came from agamized heads, but on the other hand one indi-
vidual was rather unsatisfactory to build upon. Therefore I repeated
the whole experiment in an enlarged scale.

In 1915 I agamized a good many heads of a plant of the original
H. rigidum and of another plant derived from the former after agamiza-
tion, thus both Alpha-individaah. The seeds of each of these two plants
were sown in 1916 and flowered in 1917. The offspring of H. rigidum
itself gave 635 plants like the parent and three " doubtful " ; the offspring
of the Fi of H. rigidum gave 469 typical plants and one "doubtful." The
" doubtful " plants were such as appeared to differ somewhat from the
typical, but as the plants stood rather densely, the differences might be
caused by conditions. Heads of the four " doubtful " plants and of one
typical were agamized, and the seeds were sown in 1918. The new plants
flowered in 1920, and it now appeared that the plants from three of the
" doubtful " parents did not differ from the typical. But the offspring of
the fourth " doubtful " plant (one from the seeds of the original H.
rigidum) gave 109 plants which were distinctly different from the type,
and also different from the first deviating form {Beta). We have thus
again obtained a deviation which was at once constant, gi^'ing uniform
offspring.

120 Origin of New Forms in Hieracium

This new form — H. rigidum Gamma — differs from the original Alpha
in the following characters (compare the plates) :

1. It is somewhat lower and with more divaricating branches.

2. The leaves are distinctly broader than in Alpha.

On the other hand there is no marked difference with regard to the
involucral bracts nor to the shape and size of the flowerheads, only a
slight tendency towards the Beta-iovm. The same is the case with the
colour of the involucral bracts.

As to the vigour and height the new (9am7na-form is intermediate
between Alpha and Beta, but it has broader leaves than either of them.

There is a peculiarity which may be of some importance. The original
form {Alpha) is very fertile after agamization; nearly all achenes are
developed and contain sound-looking germs. In Beta I have always
found many empty achenes and the germinating power is less, while in
Gamma the fertility is nearer to Alpha, although not reaching it.

My experiments show that it is possible to produce nevj forms apo-
gamically from, constant apogamic forms of Archieracium, these new
forms being themselves apogamic and at once constant. They may be
called " apogamic mutants." Most probably the numerous microspecies of
Archieracium found in nature have arisen in that maimer.

There seems, to judge from the experiments, to be a contrast between
the origin of species in the two subgenera ; in Pilosella the forms arise
by means of hybridization, in Archieracium by means of " apogamic
mutation." But is this contrast radical?

In Pilosella we find species which are partly apogamic and partly
sexual as they are able to produce hybrids after fertilization with other
species, and at least some of these species are themselves probably — to
judge from the polymorphy of the jPj-generation — to be considered as
hybrids. The hybrids produced are apogamic like their parents and
owing to apogamy at once constant. If we now cultivate in large
numbers such a new form produced by means of hybridization, would it
then be possible to get single deviating individuals, as I have shown
with regard to Archieracium ? I have not yet made any thorough ex-
periment, but I should think that we may answer the question in the
affirmative. A single experiment which was not completed points in
that direction : in a large culture of a race of H. aurantiacum — all from
seeds obtained after agamization of the heads of a single plant— a single
much deviating individual was found, but unfortunately it died before
flowering. Supposing that this fact is not due to any experimental
error, it speaks in favour of my hypothesis that also within the newly

C. H. OSTENFELD 121

produced species-hybrids of the Pilosella subgenus we may find " apo-
gamic mutants " if we make the cultures extensive enough. And if this
supposition is right, it follows that we can declare the " apogamic
mutants" as being after-effects of earlier hybridization.

In that way the contrast between Pilosella and Archieracium dis-
appears. Archieracium is then further advanced as regards apogamy,
and the microspecies of the present time are either the apogamically
seed-setting hybrids from the hybridization-period (as no fertilization
tackes place we must regard the apogamic propagation as a vegetative
propagation, a " klon "-propagation) or " apogamic mutants " arisen from
these. In both cases we find a basis of hybridization and in both
subgenera the real cause of the appearance of the many forms is the
hybridization, luhile apogamy is the reason that they keep constant.

Another question is how this sudden appearance of single " apogamic
mutants " as after-effects of hybridization goes on. This is really a cyto-
logical problem which I can only touch very briefly. From extensive
studies of the cytology of the divisions in the pollen mothercells in
many Archieracium species Rosenberg (1917) has shown that there are
many irregularities. E.g. it happens that single chromosomes do not
follow the others in the regular movements during the division; we
may, so to speak, say that they are rejected or " forgotten." If we suppose
that something similar happens during the division of the embryosac
mothercell, it would be but natural to think that the " somatic eggcell"
in some cases gets a different number of chromosomes, e.g. one chromo-
some less than ordinarily. The embryo formed Avithout fertilization
from such a deviating " somatic eggcell " would produce a plant different
from the others derived from the same flowerhead, and in that way
we get an explanation of the appearance of the so-called " apogamic
mutants." This is of course only a mere hypothesis which has to be
proved by cytological investigation, but it is in agreement with the
experimental facts.

LITERATURE.

CoRRENS, C. (1905.) "Gregor Mendels Briefe au Carl Nageli, 1866—1873. Ein

Nachtrag zu den veroflfentlichten Bastardierungsversuchen Mendels." Abh. d.

math.-phifs. Kl. d. k. scichsischen Ges. d. Wiss. xxix. No. 3, 187—265.
Ernst, A. (1918.) Bastardierung als Ursache der Apogamie im Pflanzenreich. Jena

(Gustav Fischer).
LoTSY, I. P. (1916.) Efolutlon by means of Hybridization. The Hague (M. Nyhoflf).
Massart, Jean. (1920.) "La notion de I'espfece en biulogie." Bull. Acad. roy. de

Belg. des sciences, pp. 366 — 381.

122 Origin of New Forms in Hieracium

OsTENFELD, C. H. (1912.) "Experiments on the Origin of Species in the Genus
Hieracium (Apogamy and Hybridism)." The New Phytologist, xi. 347 — 354.
Contains the earlier literature.

(1919.) " Kimdannelse uden Befrugtning og Bastarddannelse hos nogle

Kurvblomstrede og disse Forholds Betydning for Formernes Konstans." Kgl.
Veterincer- og Landboh'^jskolens Aarsskrift, 1919, 207 — 216, Engl. Resume, 217 —
218.

Rosenberg, O. (1917.) " Die Reduktionsteihmg und ihre Degeneration in Hiera-
cium." Svensk Bot. Tidskrift, Bd. xi. (1917), 145—206.

Samdelsson, G. (1910.) " tjber die Verbreitung einiger endemischer Pflanzen."
Arkivf. Botanik, Stockholm, Bd. ix. No. 12, 16 pp., 2 maps.

Tackholm, G. (1920.) " On the cytology of the Genus Rosa." Svensk Bot. Tidskrift,
Bd. XIV. (1920), 300—311.

Winkler, H. (1920.) Verbreitung mid Ursache der Parthenogenesis im PJlanzen-
iind Tierreiche. Jena (Gustav Fischer).

JOURNAL OF GENETICS, VOL. XL NO. 2

PLATE XVII

JOURNAL OF GENETICS, VOL XI. NO. 2

PLATE XVIII

THE INHERITANCE OF WING-COLOUR IN
LEPIDOPTERA.

V. MELANISM IN ABRAXAS GROSSULARIATA
(VAR. VARLEYATA, PORRITT).

By H. ONSLOW.

(With six Text-figures and Plate XIX.)

Several examples of melanism have already been investigated, and
their inheritance described in previous papers. The chief interest of
these melanic strains lay in the fact that they were all discovered in
the south of England, and far from any manufacturing centres. As
appears to be usual, the melanic variety in each case behaved as a
simple mendelian dominant.

That exceedingly variable species, Abraxas grossulai'iata, possesses
a black variety, known as var. varleyata, Porritt', found in Hudders-
field, and now established as a domesticated strain. Basing his opinion
on a single family, bred by Mr L. W. Newman (see the third table),
Mr Bateson'^ concluded that the black variety was not dominant, but
recessive to the type form, like the melanic variety oi Odontopera hidentata
(The Scalloped Hazel) investigated by Bowater^ More recently, Porritt^
has published the results of some breeding experiments. In 1906 he
obtained "a considerable brood," from a mating oi varleyata x varleyata,
all of which proved to be like their parents. Further, in 1907 he
obtained a pairing between varleyata % x grossulariata (^ . The latter
was a typical wild insect, and 9 cTcT and 5 $ $ resulted, all of which
were like the male parent. With a view to testing this conclusively, a
strain of varleyata was procured by the author from Mr L. W. Newman,
and both the Rev. G. H. Raynor and Mr Porritt with great kindness
furnished some ova.

1 Porritt, G. T., Ent. Mo. Mag. Vol. xli. p. 211, 1905.

- Bateson, W., MendeVs Principles of Heredity, 1913, p. 44.

3 Bowater, W. J., Journal of Genetics, Vol. in. p. 299, 1914.

* Porritt, G. T., Ent. Mo. Mag. Vol. xliii. p. 12 and p. 276, 1907.

124 TJie hilieritmice of Wing -Colour in Lepidoptera

As it was intended later to carry out experiments with var. varleyata
and the well-known pale variety, lacticolor, some individuals of the
latter variety were used in certain pairings, with the result that both
lacticolor and a completely new variety, with an exceedingly handsome
facies, appeared among the F.^ offspring. The genetics of this variety,
named var. exquisita, Raynor^ will be dealt with in a future communi-
cation. Owing to the occurrence of these varieties, the material in the
following tables has been classified into melanic and non-melanic, in
order to reduce the number of columns.

The result of mating varleyata with non-melanic, that is to say,
ordinary wild insects or lacticolor, was as follows :

Melanic x Non-7nelanic.

RB X DD.

Female x Male

Imagii

IPS i\

Melanic
Male Female Totals

Non-melanic

Family

Male

Female

Totals

'16 Z

type X varleyata

— — —

6

9

15

'18^

varleyata x lacticolor

— — —

1

—

1

'18 B

lacticolor x varleyata

— — —

19

17

36

'ISiiT

varleyata x type

— — —

1

—

1

'18 L

,, X lacticolor

:

—

3

3

'18 2^

,, X ,,

_

1

2

3

'18 0

type X varleyata"^

10

10

20

'18 fJ

,, X ,,

8

17

25

'19^

varleyata x lacticolor

._

3

2

5

'19 W

5> ^ >)

5

9

14

'19 Z

type X varleyata

—

23

20

43

'19 X

varleyata x lacticolor

.

2

—

2

'19 m

X

5

5

10

Total

178

It is seen that the black variety is recessive both to type and to
lacticolor. The amount of the black pattern in F^ individuals appears
to vary very considerably, but this point will be dealt with later.
Whether lacticolor is heterozygous for varleyata or not, the variation in
the amount of black is usually scarcely noticeable.

1 Eaynor, G. H., Ent. Rec. Vol. xxi. No. 11, p. 205, November 1919. It is possible that
the variety named albovarleyata by Porritt [Ent. Rec. Vol. xxix. p. 136, 1917) is exquisita,
but in answer to some inquiries Mr Porritt says that this form was bred from stock which
never contained a single specimen of lacticolor ; in fact, he had never bred the latter
variety. Moreover, although the resemblance is close, there appeared to be some slight but
significant difference in the appearance of the black shoulder -knot on the under surface.

* This varleyata ^ was radiated (see p. 136).

H. Onslow

The result of mating varleyata x varleyata was as follows
Melanic x Melanic.

125

RRx

RR.

Imagines
.J-

Bred by

Melanic

Totals

Noil -melanic

Family

Male

Female

Male

Female ToUlg

1906

Mr G. T. Porr

iit

?

9

28

_

'18£

Ova from Mr G

r. T.

Porritt

2

1

3

'181*

Ova from Rev. G.

H. Raynor

8

5

13

1 1

'19 F

H. 0.

1

1

'191-

H. 0.

5

3

8

'19 0

H. 0.

1

1

'19 P

H. 0.

1

1

'20 E

H. 0.

9

3

12

'20 G

H. 0.

6

3

9

'20 0

H. 0.

7

3

10

'20 S

H. 0.

8

6

14

'20 fT

H. 0.

3

2

5

—

— —

Totals ... ... 105 ... ... 1

* These families contain one or more ^ ^ which are radiated (see table on p. 135).

It is clear that the recessive form breeds true, although the numbers
are not large, owing to disease which was particularly severe in 1919.
The fact that a single specimen of grossidariata appeared in family

Fy, heterozygous for varleyata, inter se.

Imagines F.j.

Bred by

Melanic

Totals

Non-melanic

Family

Male

Female

Male

Female

Totals

—

Mr L. W. Newman

4

3

7

?

?

24

'17.1*

Ova from Rev. G.

. H.

Raynor

1

1

2

2

i

3

'17 F

H. 0.

2

3

5

5

3

8

1.17

Rev. G. H.

Raynor

7

2

9

14

6

20

2.17

^j

11

8

19

23

25

48

3.17

13

2

15

22

20

48

4.17

99

1

2

3

13

2

15

7.17

4

5

9

20

12

32

8.17

,^

5

2

7

13

10

23

9.17

99

2

7

9

8

16

24

12.17

9 9

1

1

2

5

7

12

13.17

1

—

1

14

11

25

21.17

jj

2

2

4

7

6

13

22.17

1

—

1

2

1

3

32.17

9 J

2

. — .

2

12

5

17

'18 H

H. 0.

—

1

1

—

—

—

'18 T*

H. 0.

4

2

C

9

11

20

'19 N

H. 0.

1

—

1

—

—

—

'19 V*

H. 0.

4

—

4

8

9

17

Totals

107 (23

7o)

352 (77 7 J

Expec\

tat ion

115

344

These families contain one or more ^ ^ which are radiated (see table on p. 135).

126 The Inheritance of Wing-Colour in Lepidoptera

'18/ cannot be of any significance, as the ova might very easily have
been contaminated before they were received.

By mating together the F^ heterozygotes, an F^ generation is pro-
duced, which consists approximately of one varleyata to three others,
either grossulariata or lacticolor.

I am very much indebted to Mr Raynor for the data of his 1917
families, which are included in the above table. They supplement the
other figures, and the totals are fairly close to the expected ratio of 3 :1.

All the parents are grossulariata, but several of them are dark, and
the cf parent of 1.17 and the $ parent of 22.17 approach var. hazeleigh-
ensis. Since the parents are in a number of cases heterozygous for
lacticolor, this variety occui's among the non-melanics in several families.

The result of mating varleyata to F^ heterozygotes shows an almost
equal number of melanic and non-melanic, as follows:

Varleyata x F^.
RR X DR.

Ii

uagines

Non-meli

Melanic

mlc

Family

Female x Male

Male

I<"emale

Totals

MaiT"

Female

Totals

'18 G*

lacticolor x varleyata

15

13

28

13

14

27

'18 M*

type X varleyata

1

2

3

1

2

3

'18 P*

X ,,

12

2

14

12

7

19

'ISR*

varleyata x type

5

8

13

6

8

14

'19 A*

type X varleyata

7

7

14

8

4

12

'19 ii*

lacticolor x type

14

10

24

12

8

20

'19 C*

,, X ,,

7

2

9

6

4

10

'19£*

varleyata \ ,,

9

3

12

5

1

6

'19 G

,, X ,,

2

—

2

1

—

1

'19 if

,, X ,,

—

1

1

2

—

2

'19 J

lacticolor x varleyata

1

1

2

2

1

3

'19ilf

varleyata x type

4

4

8

5

3

8

'19 y*

,, X ,,

8

6

14

12

11

23

'19 JJ

lacticolor x varleyata

—

1

1

—

—

—

'19 S*

„ X „

11

7

18

11

8

19

'19 A'*

type x varleyata

4

5

9

2

4

6

'19 r

varleyata x lacticolor

3

2

5

2

2

4

'19 a

type X varleyata

2

5

7

3

G

9

'19/3

,, X ,,

3

1

4

—

5

5

'195

,, X ,,

—

2

2

1

1

2

'19 f

varleyata x type

1

4

5

2

1

3

'19 7J

,, X ,,

10

9

19

17

9

26

'19 d

lacticolor x varleyata

—

1

1

—

—

—

'19 TT

,, X ,,

—

2

2

1

3

4

'19 X

X „

—

—

—

—

1

1

'19 p

type X „

1

—

1

—

2

2

'19 w

varleyata x type

—

—

—

—

1

1

'201

exquiitita x lacticolor

6

4

10

2

4

6

'20 J

lacticolor x exquisita

2

1

3

2

2

4

'20 L

type X varleyata

8

1

9

8

1

9

'20 ill

exquisita x type

16

9

25

9

6

15

'20 i^

varleyata x lacticolor

9

2

11

4

3

7

'20 P

,, X ,,

3

1

4

2

—

2

Totals ... 164 116 280(51%) 151 122 273(49%)

Expectation ... 274-5 274-5

These families contain one or more ^ ^ which are radiated (see table on p. 135).

H. Onslow, 127

These figures leave no possible doubt of the recessive behaviour of
the melanic variety when mated with grossulariata. It will be observed
that in both melanics and non-melanics of the last table the cTc/ are
considerably in excess of the $ $ . These numbers may be compared with
those given in the Table on p. 135, where attention is called to this sex
ratio.

Var. hazeleighcnsis, and intermediates.

The normal variety oivarleyata is shown in Plate XIX. The (quantity
of white round the base of the wings varies from the amount seen in
Fig. 1, where it includes the discoidal spot, to the amount seen in Fig. 10,
where practically all the white is gone from the fore wings. It has
already been mentioned that the F^ generation, from varleyata x grossu-
lariata, often gives individuals which are darker than the normal type
form. Fig. 5 shows a typical grossidariata ; Figs. 6, 11, 12 are forms with
increasing amounts of black pigment. The blackest individual. Fig. 12,
corresponds to var. hazeleighensis, and has fore wings all black, or with
a few flecks of white. It will be noticed that its hind wings are not
very different from those of the type insect Fig. 5, although the fore
wings have actually more black than some examples of varleyata.

In order to gain an idea of the amount of black that was present
in the intermediate varieties, an attempt was made to measure the
extent of the black areas in the different individuals, and to express
this as a percentage of the total area of the wing. The following pro-
cedure was found to give very satisfactory results. The insects are
pinned to a board in such a position that when viewed through a camera
lucida\ a magnification of two diameters is obtained, A drawing is then
made of the outline of the wing and of the black pattern. An accurate
measurement of the different areas can then be made by means of a
planimeter, the maximum error of which is not more than 1 per cent.

This method could with advantage be extended so as to apply to
other cases, where the genetics of pattern factors are being studied,
such as many problems in the inheritance of coat colour, involving more
or less elaborate patterns upon white or coloured grounds. This might
be done at least in all cases where the skin could be spread out flat,
and would surely be preferable to assigning the specimens to arbitrary
grades by means of the eye.

- A suitable instrument is sold by Lechertier Barbe, 95 Jermyn Street, S.W. A set of
lenses is supplied with the instrument, which are said to enable the operator to make
enlargements or reductions to any desired size.

128 The Inheritance af Wing-Colom- in Lepidoptera

In measuring the black pattern of grossulariata, only the fore wings
were considered, since the variation in the hind wings is very slight, as
may be seen in the plate. Moreover, only a single wing was measured,
since the labour involved is very considerable, and the variation between
the two wings of the same insect is usually insignificant. In each case, the
sum of the black areas on one fore wing was expressed as a percentage
of the total area of that wing ; the values so obtained were then plotted
as distributions ^ It at once became evident that there was a factor, L,
localising the black pigment to the normal grossulariata pattern of the
wings. The effect of I is to allow the pigment to spread, until the fore
wings are as black as in var. hazeleighensis. This localising factor L was
not always carried by varleyata, the factor I being often introduced by
matings with certain dark strains. On account of the increased prices
they command, such dark strains are prized by dealers, who would
probably pair any varleyata to dark rather than to the ordinary type
individuals, in the belief that they would be more likely to produce
varleyata in the next generation. It may here be recalled that the
factor causing the black rings round the body of var. nigrocincta^,
though apparently sometimes associated with var. hazeleighensis, has
been shown to be, like the factor I, unconnected with the black variety
varleyata.

Thus many of the F^ families coming from strains of varleyata
obtained from Messrs Newman and Porritt show an increased develop-
ment of black pigment, which is not apparent in the Fj families obtained
from Mr Raynor's strain. Moreover it has been observed that in addition
to this factor I, femaleness, or some factor associated with it, prevents
the black pigment in the wings of the females from spreading to the
extent that it does in the males. A rather similar case has already been
described in var. lutea^ in which the intensity of the yellow pigment is
considerably lower in the females than in the males.

Eeference to Text-fig. 1 shows family '19 Z, the ^i generation from
a grossulariata $ (33 per cent, black) x varleyata (/. It can be seen
from the two curves plotted for a certain number of the cT (/ and $ $ ,
purposely selected at random, that the extent of the black pattern is
practically the same in both sexes. The average amount of the black
pattern was found to be about 64 per cent, of the total area of the fore

1 The method adopted for making these distributions, as well as the percentage
frequency distributions, was exactly the same as that described by Onslow, H., Journal of
Genetics, Vol. viii. No. 4, p. 209, September 1919.

2 Onslow, H., loc. cit. ^ Onslow, H., loc. cit.

H. Onslow

129

wings. This is more black than is usually found in most type insects, but
it is considerably less than that shown in Text-figs. 4 and 6 (about 80 7^),
which are constructed from iiusects clearly belonging to dark strains.
It is of course probable that there is more than one factor, which by
modifying the pattern produce a whole series of insects, in which the
area of black varies greatly.

-:i 100

"^- 90

\ J,

V 18 G Q 9

^->^

-X--.

M9 zggr-o

$ Parent of
'19 Z

No. of Insects
Text-fig. 1 (cf. Text-fig. 2). Curves showing the distribution of the
values of the black pattern in Fi^ ^ and 5 $ from the cross
grossulariata $ x varleyata ^ ('19^), which appear to carry L,
the factor localising the black pigment. Similar curves are shown
for the black values of varleyata in a cross (Fj) $ x varleyata
c?('18G).

The above points are best seen in the frequency distribution diagram
(Text-fig. 2). The curves for both the «/'(/' and the $ ? of family 19 Z are
similar. They have their maxima at the same point, and they include
much the same area.

In the diagrams showing the frequency distributions the percentage
of the black area is given along the base line, and the percentage fre-
quency of any given black value is shown on the left-hand side. Thus,

130 The Inheritance of Wing -Colour 171 Lepidoptera

the maximum of the curve denotes that percentage of black which occurs
most frequently. In the case of the curves showing the distributions of
the black values, the figures along the base line denote the number
of individuals. Each individual is represented by a cross, a circle or
some other sign, which is the same for both sexes of any given family.
The percentage of black in the pattern is shown on the perpendicular
at the left-hand side. The name of the family and the sex are written
against each curve, whereas the percentage of black in the grossulariata
pattern of each parent is shown by an arrow at the correct point on
the right-hand side.

1) 30

1 1

1 1 --I r

1 1 r-

, 1

-

I

,

1
1

&;

\ /

\

t)^

N ,'

A \ <tk

\

O) 1

/ \ > »/' \

\

' '/

^ ^y ''

y I

^7

'V '

y

1 L i 1 L_

\

9 Parent of '19 Z Percentarje of Black

Text-fig. 2 (cf. Text-lig. 1). Diagrams showing the frequency distribution of the values
for the black pattern in the ^ ^ and ^ '^ ol the families shown in Text-fig. 1.

Text-fig. 3 shows a group of families from a dark strain of Mr New-
man's, which carries the factor, I, leading to an increase of the black
pattern, but not the factor for varleyata, thus showing conclusively the
independence of these two factors. The appearance of these curves is
very different from that of '19 Z (Text-fig. 1), and the salient points are
best seen in the frequency distribution. Text-fig. 4, in which the males and
females of all three families are combined in two curves. The average
extent of the black pattern is obviously greater than in type strains, and
this increase of black is greatest in the curve for the males, as may be seen
from the fact that its maximum falls to the right of that of the $ $ curve.
The average value for the black pattern, when both </ ^ and $ $ are con-
sicjered, is about 80 per cent. ; that of the $ % about 78 per cent., and that

H. Onslow

131

S, 100

?J 90- , V ~x

CL| 70

K VIS F9Q'.
jl8 AQQ -^i

-X- -K "

•18 C0rj

■■;i8 F ccr

1

iii8 c 99

Parent of '18 A

Parent of "18 C

Parent of '18 F
Parent of '18 A

Parent of '18 C

No. of Iimects
Text-fig. 3 (cf. Text-fig. 4). Curves showing the distribution of the values of the
black pattern in ^ ^Jand ? $ from three families of grossulariata x grossu-
/ari«te ('18 A, C, F). They carry the factor, I, leading to an increase
01 the black pattern which aflfects ^ ^ more than 9 9 , but not the factor
lor varleyata.

a, 20

'18 A,C. F. 99

'18 AC FvCTO

Percentage of Black
Text-fig, 4 (cf.Text-fig. 3). Diagram showing the frequency distribution of the values
tor the black pattern of the ^ ^ and the ? ? in the three families of Text-fig. 3.
Journ. of Gen, xi g

132 The Inheritance of Wing-Colour in Lepidoptera

of the cf </ about 85 per cent. If the families are considered separately,
the darkest parents, as might be expected, appear to have the blackest
offspring. This may be seen by a glance at the values for each parent,
shown by the arrows on the right-hand side, where families '18 A and
'18 G (Text-fig. 3) seem to have the darkest parents. The (/* parent of
'18^ was unfortunately lost, but a note remains describing it as rather
darker than type, and probably with a black value of about 60 per cent.
The curves shown in Text-fig. 5 represent families which carry both the
factor, I, leading to an increase of the black pattern, and the factor for

f.-~

^Hs^

'••..^'16 Z^x ^-C^-^^^PCTCT

+s^ '■'.': d'c? '18 n^ytj ° — »-,

Vl8\ -.'16 ZQ9

■'■■••••.\.QQ+ \

\--.... '■••.'.18 UeTo-

'i8p\ '■••.;

99

■■,'18 u99

Parent of '18 R

Parents of '18 P,
and '16 Z

Parent of '18 U

No. of Insects

Text-fig. 5 (cf. Text-fig. 6). Curves showing the distribution of the values for
the black pattern inFi ^ ^ and $ $ , from the cross grossulariata 5 x var-
leyata ^ {'16 Z and '18 U). Similar distributions are given for grossu-
lariata ^ ^ and $ $ from the cross F^ x varleyata ('18 P and '18 jR). The
effect of the factor I, which leads to an increase of the black pigment, can
be seen in both sexes, but more so in the ^ ^ than in the $ $ .

varleyata. Families '\Q Z and '18 C/" are the F^ generation from the cross
grossulariata % x varleyata (/ , and their average black value is greater
than in type specimens, and also greater in the </ cT than in the $ $ . The
two other families in the same figure, '18 P and '18 i^, are grossulariata

H. Onslow

133

from the cross (F^) x varleyata, and they illustrate practically the same
points as the previous insects. This is best seen in the frequency
distribution diagram (Text-fig. 6) in which the (/•(/ and ? % of all four
families have been combined in two curves. The average black value
is about 80 per cent., that of the % % being 75 per cent, and that of
the cTc/ ^'5 per cent., that is to say, the percentage of black on the fore
wings is (1) greater than in the type specimens, and (2) greater in
the c/c/ than in the %%.

1

30

-

-

5-

25
20
15
10

-

A

1

1

\ 1

\
1

1 »

1 *

\
\
\

5

■

1

,/

•16 Z

'18 P.

R

U

99/

'-16 2

•18 P. f
R U.'

\

-

0

5

-lo"

20

30

40

'50

60

70 '

80

90

10

Percentage of Black

Text-fig. 6 (cf. Text-fig. 5). Diagram showing the frequency distribution of
the values for the black pattern of (J (J and 5 2 in the four families
in Text-fig. 5, combined together.

As might be expected, it makes no apparent difference which parent
is the varleyata; for instance, family '18 R comes from a varleyata ?
X grossulariata (/ (88 per cent, black), and- family '18 P comes from a
grossulariata $ (75 per cent, black) x varleyata ^ (see Text-fig. 5). The
correlation between the black values of the parents and the offspring is also
clearly seen. The arrows on the right-hand side of Text-fig. 5 show the
parental values, which in the case of '18 JJ is 57 per cent. The offspring
of this mating are not so dark as the other two families, the parental
values of which are 75 and 88 per cent. Readings were made from
only a small number of families, owing to the great labour involved,
but the curves given are representative of all the families in the pre-
ceding tables, in which the two classes — those which inherit the localising
factor Z, and those which do not — can be very clearly differentiated.

9—2

134 The Inheritance of Wmg-Coloiir in Lei^idoptera

A consideration of families 1.17 and 22.17 {hazeleighensis x grossulariata)
seems to show that the factor I is probably recessive when crossed with
the type form, but the fact that there may be more than one modifying
factor, no doubt, complicates the case considerably.

It has already been mentioned that there is a considerable variation
in the size of the white collar ofvarleyata (see Plate XIX, figs. 1 — 4 and
7 — 10). Not improbably, this variation has been introduced from the
dark strain of grossulariata, either by or in association with the factor I.
However, the variation is not found to be most highly developed in
those families which show the greatest increase of the black pattern.
If the restriction of the white collar of varleyata were clearly associated
with an extension of the black pattern of grossulariata, it would un-
doubtedly behave in the same way. Measurements were made of certain
varleyata ^ ^ and $ $ of family '18 G from a cross, F^%y. varleyata ^ .
Text-fig. 1 shows the percentage of black on the fore wings of these var-
leyata,. This varies between 80 and 100 per cent., and moreover there
is a small difference between the black values of the (/ ^ and $ % ,
amounting to about 4 per cent, in favour of the (/(/, a value which
should be above the experimental error, though not of much significance.

The relationship of the curves for the two sexes may best be seen
by reference to Text-fig. 2. It seems possible that even if there were an
inhibiting factor associated with femaleness, its effect might be very
difficult to measure, on account of the small area of white present in
the variety. It was accordingly considered not worth while to under-
take a more extensive series of measurements.

Vars. actinota and leucosticta.

Another interesting variety of varleyata, which appeared during the
experiments, was that showing a varying amount of radiation (Plate XIX,
figs. 7, 8 and 9). This variety, which shows a white radiation upon a
black ground, must not be confused with var. radiata\ in which the
radiation consists of a prolongation of the black marginal spots. The
amount of radiation varies from a mere fleck of white on the hind wings,
as in Plate XIX, fig. 7, to the striking wedge-shaped markings of fig. 9,
which can be seen to extend even to the fore wings. For the first form
Mr Raynor has suggested the name var. leucosticta, and for the fully
radiated form var. actinota. This character behaves quite independently
of the factor L which leads to a restriction of the black pigment.

1 Onslow, H., Journal of Genetics, Vol. viii. No. 4 (see p. 221, and Plate IX, fig. 56).

H. Onslow

135

In the preceding tables the families in which these radiated in-
dividuals occur have been indicated by an asterisk. The number of
radiated individuals in these families is now given in the following table:

Fawdlies containing the radiated varieties actinota and leucosticta.

Imagines
Melanic Non-MelaDic

Female x Male

Radiated
Male

Non-radiated
Male Female

Totals

Family

Male

Female

Totals

'18i

varleyata x varleyata

1

7

5

13

-

1

1

'17.-1

type X type

1

(1)

1

2

2

1

3

'18 r

„ X ,,

1

3

2

(J

9

11

20

'19 F

,, X ,,

4

w

—

—

4

8

9

17

'18 G

lacticolor x varleyata

1

14

13

28

13

14

27

'18 M

type X

1

—

2

3

1

2

3

'18 P

„ X ,,

7

(3)

5

2

14

12

7

19

'18 ii

varleyata x type

3

2

8

13

0

8

14

'19^

type X varleyata

2

5

7

14

8

4

12

'19 B

lacticolor X ,,

3

i^)

11

10

24

12

8

20

'19 C

„ X

2

5

2

9

6

4

10

'19 £

varleyata x type

9

{9)

—

3

12

5

1

6

'19 Q

M X 5»

5

3

G

14

12

11

23

'19 S

lacticolor x varleyata

1

W

10

7

18

11

8

19

'19 X

type X ,,

2

2

-5

9

2

4

6

Totals

43 (16)

67

73

183

107

93

200

The broods come from three different types of mating. The figures in
brackets against the radiated specimens denote the number of those
heavily radiated (var. actinota), the remainder are all var. leucosticta.
The most interesting fact to be noted is that all 43 radiated insects are
(^ (/ , not a single $ showing any trace of this character. The association
of this variation with maleness seems to be complete, for I am informed
by Mr Raynor that he has bred 104 specimens of var. actinota in 1920..
and almost as many of var. leucosticta, and they do not include a single
female. Moreover, Porritt^ mentions a similar case (see Family 1906 in
the second table, p. 125) in which all the (/•(/ were radiated, but not a
single $ . Recently, however, this observer has bred fi^om radiated stock
a leucosticta $ .

The greatest number of radiated individuals is found in family '19 E
which has 9 actinota ^^ ^ , but the study of the pedigree of this and the
other families gives no clue to the type of inheritance followed. When
a radiated insect is mated to grossulariata (i.e. family 18 0, see first

1 Porritt, G. T., Ent. Mo. Mag. London, Vol. xliii. p. 12, 1907.

136 The Inheritance of Wing -Colour in Lepidoptera

table, p. 124), the character cannot appear in the type offspring, even
if the necessary factors were present, becuase radiation is essentially a
variation oivarleyata, and could not show on the grossulariata pattern.

It would seem as if the white radii ought to be a remnant of the
dominant white ground of grossulariata, which shows through the black
of varleyata, and therefore the radiated varieties should be dominant to
varleyata. Inspection of the last table, however, shows that in family
18 / one radiated insect was bred from two varleyata parents, but even
if this radiated insect is accounted for, by supposing it to have been
introduced accidentally like the single grossulariata in the same family,
it is seen that the black parents of all the other radiated insects are
normal. Supposing radiation to be dominant the grossulariata parents
would have to be heterozygous for the factor, in which case the ratio
between radiated and non-radiated varleyata should have been one of
equality, whereas it is 43 radiated to 140 non-radiated.

From the scanty data in the last table it is impossible at present to
determine the inheritance of this variation with any certainty, but the
fact that there is in Abraxas a factor linked to maleness, as well as one
linked to femaleness (lacticolor-), is of some interest, for, except in
Drosophila this has not been recorded. Suppose in the first place it is
assumed that the case is as simple as possible, the factor for radiation
being sex-linked to the male, and recessive to varleyata. Then if both
parents are heterozygous for the recessive factor, considering the melanic
offspring, which alone can show the radiation, there should be one
radiated male, one normal male and two normal females, or 4575 radiated
to 137"25 non-radiated. This does not differ much from the ratio actually
found, which was 43 radiated to 140 non-radiated. The suggestion,
however, takes no account of a very important point in connection with
the ratio of the sexes, to which attention has not as yet been drawn.

Among the type insects the sexes approach equality, but among the
melanics there are 110 (/</ to 73 $ $. This looks as if there is a
missing group of $ $ which correspond to the radiated cTcf, a fact
suggesting the operation of a lethal factor. It is curious, however, that
the missing $ $ should be restricted to the melanics, and it looks as if
the lethal factor is only effective when a female is both radiated and
melanic. If this were so, the ratio of the sexes would be much closer to
the figures actually obtained, but on the other hand the number of
radiated (/(/ should be only 26, instead of 43 as recorded.

In trying to account for this unusual sex ratio, if only the factor for
radiation were in spite of the evidence found to be dominant, an analogy

H. Onslow 137

suggests itself with the inheritance of "notch" wing in Drosophila\ This
variation is dominant and sex-linked to the female ; associated with the
homozygous condition is a lethal factor, causing homozygous females to
die, but the heterozygous females live. Moreover, all males carrying
" notch " die, because the Y chromosome is supposed to be empty, and
therefore the insects do not possess the normal allelomorph for " notch "
which, in the heterozygous females, counteracts the lethal factor. If in
the case of radiation, conditions were the same, except that the factor
was linked to the male sex, then varleyata % % (het. for radiation) x
varleyata (/(/ should give equal numbers of varleyata cfcT and $ $ , and
the same number of radiated c/cT, but no radiated ? %.

When more data are available, it will be easier to decide whether a
lethal factor is present or not ; also whether " notch " wing, colour-
blindness, or some other mode of inheritance is the most correct type
on which to base the explanation of the anomalous sex-ratio observed.
At present only some of the possibilities can be indicated, but it should
be remembered that, whatever explanation is adopted, it must take into
account the fact that at least one radiated female has been produced.
Moreover, attention may be called to the sex ratio in the varleyata x F^
families (see p. 126) which show a considerable excess of males in both
black and type insects, no matter whether the radiated families, marked
with an asterisk, are included or not.

If the factor for radiation is supposed to be recessive, and sex-linked
to the male, Professor Punnett has raised an interesting question.
Supposing the missing chromosome observed by Doncaster in the all-
female strains of grossulariata to be a sex chromosome, then according
to the chromosome hypothesis, the factor for radiation Avould be carried
by the Z chromosome. A female, heterozygous for radiation, would have
that factor in her Z chromosome. Thus if Morgan's suggestion that the
W chromosome is a "dummy" without factors, is accepted, then all such
heterozygous females must be radiated in appearance. To avoid this
difficulty, either the W chromosome must carry the normal allelomorph
which prevents the development of the white radiation, or the W
chromosome itself must be able to inhibit radiation in a female that is
genetically actinota, which is after all much the same thing. From his
work on poultry. Professor Punnett- is inclined to believe that if sex
chromosomes exist in these cases, they can carry factor.s.

1 Morgan, T. H., Publications of the Carnegie Institute, No. 343, 1919.
■^ Punnett, E. C, Journal of Genetics, Vol. xi. No. 1, p. 37, April, 1921.

138 The Inheritance oj Wing-Colour in Lepidoptera

Conclusions.

1. In Abraxas grossulariata the factor for the black pattern of
var. varleyata behaves as a simple mendelian recessive both to the type
form and to var. lacticolo7\ A new variety, exquisita, which probably
contains both recessive factors, that for varleyata and that for lacticolor,
will be dealt with in a subsequent communication.

2. The type grossulariata carries a factor L, which localises the
black pigment to the normal pattern. The factor I allows the black
pigment to spread until it almost covers the fore wings as in var.
Iiazeleighensis. This variation appears to behave as a recessive when
mated to type insects. It is inherited independently of the factor for
varleyata, but is often introduced by crossing this variety with dark
strains of grossulariata.

3. A method is described by which the black pattern can be
measured, and its area expressed as a percentage of the total extent
of the fore wings. The values of the black pattern in several families
were obtained, and were plotted as distributions and as percentage
frequency distributions,

4. These diagrams show that in certain families the amount of
black in both sexes is considerably greater than in the type form, and
that the darkest parents have as a rule darker offspring than lighter
parents. Moreover, the percentage of black in the (/c/* is considerably
greater than in the $ $ , suggesting that femaleness or some factor
associated with it prevents the full development of the pigment.

5. The size of the white collar in var. varleyata varies considerably.
This is possibly in some way connected with the factor I, for measure-
ments of one family show that the ? $ have slightly less white than
the (/(/, but the amount can scarcely be considered significant.

6. A variety showing a fluctuating amount of white radiation on
the black ground of varleyata has been found, which appears to be
linked to the male in a manner analogous to that in which lacticolor is
linked to the female. At first sight it appears that this white radiation
should be part of the dominant white ground o^ grossulariata. Never-
theless, there seems to be some evidence to show that it is recessive to
varleyata. Since this radiation is a modification of the black ground, it
is clear that it cannot show in grossulariata insects, even if they carried
the factors necessary to produce it. The ratio of one radiated </ to

JOURNAL OF GENETICS, VOL XI. NO. 2

PLATE XIX

• ■■* •

11

\^:)i:^

H. Onslow 139

three non-radiated ^^ and % % , which would result from the hypothesis
that radiation is recessive, and that both parents were heterozygous, is
closely approached by the actual figures. It is, however, an hypothesis
that takes no heed of the remarkable sex ratio which shows a large
preponderance of males.

There are not at present enough data to determine whether this
disturbance of the normal sex ratio can be explained better by the action
of a lethal factor or not, but the work is being continued.

In conclusion my thanks are due to the British A,ssociation for
defraying part of the expenses of this research by a grant during the
year 1920. I am also much obliged to the Rev. G. H. Ray nor both for
the data and for the living material he has so kindly allowed me to use,
and to the editors for their kindness in reading this paper. I am especially
grateful to Miss Helen Moodie for her care of the larvae, and the
resulting large families.

DESCRIPTION OF PLATE XIX.

Abraxas grossnlariata and varieties. Natural size.

1. Var. varleyata $ with broad white collar showing discoidal spots.

2. Var. varleyata ^ with black stripe crossing collar from basal shoulder knot to
discoidal spot.

3. Var. varleyata $ with narrow collar showing spread of black pigment outwards from
basal shoulder knot.

4. Var. varleyata 9 except for a white fleck the collar is restricted to the hind wings.

5. Grossnlariata 5 > type, showing normal pattern.

6. Grossnlariata $ showing increase of black pigment, also stripe as in No. 2.

7. Var. leucosticta ^, a black insect showing traces of two white radii near margin of
hind wings.

8. Var. leucosticta ^ showing number of incomplete radii on hind wings.

9. Var. actinota ^ showing fully radiated hind wings with traces of radiation on fore
wings.

10. Var. varleyata ^ showing, like No. 4, almost complete absence of white collar on
fore wings.

11. Var. hazeleighensis ^ showing considerable increase of black piginent, though the
white areas are rather too extensive for this variety.

12. Var. hazeleighensis ^ showing black pigment almost entirely covering the fore wings.

Note. Those insects whose fore wings show a great increase of black pigment have
little or no corresponding increase iu the hind wings.

SEX-REVERSAL IN FROGS AND TOADS. A REVIEW
OF THE RECORDED CASES OF ABNORMALITY
OF THE REPRODU(JTIVE SYSTEM AND AN
ACCOUNT OF A BREEDING EXPERIMENT.

By F. A. E. CREW.

{Papers from the Animal Breeding Research Department,
The University/, Edinburgh.)

(With Twenty-three Text-figures.)

The abnormalities which have been recorded can be so tabulated
that the first case most nearly approximates to the normal female and the
last the typical male, with respect to the nature of both primary and
secondary sexual characters. Thus arranged, it is seen that the cases
furnish an almost complete series of gradations which range from an
individual almost completely female, to one almost completely male, and
that the conditions found readily appear to be merely graded stages of
a single process.

A. RANA.

Division I. Those Cases in which the Secondary Sexual
Characters were not Typically Male.

Group A. Those cases in whicJi, on inspection, one gonad was an
ovary and the other an ovo-testis.

Sub-group 1. The ovary was normal in appearance.

1. Bourne. R. temporaria. Adult. Secondary sexual characters
were not described.

Right gonad. An ovary.

Left gonad. An ovo-testis. The larger ovary-portion bore along the
anterior half of its inner border the smaller testis portion.
■ Vasa efferentia could not be observed.

Seminal vesicles not present.

Miillerian ducts well-developed.

On section. Ovary-portion of the ovo-testis contained well-developed
ova; testis portion contained motile spermatozoa; and there was no
distinct line of separation between ovary- and testis-portions.

142

Sex-Reversal in Frogs and Toads

2. YouNGMAN. R. te)iipo7^aria. Adult. Killed June. Secondary
sexual characters male but poorly developed.

Right gonad. An ovo-testis. An ovoid non-pigmented testis-portion
suspended from the ventral surface of a small but apparently normal
ovary.

Left gonad. An ovary of normal appearance. •

Vasa efferentia. " There appears no means of exit to the exterior for
the spermatozoa, unless they burst into the coelom, for I can make out
neither macroscopically nor microscopically by sections, any trace of vasa
efferentia, and sections of the ureter do not show the presence of sper-
matozoa therein."

Seminal vesicles not present.

Fig. 1.

Fig. 2.

Mlillerian ducts exceedingly well-developed with eggs within the
convoluted portions of each and with uterine segments crammed to dis-
tension.

On section. Left ovary and the dorsal ovary portion of the right
gonad showed the following structure. Numerous normal ova were pre-
sent, but there were also cavities among the ordinary egg-containing
chambers, and of the same size, which were filled with black pigment
(" representing perhaps degenerate eggs "). The ventral testis-portion of
the ovo-testis showed testicular structure, and contained immature and
mature spermatozoa and, in addition, peculiar eosinophilous bodies.

F. A. E. Crew

143

Sub-group 2. The ovary tuas abnormal in appearance.

3. Huxley. (1.) R. temporaria. Adult. Killed April. Secondary
sexual characters as those of a Avell-developed female. No male characters
were present.

Right gonad. An ovary of usual size but more deeply pigmented
than usual.

Left gonad. An ovo-testis consisting of an ovary of unexceptional
size, which bore upon its ventral surface a small nodule of testis material.

Vasa efFerentia were not present in connection with the right gonad
but there were two in association with the testis-portion of the ovo-
testis.

Seminal vesicles were present, being very small and spindle-shaped.

Miillerian ducts were well-developed and ova were contained therein.

Fig. 3.

Fig. 4.

On section, the right gonad was seen to be entirely ovarian in struc-
ture, the fibrous tissue of the ovarian wall was everywhere hyperplastic
and there was excess of pigment.

The left gonad consisted of two portions. The smaller was composed
of spermatic and the larger of ovarian tissue similar in structure to the
ovary of the opposite side. The spermatic tissues were healthily normal
for the most part, but ova were found amid them, within and between
the seminal tubules. An ovum within a tubule lay among spermatozoa
which were deformed and degenerate, a consequence possibly of the
pressure exerted by the ovum. The seminal tubules in the neighbourhood
of an ovum which lay between the seminal tubules were contorted and

144 Sex- Reversal in Frogs and Toads

misshapen and the intertiibular connective-tissue in this area was
hyperplastic.

There was ciliated epithelium upon the peritoneum, and the ova
within the Mullerian ducts were all degenerate.

4. Crew^. (1.) R. temporaiia. Adult. Killed April. Secondary
sexual characters male but imperfectly developed.

Right gonad. An ovary, somewhat firmer and more strongly pig-
mented than usual.

Left gonad. An ovo-testis, having the appearance of an ovary rather
more deeply pigmented and firmer to the touch than usual, which bore
along its inner border three small nodules of testis-substance.

Vasa efferentia^few in number— were present on either side.

Seminal vesicles were present, being small and fusiform.

Mullerian ducts very well-developed, ova within uterine segments
and convoluted portions.

On section. Right ovary and ovary -portion of the left gonad entirely
ovarian in structure, but practically all of the more mature ova showed
signs of degenerative changes. There was widespread increase of the
pigment normally present in an ovary and the connective-tissue was
distinctly hyperplastic. The three nodules upon the inner border of the
left gonad had the structure of normal spermatic tissue.

There were ciliated cells upon the peritoneum. The Mullerian ducts
contained only degenerate and spurious ova.

5. Marshall. (D.) R. temporaria. Adult. Secondary sexual cha-
racters not described.

Right gonad. An ovo-testis. The larger testis-portion had the
ovarian tissue disposed along its outer border. The body of the testis-
portion presented upon its ventral surface a very deep transverse groove.
At the anterior end of the ovary-portion was a deeply pigmented lobule
which was bent back over the dorsal surface of the gonad.

Left gonad. An ovary more strongly pigmented than usual and
divided into six lobes by well-marked constrictions.

Vasa efferentia on the right side as those of the normal male ; two
or three thin-walled tubules connected the left ovary to its kidney.

Seminal vesicles not present.

Mullerian ducts well-developed.

On section, the greater part of the right gonad had the structure of
normal testis. The pigmented band, which ran along its outer border,
extended inwards about a quarter of the way through the substance of
the testis, and consisted of masses of pigment which followed the lines

F. A. E. Crew 145

of the intertubular connective tissue, this being hyperplastic in these
{ireas. The polygonal masses of pigment were, in most cases, entirely
outside the seminal tubules which had become crushed together and
obliterated by the increase of the intertubular connective-tissue, and
only in the last stages of atrophy of the tubules were pigment found
within them. The dorsal lobe of the ovo-testis and the left gonad
showed exactly similar structure, that of ovary with hyperplastic con-
nective tissue and great abundance of pigment. Few ova were found
which were in a thoroughly healthy condition and the great majority
were in various stages of degeneration, the protoplasm having shrunk
from the follicle wall, the nuclei being of small size, and the follicle in
some cases invaded by pigment.

" It is worthy of note that while the male portion of the essential
glands is absolutely normal throughout, the female portion, though
more bulky, forming the whole of the left gland and part of the right
one as well, is not normal in any part, the majority of the ova showing
more or less well-marked signs of degeneration, and the whole structure
exhibiting very obvious sclerotic induration, which must be regarded as
pathological."

Fig. 5. Fig. 6.

Group B. Those cases in which, on inspection, both gonads were ovo-
testes.

6. Hooker. (B.) R. fasca. Adult. Killed October. Secondary
sexual characters imperfectly developed male, but in addition the skin
of the back was warty.

146

Sex-Reversal in Frogs and Toads

Right gonad. An ovo-testis. To the outer border of the larger
testis-portion the smaller ovary-portion was attached.

Left gonad. An ovo-testis, having the form of a normal-sized six-
lobed ovary which had a small nodule of spermatic tissue attached to
the middle of its inner border.

Vasa efferentia of the usual male pattern.

Seminal vesicles present, being small and spindle-shaped.

Mullerian ducts well-developed.

On. section of the left gonad (it was assumed that the right was
similar), the testis-portion, discrete from the ovary-portion, showed the
structure of normal testis ; the ovary-portion contained large numbers
of ova, some immature, some mature, and many showing various stages
of degeneration, being associated with black pigment.

7. Smith. R. temjwraria. Adult. Secondary sexual characters male
but imperfectly developed. (The skin below the throat was lighter on
the left side ; the typically male pad was present on the index finger of
the right side only.)

Fig. 7. Fig. 8.

Right gonad. An ovo-testis. The larger testis-portion presented
several constrictions upon its surface and bore along the outer border
the smaller ovary-portion, on the surface of which the outlines of ova
could be distinguished.

F. A. E. Crew 147

Left gonad. An ovo-testis consisting of a larger ovary-portion bear-
ing upon its inner border a nodule of spermatic tissue. There was a
second nodule of this tissue upon the outer border,

Vasa efferentia (of the usual male pattern ?).

Seminal vesicles not present.

Miillerian ducts — right moderately, left well-developed.

No sections made.

Group C. Those cases in which, on insj^ection, one gonad was an ovo-
testis and the other a testis. None.

Group D. Those cases in which, on inspection, both gonads were testes.
None.

Division II. Those Cases in which the Secondary Sexual
Characters were Typically Male.

Group A. Those cases in tvhich, on inspection, one gonad was an ovarg
and the other an ovo-testis. None.

Group B. Those cases in luhich, on inspection, both gonads were ovo-
testes.

8. BouLANGE. R. fusca. Adult. Killed December. Secondary
sexual characters typically male.

Right gonad. An ovo-testis of which the testis-portion constituted
three-quarters of the whole gonad and was situated anteriorly.

Left gonad. An ovo-testis of which the three-lobed ovary-portion
formed four-fifths of the whole and encompassed the testis-portion which
lay upon its inner border.

Vasa efferentia were present in association with the testis-portions
of both gonads.

Seminal vesicles were not present.

Miillerian ducts well-developed.

On section, the ovarian portions showed the structure of normal im-
mature ovaries : the testis-portions, that of normal testis.

9. Punnett. R. temporaria. Adult. Killed February. Secondary
sexual characters male.

Right gonad. An ovo-testis with a very large testis-portion bearing
a deep constriction upon its inner border. The ovary-portion was merely
a patch of ovarian tissue situated upon the outer border of the testis-
portion exactly opposite the constriction upon the inner border, and it
lay within a deep angle of testicular material.

Journ. of Gen. xi 10

148

Sex-Reversal in Frogs and Toads

Left gonad. An ovo-testis with a larger six-lobed ovary-portion,
upon the inner border of which, at about its mid-point, was situated a
small spherical lobule of testicular substance.

Vasa efiferentia entered into functional relationship with the sper-
matic and ovarian tissues of both gonads.

Seminal vesicles small and fusiform.

Miillerian ducts well-developed.

On section, the ovarian patch upon the right gonad contained but a
single ovum, and the rest of this gonad was purely testicular. The tes-
ticular patch upon the left gonad had the structure of normal testis but
at one point it contained a well-developed ovum.

Fig. 9. Fig. 10.

10. Crew. (2.) R. temjporaria. Adult. Killed June. Secondary
sexual characters typically male.

Right gonad. An ovo-testis having the appearance of a malformed
testis, irregular in outline and with its surfaces scored with deep grooves,
which bore along its outer border a broad band of dense black pigment,
which in some places was flatly applied to the surface of the testis and
in others was piled up into nodular prominences resembling ova.

Left gonad. An ovo-testis consisting of three lobes, of which the
anterior and middle were testicular, but bore a band of dense black
pigment similar to that of the opposite gonad along the outer border,
while the posterior lobule was ovarian, the j)igraent of which was con-
tinuous with the band upon the outer border of the other two.

Vasa efferentia present on both sides.

F. A. E. Crew

149

Seminal .vesicles, pigmented and of moderate size were present.

Mlillerian ducts were well-developed.

On section, the right gonad consisted of normal spermatic tissue in
all parts, save those involving the pigment upon its outer border. The
band of pigment was composed of dense masses of polygonal pigment
granules within a fibrous tissue matrix and included many ova, all of
which showed signs of degenerative changes. Ovary- and testis-portions
were quite discrete but the line of demarcation was very irregular and
the pigment appeared to ramify along the lines of the intertubular con-
nective-tissues of the testis-portion. The anterior and middle lobes of
the left gonad had a structure similar to the above, whereas the posterior
lobe was entirely ovarian but pathological.

There was ciliated epithelium upon the peritoneum.

11. Marshall. (B.) R. temporaria. Adult. Killed during the
winter. Secondary sexual characters typically male.

Right gonad. An ovo-testis having the appearance of an elongated
and irregularly shaped testis bearing an irregular band of black pig-
mented material along the anterior two-thirds of the anterior border, and

Fig. 11.

Fig. 12.

at the anterior pole, a clump of spherical bodies strongly pigmented and

closely resembling ova in appearance. At the extreme anterior end there

was a pigmented lobe separated from the main body of the gonad by

a deep constriction.

10—2

150 Sex-Reversal in Frogs and Toads

Left gonad. An ovo-testis similar in appearance to the right save
that there was no pigmented lobe at the anterior pole.

Vasa efferentia on both sides as those of the normal male.

Seminal vesicles small and spindle-shaped.

Miillerian ducts well-developed.

On section, the greater part of each gonad had the structure of
normal testis, but ova were found in considerable numbers lying between
the seminal tubules. The great majority of these ova were immature
and showed signs of degenerative changes. The pigmented bands were
composed of densely packed granules of pigment and hyperplastic
connective-tissue. The anterior lobe of the right gonad consisted partly
of normal testicular tissue and partly of ova in a state of extreme de-
generation. The ovarian tissues were not discrete from the spermatic
and were pathologicab

12. Crew. (3.) R. esculenta. Adult. Killed November. Secondary
sexual characters typically male.

Right gonad. An ovo-testis having the appearance of an irregularly
shaped testis with bands of dense black pigmented material running in
deep grooves along its twisted outer border.

Left gonad. An ovo-testis, consisting of six lobes, five of which were
ovarian, and the other, testicular in appearance. The testicular lobule
was placed centrally and medially in relation to the others.

Vasa efferentia, of the ordinary male pattern on the right : on the
left four in number and connected with both ovarian and spermatic lobes
of the gonad.

Seminal vesicles present but small and spindle-shaped.

Miillerian ducts well-developed.

On section, all those parts of the right gonad not including pigment
had the structure of normal testis. The pigmented nodules consisted of
densely packed polygonal masses of pigment granules and hyperplastic
connective-tissue. No ova were found amid this pigment, but near to
this two ova were seen lying between the seminal tubules of the sper-
matic tissue. The pigment had the appearance of following the lines
of the intertubular connective-tissue. The left gonad consisted of two
distinct portions. The five pigmented lobes were composed of patho-
logical ovarian tissue in which degenerate ova were included among
dense polygonal masses of pigment and hyperplastic connective-tissue.
The spermatic lobe, discrete from the rest, was entirely testicular in
structure. The renal vessels of both sides lay in deep pigmented
channels upon the surface of the kidneys and the pigment was identical
microscopically with that of the ovary-portions of the ovo-testes.

R A. E. Crew

151

13. RiDEVVOOD. K temporaria. In first year. Killed November.
Secondary sexual characters typically male.

Right gonad. An ovo-testis having the appearance of an excep-
tionally large testis bearing a small pigmented excrescence antero-
externally.

Left gonad. An ovo-testis with a 3-lobed ovary-portion situated
externally.

Vasa efferentia as those of the normal male.

Seminal vesicles present.

MuUerian ducts well-developed, more especially the left.

No sections were made.

14. Crew. (4.) R. temporaria. Adult. Killed May. Secondary
sexual characters typically male.

Right gonad. An ovo-testis which save for a small pigmented pit
upon the outer border had the appearance of a testis.

Left gonad. An ovo-testis having the appearance of an irregularly
shaped testis which bore along its outer border a prominent pigmented
crest, jet-black in colour and consisting of four spherical nodules in its
anterior part.

Fig. 13.

Fig. 14.

Vasa efferentia on both sides as those of the normal male.
Seminal vesicles well-defined.
Mullerian ducts moderately well-developed.

On section, the greater part of the right gonad had the structure of
normal testis. Sections involving the pigmented pit showed the presence

152 Sex-Reversal in Frogs mid Toads

therein of polygonal masses of pigment and broken down granular
material. The left gonad was mainly spermatic in structure but the
pigmented parts consisted of masses of pigment, hyperplastic connective-
tissue, and a few very degenerate ova. There was ciliated epithelium
upon the peritoneum.

Group C. Those cases in which, on inspection, one gonad was an ovo-
testis and the other a testis.

Sub-group 1. The testis was abnormal in appearance.

15. Ognew. R. temporaria. Adult. Secondary sexual characters
typically male.

Right gonad. An exceptionally large testis.

Left gonad. An ovo-testis with a small ovarian portion placed
laterally.

Seminal vesicles were present.

Miillerian ducts. Right one moderately, left one well-developed.

No sections were made.

16. Cole. R. temporaria. Young and apparently immature. Killed
during the summer. Secondary sexual characters typically male.

Right gonad. A testis somewhat enlarged and non-pigmented.

Left gonad. An ovo-testis with a larger densely pigmented ovary-
portion and smaller non-pigmented papilliform testis-portions at its
poles.

Vasa efferentia typically male on the right : on the left, only two
but " the divisions of the renal artery to the ovo-testis might well have
been mistaken for efferent canals."

Seminal vesicles not present.

Miillerian ducts. That of the right side was rudimentary and not
convoluted, but had a distinct and densely pigmented uterine segment.
That of the left side was pigmented, convoluted, and moderately well-
developed, although its uterine segment was non-pigmented.

On section, the right gonad was a normal testis and with the excep-
tion of one small spherical patch and two or three problematical granular
bodies, similar to those found in the opposite ovo-testis, pigment was
entirely absent. One ovum at " a stage of development corresponding
to that found in female frogs of between one and two summers " was
found near the surface close to the junction of the anterior and middle
thirds of the gonad, and the neighbouring portions of the testis were
less conspicuous than the remainder of the gland and contained the
most immature spermatozoa.

F. A. E. Crew

153

The left gonad was " for the most part made up of the polygonal
pigment cells which occur normally in the frog's ovary, with, however,
patches of normal testis at its poles. The anterior pole consists of testis
with a thin zone of the pigment cells invading one side, whilst the
posterior pole is somewhat over half testis. The remainder of the gland
is occupied by the pigment cells alone. Further a horizontal section
through the middle of the gland of this side, exposed five circular spaces
in the pigment mass, containing, surrounded by fibrous capsules,
spherical masses of lightly pigmented granular matter. Three out of the
five completely filled their respective spaces — the other two did not, one
of the latter also containing a large vacuole. The significance of these
bodies, which may be free or lie indiscriminately among the pigment
cells, and of the abnormally developed pigment mass, is, of course, on
the available data, impossible to determine. The former may plausibly
be looked upon as disintegrated ova (undergoing resorption)— the latter
as preceding the development of ova and the formation of a true ovo-
testis."

Fig. 15.

Fig. 16.
Killed January. Secondary

17. Kent. R. temporaria. Adult
sexual characters typically male.

Right gonad. An ovo-testis consisting of an inner lestis-portion
divided by a well-defined transverse constriction into anterior and

154 Sex- Reversal in Frogs and Toads

posterior halves, and bearing along the anterior part of its outer border
the strongly pigmented ovary-portion.

Left gonad. A testis of unusual shape.

Vasa efferentia of the male pattern present on both sides.

Seminal vesicles were present but were of unequal development, the
left one being the larger.

Mlillerian ducts. The right one was well-developed and much con-
voluted, but its uterine segment was small and non-pigmented. The left
one was rudimentary and straight, but possessed a well-defined though
small non-pigmented uterine segment.

On section the right gonad was seen to be composed of two discrete
portions. The testis-portion consisted everywhere of normal spermatic
tissue, but the ovarian was pathological. The left gonad was spermatic
in structure, but throughout the gland ova were found some within and
others between the seminal tubules. The presence of an ovum within
a tubule was associated with deformity of the spermatozoa, which were
obviously compressed. The presence of an ovum between seminal
tubules was associated with fibrous overgrowth in the intertubular con-
nective-tissue, and this produced deformity of the tubules in the
neighbourhood. There was ciliated epithelium upon the peritoneum.

18. Huxley. (2.) R.temporaria. Adult. Killed October. Secondary
sexual characters typically male.

Right gonad. An irregularly shaped testis with uneven outlines and
a scarred surface.

Left gonad. An ovo-testis. A three-lobed testis-portion twisted
upon itself, bore a prominent crest of pathological ovarian tissue upon
its outer border.

Vasa efferentia of the usual male pattern on both sides.

Seminal vesicles were present but were small and fusiform.

Miillerian ducts. Right one weakly developed and convoluted and
with a well-defined and pigmented uterine segment. The left one was
well-developed.

On section, the right gonad had the structure of ordinary spermatic
tissue save that six ova were found, two actually within and the rest
between seminal tubules. Some of these ova had a normal appearance,
others were degenerate. Accumulations of jjigment, ovarian in origin,
were found also between seminal tubules.

The left gonad consisted of an inner normal testis-portion and an
outer pathological ovary-portion. The two were quite discrete and the
ovary-portion contained but few ova and all of these were degenerate.

1

F. A. E. Crew 155

The greater part of the ovary-portion was composed of polygonal masses
of pigment and hyperplastic connective-tissue. The cavities of the lobu-
lated ovary were obliterated by this pigment but the divisions of the
lobules were still distinct on section. There were ciliated cells upon the
peritoneum.

Fig. 17. Fig. 18. Fig. 19.

Sub-group 2. The testis was normal in appearance. None.
Grouj) D. Those cases in which, on inspection, both gonads were testes.
Sub-group 1. The testes were abnormal in appearance.

19. Latter. R. temporaria. Adult. Killed September. Secondary
sexual characters typically male.

Right gonad. A faintly pigmented testis smaller than usual.

Left gonad. A faintly pigmented testis larger than usual and
irregular in shape.

Seminal vesicles present.

Miillerian ducts only slightly developed, not convoluted, and solid in
their anterior portions, but expanded in their posterior portions.

On section, ova were found in both testes, some within and others
between the seminal tubules. Besides these ova degenerate cell-masses
were found in similar situations.

20. MiTROPHANOW. R. esculenta. Young. Secondary sexual cha-
acters male.

Right gonad, A testis much smaller than usual.
Left gonad. A testis smaller than usual.
Seminal vesicles present and well-developed.

156 • Sex- Reversal in Frogs mid Toads

Miillerian ducts well-developed.

On section, the right gonad contained one true ovum and many
doubtful ones. A lobe at the anterior pole of the gonad was regarded as
a rudimentary Bidder's organ. The true ovum lay within a seminal
tubule. The left gonad was also a testis but contained doubtful ova only.

21. Pedaschenko. R. tcmporaria. Adult. Secondary sexual
characters typically male.

Both gonads missing.
Mtillerian ducts well-developed.

22. Marshall. (E.) It. temporaria. Adult. Secondary sexual
characters typically male.

Both gonads missing.
Seminal vesicles present.
Miillerian ducts well-developed.

23. Hooker. (A.) R. fusca. Adult. Secondary sexual characters
typically male. Killed April.

Left testis missing.

Seminal vesicles small.

Mtillerian ducts weakly developed with narrow uterine segments.

24. Marshall. (C.) R. temporaria. Adult. Secondary sexual
characters typically male.

Right testis replaced by fat.

Vasa efiferentia normal on the left ; absent on the right.

Seminal vesicles present.

Miillerian ducts well-developed.

Sub-group 2. The testes were normal on inspection,

25. Marshall. (A.) R. temporaria. Adult. Secondary sexual
characters typically male.

Seminal vesicles were small.
Miillerian ducts well-developed.

26. Gerhartz. R. esculenta. Adult. Secondary sexual characters
typically male. In addition to the usual male accessory sexual apparatus,
well-developed Mtillerian ducts were present.

27. Tarnanl R. esculenta. Adult. Secondary sexual characters
typically male.

Seminal vesicles were not present.

Miillerian ducts were only moderately developed.

28. TiCHOMiROW. R. esculenta. In an otherwise normal male
moderately well-developed Miillerian ducts were present.

F. A. E. Crew 157

29. Sutton. R. temporaria. In an otherwise normal male, slightly
developed Mullerian ducts were present.

30. KoRTHCHAGiN. R. esculentci. In an otherwise normal male,
slightly and imperfectly developed Miillerian ducts were still present.
The right one had separate anterior and posterior portions, it had no
anterior opening and was solid.

Fig. 25. Fig. 31.

31. Sumner. (1.) R. virescens. Adult. In an otherwise normal
male, Mullerian ducts slightly developed and with no anterior openings
were present.

32. Sumner. (2.) R. virescens. Adult. In an otherwise normal
male, very slightly developed Miillerian ducts with no posterior open-
ings were present.

B. BUFO.
Division II. Secondary Sexual Characters Typically Male.
Group A. On inspection one gonad an ovary, the other an ovo-testis.
Sub-group 2. The ovary abnormal in appearance.

33. Cerruti. (A.) Bufo vulgaris. Adult. Killed January. Secon-
dary sexual characters typically male. (When captured was actually in
copulation with a large female and appeared strongly excited.)

158

Sex-Reversal in Frogs and Toads

Right gonad. From behind forwards this consisted of (1) a pyriform
testis-portion ; (2) an ovary-portion ; and (3) a Bidder's organ.

Left gonad. Posteriorly, an anomalous ovary ; and in front of this, a
Bidder's organ.

Vasa efferentia. " Serial sections revealed the lack of a rete testis."

Miillerian ducts. Right one relatively long and thin, terminating
blindly ; left one relatively shorter and also terminating blindly. On
section the testis-portion of the right gonad showed the presence of
sperms of normal appearance mingled with other cellular elements
having variable form and size, " many of which are to be considered as
spermatids or spermatocytes in more or less advanced stages of de-
generation." The intercanalicular connective-tissue was abundant around
the tubules towards the centre of the testis-portion, and between the
tubules were numerous cells pigmented yellowish-brown.

The ovarian tissue of both gonads presented common characters.
The smaller oocytes presented a normal structure but the larger ovules
exhibited signs of degeneration and contained numerous granules of
yellow-brown pigment.

Since the Miillerian ducts ended blindly, the ova, even if produced,
could not be emitted.

Fig. 33.

Fig. 34.

Group B. On inspection both gonads ovo-testes.

34. Cerruti. (B.) Biifo vulgaris. Adult. Secondary sexual charac-
ters not described.

Right gonad. Posteriorly, a pyriform testis-portion ; anteriorly, an
ovary- portion. From the anterior part of the testis-portion "one sees

F. A. E. Crew 159

projecting from the testicular stroma some ovules of notable size and
browned with darkish jjigment." Bidder's organ missing.

Left gonad. Similar to the right.

Vasa efierentia. " Relation to the kidneys normal."

Miillerian ducts had been cut away before the examination was
made. On section the testis-portions had the normal structure. The
ovules seen projecting from the testicular stroma were in an advanced
stage of degeneration, being full of a granular substance — " in which
are present numerous granules of a yellowish-brown pigment and a rare
nucleus belonging to immigrant cells (perhaps leucocytes)." The ovary-
portions contained ova of all stages of development. The young and
medium were mostly normal though " in some the cytoplasm does not
appear homogeneous but in sections presents striae which stain deeply
with acid dyes. The nucleoli appear homogeneous or vacuolated and
the chromatic threads stain with great difficulty." " The larger ova
850-900 M. seem like normal cocytes of equal size, but sections of them
show easily that they are in a very advanced stage of degeneration, being
invaded by immigrant elements, probably leucocytes, and containing
numerous granules of yellow-brown pigment."

(In one figure is shown a dark mass due to an accumulation of
granules of normal yellow-brown pigment, which Cerruti suggests re-
presents an egg in a more advanced stage of degeneration.)

35. Spengel. Bufo cinereus {vulgaris). Adult. Secondary sexual
characters typically male.

Each gonad consisted from behind forwards of a testis-portion, an
ovary-portion of several ovarian chambers and containing ova as large
as those seen in a female of 2 or 3 years.

36. Knappe. Bufo vulgaris. Adult. Secondary sexual characters
typically male.

Each gonad consisted from behind forwards of a testis-portion, a
Bidder's organ, an ovary-portion, another Bidder's organ.

37. King. Bufo lentigijiosics. Adult. Killed March. Secondary
sexual characters male.

Right gonad. An ovo-testis with testis-portion posteriorly. No
Bidder's organ.

Left gonad. An ovo-testis, similar to that of the right side. The
ovary was placed where normally the Bidder's organ would have been.

Vasa eflferentia present.

Mullerian ducts still present but without terminal dilatations.

160 Sex- Reversal in Frogs and Toads

On section, the testis-portions had the structure of normal spermatic
tissue ; while all the larger ova within the ovary-portions showed de-
generative changes.

Group D2. Both gonads testes, normal in appearance.

38. Spengel. Bomhinator igneus. Adult. In an otherwise normal
male, the right Miillerian duct was as well-developed as the oviduct of
a female and the left, though smaller, was abnormally large.

C. PELOBATES.
Division II. Secondary Sexual Characters Typically Male.

Group (72. On inspection one gonad an ovo-testis, the other a testis of
normal appearance.

39. Spengel. Pelohates fuscus. Adult. Secondary sexual characters
typically male.

Right gonad. A normal testis.

Left gonad. An ovo-testis of which the posterior part consisted of a
2-lobed ovary, containing pigmented ova of the size of normal mature
ova.

Accessory sexual apparatus as that of a normal male.

Miillerian ducts extremely rudimentary.

D. HYLA.
Division II. Secondary Sexual Characters Typically Male.
Group D2. Both gonads testes, normal in appearance.

40. Sweet. Ilyla aurea. Adult. Secondary sexual characters
typically male.

Right and left gonads both testes normal in appearance.

In addition to the usual male accessory sexual apparatus, Miillerian
ducts were present, being well-developed, but having no posterior
openings.

On section, ova were found within both testes.

The species and the extent of the abnormality.

The species which contributed the abnormalities and the nature of
the abnormality are shown in the following Table :

F. A. E. Crew

161

TABLE I.

Division I.

The secondary sexual characters were
not typically male

Group A. One ovary ; one ovo-testis ...
Sub-group 1. The ovary was normal iu

appearance
Sub-group 2. The ovary was abnormal in

appearance
Group B. Two ovo-testes
Group C. One ovo-testis ; one testis
Sub-group 1. The testis was abnormal in

appearance
Sub-group 2. The testis was normal in

appearance
Group D. Two testes
Sub-group 1. The testes were abnormal in

appearance

Sub-group 2. The testes were normal in
appearance

a: Totals

= 1

Division II.

The secondary sexual characters were
typically male

Group A. 1.
o

As above

—

—

—

-

1 —
3 1

—

—

— 0

— 1

— 11

— 4

— 1

Group B.
Group C. 1.
„ 2.

5
4

1

1

—

1

—

Group D. 1.
„ 2.

4
2

1
4

1

2

— —

—

1

— 6
1 10 = 33

Totals

21

6

3

2

4 1
5

1
1

1
1

1

~~?2

1 =40

Group A ... 6

Group

C

5

,, B ... 13

D

Ifi

Of the 40 cases described, 21 were furnished by R. temporaria. This
does not mean, necessarily, that this species is more prone to abnor-
mality than any other, however, and the explanation of these figures
possibly lies in the fact that the Common Grass Frog is used for labora-
tory purposes more commonly than others. Only 40 cases have been
thought worthy of detailed description, but there is little doubt that
similar cases have been encountered in every Zoological Laboratory
though unrecorded. Nevertheless, such abnormality must be regarded
as very rare, though certainly occurring in all species of the Anura.

162

Sex- Reversal in Fror/s and Toads

The primary sex-glands of these abnornud individuals.

Spermatic and ovarian tissues were present, as noted in the tabu-
lated List of Cases, separately and in combination, in the following
grades.

Sp. Spermatic tissue. Ov. Ovarian tissue. ? Details not available. ^J As in the male.

5 As in the female. - Less than. Much less than. Very much less than.

Abs. Absent. Et Eight, Lt Left.

Right gonarl

Left gonad

Vasa

o6TTlin3(L

iviiiiier ^^^

Case

Sp.

Ov.

Sp.

Ov.

eflferentia

vesicles

ducts

A.

IA\

i

—

1

1

1

?

?

9

2

1

2

—

2

?

?

9

A '2

3

—

2

2

2

Et $ : Lt -

-3

- -c?

9

4

—

3

1

3

- <S

-c?

9

5

1

3

—

3

Et <J : Lt -

- S

?

9

B

6

1

2

1

2

c?

-<?

9

7

9

9

?

?

?

?

Et- 9:

Lt 9

\IB

8

1

1

1

1

o

?

9

9

1

3

2

?

3

-c?

9

10

1

3

1

3

S

3

9

11

2

3

2

4

3

- 3

9

12

2

4

1

3

Et,^: Lt-

3

-3

9

13

?

?

?

?

c?

3

9

14

1

4

1

3

c?

3

-9

G\

15

?

—

?

?

(?

3

Et- 9

Lt 9

16

2

—

1

3

c?

-3

Et- - 9

Lt - 9

17

1

3

2

—

o

3

Et- 9

Lt - - 9

18

2

—

1

3

^

-3

Et- 9

Lt 9

D\

19

2

—

2

—

c?

3

- - 9

20

2

—

2

—

(?

3

9

21

— Abs.

—

— Abs.

—

Abs.

3

9

22

— Abs.

—

— Abs.

—

Abs.

3

9

23

1

—

— Abs.

—

Et(J : Lt Abs.

- 3

- - 9

24

— Abs.

■-

1

—

Et Abs. : Lt $

3

9

D2

25

1

1

.

(?

-3

9

26

1

—

1

—

(?

3

2

27

1

—

1

—

o"

3

-9

28

1

—

1

—

c?

3

- 9

29

1

—

1

—

(?

3

--9

30

1

—

1

—

•3

3

--9

31

1?

—

1?

—

S

3

--9

32

1?

—

1?

—

.^

3

__ 9

B.

II A 2

33

2

2

2

2

9

3

--9

II B

34

1

2

1

2

(?

3

?

35

?

?

?

?

6

3

?

36

?

?

9

?

S

3

9

37

i

2

i

2

6

3

--9

II D2

38

1

—

1

—

6

3

Et 9

Lt - 9

c7

IIG2

39

?

—

?

?

c?

3

3

d7

II D 2

40

2

—

2

—

S

3

-9

F. A. E. Crew 163

Spermatic tissues.

1. Normal healthy spermatic tissue (represented as Sp. 1).

2. Healthy spermatic tissue with ova and pi(,nnent included within
its structure. (Sp. 2.)

Ovarian tissues.

1. Normal healthy ovarian tissue. (Ov. 1.)

2. Ovarian tissue in which a considerable proportion of the ova were
degenerate and in which there was overgrowth of the connective-tissue
and an increase of the pigment. (Ov. 2.)

3. Ovarian tissue in which the degenerative changes were more
pronounced, the ova being all degenerate and the hyperplasia of the
connective-tissue and the density of the pigment being more marked.
(Ov. 3.)

4. Ovarian tissue in the extreme stages of degeneration. No ova
remained and all that was left was a nodular mass of pigment and
fibrous tissue. (Ov. 4.)

Using these grades, it is possible to arrange the cases as on p. 162.

Reviewing only the cases of abnormality in Rana it is seen that
the various forms of ovarian and spermatic tissues furnish gonads as
follows :

Case and gonad Totals

Ov. 1
Ov. 2
Ov. 3
Ov. 4

1 Rt

3Rt, 2Lt ...
4 Rt, 5 Lt ...
Nil

1
2
2
0 =

= 5

Sp. 1
Sp. 2

23 Rt, 24 Lt,
16 Rt, 17 Lt,

25—32 both

18 Rt, 19 and 20 both

18

7 =

.2.5

Sp. 1, Ov. 1
Sp. 1, Ov. 2
Sp. 1, Ov. 3
Sp. 1, Ov. 4

11 Lt, 8 both

2Rt, Oboth

.5 and 9 Rt, 4 Lt, 10 both, 12 Lt, 14 Lt, 16 Lt, 17 Rt, 18 Lt
14 Rt

3
3

10
1 =

^17

Sp. 2, Ov. 1
Sp. 2, Ov. 2
Sp. 2, Ov. 3
Sp. 2, Ov. 4

Nil

3 Lt

11 Rt

11 Lt, 12 Rt

0
1

1
2-

- 4

In the case of 13 gonads sections could not be or were not made

13.

13

Total

64

One gonad

Ov. 1
Ov. 2
Ov. 2

Ov. 3

The other Case Totals

Sp. 1, Ov. 1 1 1
Sp. 1, Ov. 2 2 1
Sp. 2, Ov. 2 3 1
Sp. 1, Ov. 3 5 and 6 2

Sp. 2

Sp. 1, Ov. 3 16, 17, 18 3

Jouru. of Gen

. XI

11

1 64 Sex-Reversal in Frogs and Toads

When the abnormalities of the reproductive system are thus arranged
in a graduated series, they appear to be merely phases of one process at
the beginning of which the gonad has the appearance of a normal typi-
cal ovary, and at the end of which the same gonad is in every way a
typical testis. At the beginning, the gonad is an ovary of considerable
size, and is capable of functioning as such and of producing mature ova
(Case 1 Rt) ; it contains many ova in various stages of development,
and a considerable amount of undifferentiated germ-tissue, much of
which is situated upon the inner border of the gonad. Then upon this
inner border, one or several patches of spermatic tissue make their
appearance (Cases 3 and 4 Lt). These have developed from the hitherto
undifferentiated germinal tissue of this region. Very soon after this,
the ovarian tissue of both gonads begins to show signs of degenera-
tive changes, such as an increase in the relative amount of pigment, a
diminution in the size of the nuclei of the ova, and a shrinking of the
cytoplasm with vacuolaticm (Case 2). The spermatic tissue progres-
sively increases in amount, and the ovarian undergoes progressive
degeneration, becoming less and less in amount, and more and more solid
in consistence, while the pigmentation becomes more and more, intense.
Fewer and fewer ova are found within its shrinking bulk (Cases 4 Lt,
5 Rt), until at length the gonad assumes the form of a testis irregular in
shape, and with its surface scarred with deep grooves, which bears
upon its outer border a sub-peritoneal crest of jet-black nodules of pig-
ment cells and fibrous tissue (Case 14 Rt). This pigment, now a harmful
foreign body, is slowly removed by the blood-stream, and can be identi-
fied within the renal veins, and the testis, becoming healed, ultimately
assumes the form and appearance of the normal gonad of the typical
male (Cases 25-32).

It seems clear that at the beginning one gonad — usually the right
— is the seat of this transformation, for when both gonads are affected,
they exhibit two different phases of the process, that in one being more
advanced than that in the other, and it seems that the occurrence of
this phenomenon in one gonad favours its onset in the other. Since the
degeneration of the ovarian tissue is often equally pronounced in the
ovary of the opposite side, it is reasonable to assume that in every case
in the lists, both gonads have been, are, or would have been affected.

Ova are often found amid the spermatic tissues of the testis-portion
of an ovo-testis and of an otherwise normal testis. Normally ova are
extruded from the surface of an ovary. In an ovo-testis they are
extruded from the ovarian portion into the testis-portion (or the actively

F. A. E. Crew 165

growing spermatic tissues envelop ova and pigment along the line of
junction of ovary — and testis-portions), where they remain more or less
healthy for a considerable time, resisting destruction by virtue of their
maturity and being nourished and protected by the surrounding tissues,
while the other ovarian structures are destroyed and removed. The
presence of such ova is sufficient indication that the testis which
includes them has been an ovo-testis. These ova are doomed and
ultimately become resorbed, leaving the testis perfectly normal.

The question as to whether these ova are to be found actually
within the seminal tubules is of some importance, since if they are
found in this situation, it is not unreasonable to suggest that such cells
may be the results of the division of spermatogonia. There is, however,
no reason why ova produced by the ovary-portion of an ovo-testis
should not be found within a seminal tubule of the testis-portion of the
gonad, and if these cells are also found between the tubules, then force
is given to the contention that they are ova, and that they are the pro-
ducts of ovarian and not of spermatic tissue.

When such an ovum is found within a seminal tubule, the sperma-
tozoa also present therein are greatly compressed. The ovum thus may
have the normal structure, while the spermatozoa may be deformed and
degenerate. But this djegeneration is the result of local pressure, and is
not an indication that the ovarian tissue generally is healthy and
spermatic degenerate. When the ovum is situated between the seminal
tubules, those in the neighbourhood are contorted and misshapen.
This is the result of local hyperplasia of the intertubular connective
tissue, which follows the irritation set up by the presence of the foreign
body — the ovum and its pigment.

It is seen that Ov. 4 type ovarian tissue does not occur alone, but
only in association with spermatic tissue. Sp. 1 tissue is such as is
found in the very earliest stages, and in the closing stages of this pro-
cess : in the first it is associated with type Ov. I ovarian tissue. But
in the closing stages, ova which have been extruded from the ovarian
tissues, so that these are such as have been designated by the formula
Sp. 2, and all that remains of the other ovarian tissues, is the pigment
which is massed along the outer border of the gonad.

There is no gonad with the constitution Sp. 2, Ov. 1. This is because
in the case of type Ov. 1 ovarian tissue, there is, as yet, no appreciable
degeneration and no considerable degree of growth on the part of the
spermatic tissue. It is by the active enveloping growth of the spermatic
tissue that ovarian ova and pigment become included within them.

11—2

166 Sex- Reversal in Frogs and Toads

In those cases in which one gonad was a testis and the other an
ovo-testis, it is seen that the tissues of the testis are of the formula
Sp. 2. This definitely proves that the testis, which contains ovarian
pigment and ova, is a gonad in which the transformation from ovary to
testis is more advanced than in the gonad of the opposite side.

Although the process may begin in either gonad apparently, yet in
the 18 cases in which anything approaching a true estimate can be
made, the right gonad was first affected in 13 cases: the left one in 5
only.

It is the rule that the testis-portion of an ovo-testis is situated upon
the inner border of the gland, a position in every way convenient for its
association with vasa efferentia. The ovary-portion is always placed ex-
ternally to the testis-portion — and this fact has to be remembered when
the question as to whether a crest of pigment containing no ovum is
ovarian tissue or not is raised.

It would seem then, that if spermatic tissue becomes developed in a
gonad which previously has possessed the characters of an ovary, it
makes its appearance in a definite situation. This points to a localisa-
tion in the germinal tissue.

The process thus outlined can be depicted graphically by the use of
the formulae employed above. (Ov. 1, it' will be remembered, is ovarian
tissue which, on microscopical examination, has the normal structures.)
Ov. 1. (Case 1 Rt)
Sp. 1, Ov. 1. (Case 1 Lt, and Case 8 both)

Sp. 2, Ov. 1. (Case 9 Lt ?)
I I

Sp. 1, Ov. 2. (Case 2 Kt, Case 6 both) Sp. 2, Ov. 2. (Case 3 Lt)

Sp. 1, Ov. 3. (Case 4 Lt, ^Et, 10 both, Sp. 2, Ov. 3. (Case 11 Rt)
12 Lt, 14 Lt, 16 Lt, 18 Lt)

Sp. 1, Ov. 4. (Case 14 Rt) Sp. 2, Ov. 4. (Case 12 Rt, 11 Lt)

Sp. 2. (Case 16 Rt, 17 Lt, 8 Rt, 19 and 20 both)

Sp. 1. (Cases 25—32 both)
The degenerative changes in the ovarian tissue of both gonads j
quickly follow the development of spermatic tissue in one and the degree I
of this degeneration is more or less equal on the two sides. There is no

F. A. E. Crew 167

case in which the ovarian tissue was thoroughly healthy in one gonad
whilst degenerate in the other. If the presence of the spermatic tissue
in one gonad is responsible for the ovarian degeneration, then the agent
which produces this must be blood-borne, after the manner of a
hormone.

The Accessory Sexual Apparatus.

Vasa efferentia. It is seen that every gonad described as a testis,
whether it was normal or abnormal, was equipped with these efferent
ducts. The testis-portion of every ovo-testis also was linked up with
its kidney, but in these cases the vasa efferentia were commonly reduced
in number, when compared with the same structures of the typical
male. In those cases in which one gonad was an ovary, the findings of
different investigators have differed, but in some cases efferent ducts
were found cotinecting up gonad and kidney. Similarly, in the case of
an ovo-testis, vasa efferentia were found, not only in connection with the
testis but with the ovary-portion of the gonad too. Spermatic tissue
in a gonad, therefore, is invariably equipped with vasa efferentia, even
though its amount is relatively small, and ovarian tissue of an ovo-testis
and of an ovary is frequently found to be supplied with these efferent
ducts. The less the relative amount of spermatic and the greater the
amount of ovarian tissue in the gonads, the less likelihood there is that
vasa efferentia of the male pattern will be found.

It is very exceptional to find any suggestion of efferent ducts in a
typical female. During the examination of many hundreds of individuals,
no trace of such has been found. Therefore, it is reasonable to assume
that whilst the gonads have the constitution represented by the
formula Ov. 1, no efferent ducts will be present, and that they make
their appearance at the time of the early growth of the testis-portion
of the ovo-testis (Sp. 1, Ov. 1), and thereafter keep pace with the
development of the testis.

The seminal vesicles. These were absent in 5 cases (I, 2, 5, 7, and
8), and less developed than in the typical male in 10 others (3, 4, 6, 9,
11, 12, 16, 18, 23, and 25). Most of these occur towards the beginning
of the table, so that the nearer the individual approximates with
respect to the nature of the gonads, and to the nature of the secondary
sexual characters, to these characters in the typical female, the more
likely are the seminal vesicles to be either smaller than those of the
typical male or even absent altogether. Conversely, as has been pointed
out previously, in those cases in which the gonads and the secondary

1 68 Sex-Reversal in Frogs and Toads

sexual characters generally approximated those of the typical male, the
seminal vesicles were also, almost without exception, as those of the
typical male.

These facts suggest that there is some definite relation between the
nature of the gonads and the development of the accessory reproductory
apparatus. Cole's conclusion that "a good diagnostic feature as regards
sex is the seminal vesicle, which has only been observed in herma-
phrodites predominantly male" is justified, for it is seen that the
vesicles are commonly poorly developed or absent altogether in those
cases in the lists which were more female than male as regards their
secondary sexual characters and the nature of their gonads, in which a
considerable amount of ovarian tissue was present. In several cases the
vesicles were small or absent although there was no ovarian tissue pre-
sent in the gonads which were comjDOsed entirely of spermatic tissue.
In must of these cases unfortunately, the age of the individual and the
season of the year are not stated, and consequently a true estimate of
the importance of the condition of the vesicles cannot be made. Review-
ing the whole of the cases, however, it is seen that there is some con-
nection between the development of the seminal vesicles, and of the
vasa efferentia also, and the presence of spermatic tissue in the gonads.

In this connection, it is noteworthy that Nussbaum in 1905 found
that the size of the seminal vesicles was controlled by the internal
secretion of the testis.

It will be noticed that the degree of development of the seminal
vesicles and that of the vasa efferentia are connected, for in Cases 1 , 2,
3, 4, 5, 7, 12, both of these structures differ in degree of development
from those of the typical male, in that both were poorly developed.
Whatever affects the development of the one, therefore, also affects that
of the other.

MiXllerian ducts. The degree of the development of these typically
female ducts in the cases tabulated varies from almost the rudimentary
condition normally found in the typical male— fine straight strands of
tissue as described by M tiller. Burro, and Rathke — to that of the fully
developed oviduct of the typical adult female. In 18 cases these ducts
were found to be well-developed, moderately well-developed in 2,
slightly or weakly developed in 6, in 5 other cases the development of
the ducts of the two sides was unequal.

It will be seen that it is the rare exception for these ducts to be
anything but well-developed in those cases in the first half of the table,
that is, in those cases in which a considerable amount of ovarian tissue

I

F. A. E. Crew 169

still was present, and further that towards the end of the table the
condition of those ducts is more and more frequc^ntly described as
moderately, slightly or weakly developed. In the last 14 cases, th(j ducts
were well-developed in 6, moderately developed in 2, and slightly
developed in 6, and in all these only spermatic tissue was present in the
gonads.

So that in an otherwise normal male, Mullerian ducts in form and
sizie equal to the oviducts of the adult female are found in the presence
of actively functioning spermatic tissue. Harms and Meisenheimer
claim to have demonstrated that the internal secretion of the male
gonad can encourage the development of the accessory reproductive
apparatus of the opposite sex, and if this is so, then in such cases the
unusual development of the Mullerian ducts may have been a response
to the action of the internal secretion of the testes. Experiments con-
ducted to test the findings of Harms and Meisenheimer have failed to
confirm their results, and there is reason to believe that the occurrence
of Mullerian ducts in form and size equal to the oviducts of a mature
female, in an otherwise typical male, is an indication that functional
ovarian tissue of considerable amount has been present in the gonads of
the individual.

It is of interest to note that, in 1885, Sutton made the following
generalisation :

" When a male frog develops a Bidder's organ or ovary in conjunction
with a testis, the Mullerian duct or oviduct then assumes some consider-
able size," and from the results of the examination of more than 250
specimens of R. temporaria he concluded " that as a rule the amount of
development of these ducts is in direct proportion to the size of the
Bidder's organ."

It is gathered from the tables that the presence of spermatic tissue
in the gonads is associated with a fuller development of vasa efFerentia
and seminal vesicles, and that the Mullerian ducts attain their fullest
development only in the presence of ovarian tissue. It is suggested,
therefore, that the presence of well-developed oviducts is, in itself, an
indication that ovarian tissue either is or has been present in the gonads
of the individual possessing them. When once they have attained a
considerable size, as they do in association with the development of
ovarian tissue, the oviducts persist, even though the ovarian tissue is
removed. This can be demonstrated by extirpation of the ovaries, after
which the oviducts retain their form and size certainly for several
months. They do not attain a fuller development during the breeding-

170 Sex-Reversal in Frogs and Toads

season in cases thus operated upon, but they certainly do not atrophy.
In the cases in the lists, it is seen that only in those in which there
still remained some ovarian tissue capable of producing ova, though
they were abnormal and degenerate, did the oviducts exhibit a seasonal
increase in size and activity. In the majority their glandular portions
still retained the power of manufacturing their particular secretion, -for
in water the ducts became much swollen, and this is a further proof that
the ducts do not undergo any considerable degree of atrophy.

There is no evidence that the seasonal activity of the testis in these
cases produces a corresponding increase in the size and activity of the
Mtillerian ducts, as would be expected from the results of the experi-
ments of Harms and Meisenheimer. It is seen, as has been stated above,
that this increase is only shown in those cases in which ovarian as well
as spermatic tissue was present in the gonads.

It seems probable, therefore, that when the gonad has the constitu-
tion represented by the formula Ov. 1, oviducts become developed up
to the female standard ; that when the gonad's composition becomes
that represented by Sp. 1, Ov. 1, vasa efferentia and seminal vesicles
make their appearance ; and that coincidently with the fuller develop-
ment of the testis-portion of the ovo-testis, the male accessory repro-
ductory apparatus attains the form and size of that of the typical male,
while the oviducts, having reached a considerable size already, retain
their form throughout the process which is converting a young female
into a somatic male.

Granting that the hormones of the gonad are responsible for the
stimulus which calls forth the development of the accessory reproductive
apparatus, it is seen that the exhibition of the specific male hormones
does not cause the development of any structure which is not repre-
sented embryologically in the normal female, and that hormones act
merely by inhibition or stimulation of normal embryonic rudiments.

The Miillerian ducts were of unequal development in a few cases,
and in these the better developed duct was on the side of the body on
which more ovarian tissue was still present (cf. Aves).

Jase

Right gonad

M.D.

Left gonad

M.E

Sp.

Ov.

Sp.

Ov.

7

?

?

- ?

9

?

9

15

?

—

- ?

?

9

?

16

2

—

--?

1

3

-?

17

1

3

-?

2

—

--$

18

2

—

-?

1

3

9

F. A. hi Crew 171

There is, therefore, some evidence that the degree of development of
the Mullerian duct is controlled to some extent by the activity of the ovary
of the same side. It would seem that the Mullerian ducts are developed
under the direction of the ovaries, but that the degree of their develop-
ment in these cases is determined by the time during general development
at which the spermatic tissues first become expressed. If this is expressed
relatively early, then these ducts will not have attained their full de-
velopment, and their conditions will be as in Cases 14, 27, and 28. If
spermatic tissue becomes expressed even earlier still, then the ducts
will be as in Cases 20, 23, 29, 30, 31, and 32. If, on the other hand, this
expression is relatively later, then the ducts will have already attained
their full development and will retain their characters.

In the cases in which these ducts were of unequal development, it
would seem that the process of transformation of ovary into testis began,
as is usual, in one gonad, and later extended to the other, and that the
interval between the beginning of the process in the two gonads was
prolonged, so that the duct of the side on which spermatic tissue had
yet to become expressed in the gonad would be permitted to develop a
little further, as the ovarian tissue of this side would be affected to less
extent than that of the other.

A point of interest arises in Cases 30, 31 and 32. In Case 32 these
ducts had no posterior openings, while in the others they were solid in
their anterior portions.

The Secondary Sexual Characters:

Of the 30 frogs in the tables of which sufficient details were given
as to their secondary sexual characters 25 (83'3 "f^) were definitely and
typically male. Four others were definitely but imperfectly male
(13'3 °/J, and in the remaining case the secondary sexual characters
were female (3*3 "/J.

In every case functioning spermatic tissue was present, and in those
cases in which ovarian tissue was also present, and in which sufficient
description of its histological structure was given, it has been shown
that this tissue was pathological and undergoing removal.

The presence of imperfectly developed male secondary sexual cha-
racters is not associated with the development of female characters as is
seen in Cases 2 and 4, and the secondary sexual characters can remain
fully and typically male in the entire absence of the gonads, as is illus-
trated by Cases 21 and 22.

172 Sex-Reversal in Frogs and Toads

These characters do not become at all obvious while the individual
is still young, and the female characters are, for the most part, negative
in nature. Assuming that there is some intimate relation between the
primary and secondary sexual characters, a review of the cases suggests
that the super-imposition of male upon female secondary sexual cha-
racters is a simple matter in the Anura. The presence of patches of
ciliated epithelium upon the peritoneum round about the suspensory
ligament of the gonad does not interfere with the assumption of the
male characters which are developed in quite different situations. The
wartiness of the skin is a seasonal phenomenon and only occurs in the
presence of functioning ovarian tissue.

It has been shown that the development of spermatic tissue upon
the inner border of a gonad, which previously had every appearance of
an ovary, occurs at an early stage in the life-history of the individual,
and that coincidently with its appearance the ovarian tissue begins to
degenerate. If, then, the female secondary sexual characters are not
developed to any extent, until the normal female is well-matured, the
development of the spermatic tissue in these abnormal individuals and
the consequent destruction of the ovarian tissue will prevent the as-
sumption of the female characters.

The degree of development of these characters bears no relation to
the amount of germinal tissue present, but this tissue must be healthy
and functioning. A small amount of healthy spermatic tissue is asso-
ciated with well-defined male secondary sexual characters, although in
the individual there is a greater amount of ovarian tissue which is
pathological. It is seen that the exhibition of the male sex-hormones is
not attended by the development of any structure which is not repre-
sented cmbryologically in the normal female, and that as in the case of
the accessory reproductive apparatus, the hormones act by inhibition
and stimulation of normal embryonic rudiments.

The cases of abnormality in Bufo, Bomhinator, Pelohates and Hyla
are too few to permit of a similar treatment to that adopted in the case
of Rami. They are, however, exactly similar in their nature and there
can be no doubt that the conditions found are, with minor differences,
identical with those described in Rana.

The relative position of ovarian and spermatic tissues in an ovo-testis
is somewhat different, in that the ovarian tissue is placed anteriorly and
the spermatic posteriorly. Further, it would seem that the comparatively
smaller amount of spermatic tissue is associated with the assumption of
male, secondary, sexual characters in the case of Ba/o than in Rana.

F. A. E. Crew 173

The gonad ofBafo is subject to considerable variation. Cerruti figures
a Bu/o vulgaris in which he observed, in front of each testis, a Bidder's
organ, another piece of testis, another Bidder's organ and histly the fat-
body; and in another case, quoted in the lists, in which in place of
Bidder's organ there was found in front of each testis an ovary. Knappe
describes a case, quoted also in the lists, in which he observed in front
of each testis a Bidder's organ, an ovary, another Bidder's organ and
lastly the fat-body. King has dealt with such abnormalities at length,
and she is of the opinion that the component cells of an embryonic
testis or ovary, in response to the stimulus of altered blood supply, can
assume the characteristics of a Bidder's organ. It certainly does appear
that cells which normally should develop into typical testis or ovary or
Bidder's organ can, on occasion, develop into another of these tissues.

In Cases 34 and 37 the position of the ovarian portion, coupled with
the fact that the Bidder's organ was missing, strongly suggests that
these had been derived, in part at least, from those cells which normally
would have given rise to Bidder's organ. Further, as King points out,
as the testis-portion in these cases was shorter than the normal testis of
the typical male, it may be assumed that the primordial germ-cells in the
anterior part of the germinal ridge, which normally would have developed
into spermatic tissue, had also taken part in the formation of the ovary.
She suggests that the causal agent, whatever it may have been, must
have acted at a very early period in the life-history of the individual,
since in the normal toad the cells which develop into Bidder's organ
become differentiated when the tadpole is about two weeks old.

In Cases 33, 35 and 36, an ovary was found between the Bidder's
organ anteriorly and the testis posteriorly. This can be interpreted as
the condition in which the most anteriorly situated cells of the germinal
ridge have developed, as is usual, into a Bidder's organ, those situated
posteriorly into a testis, while the cells of the more anterior part of the
middle region of the ridge have developed, for some reason or other,
into an ovary. In Case 39 one gonad was an ovo-testis with the ovary-
portion posteriorly situated. There is in this a strong suggestion that
either the forerunners of testis, ovary, and Bidder's organ are normally
present in the apparently undiiferentiated primordial germ-tissue, or
perhaps, that the foundation primordial germ-cells, in response to the
appropriate stimulus, can become any one or all of the three types of
adult germinal tissue.

The cases illustrate the process by which an individual with every
female character ultimately comes to possess the typical male organisa-

174 Sex-Reversal in Frogs and Toads

tion. But these transformed individuals not only assume the male
characters but also can behave and function as males.

The opportunity neglected by Cerruti (Case 33), of applying the
test of breeding in order to examine the exact nature of an individual
with an abnormal reproductive system, presented itself to me.

It can be assumed that in the frog sex-determination is effected by
means of a sex-chromosome mechanism. The sex-chromosomes have not
yet been demonstrated, but Huxley has given reasons for the belief that
the chromosome-constitution of the frog is of the A'^ Y, XX type.

It so happened that the specimen referred to in the lists as No. 14,
when received, was actually in copulation. This pair, with many others,
was retained for breeding purposes. The individuals were marked and
full details collected during the course of the breeding. Later, when the
tadpoles had hatched, the parents were killed and dissected. Then it
was that the abnormalities were first noticed and in consequence every
attention was paid to the examination of the female with which this
abnormal individual had mated and to the rearing of the offspring. The
female proved to be normal in every way anatomically ; she died before
she could be bred from again.

The eggs and tadpoles resulting from this union were raised accord-
ing to the directions given by H. D. King and every effort made to rear
as many individuals as possible. The fertilised eggs from seven other
couples were kept under exactly similar conditions to furnish controls.

The eggs nearer the centre of the egg-mass gave rise to few tadpoles
and the eggs that were fertilised did not all hatch, for numbers of such
were arrested in their development by death. The question of selective
mortality naturally enters into a consideration of the sex-ratio of the
individuals which did survive therefore. But there is no evidence to
show that the sex-ratio would have been different if every Q^g had
hatched and every tadpole had survived.

As the tadpoles passed the metamorphosis, they were removed to
outdoor pens under conditions as natural as possible, the different lots
being kept separate.

During twelve months, the frogs were examined in batches at in-
tervals of two months. Every one that died was examined and included
in the records. The controls showed no great variation in the relative
numbers of the sexes and towards the end of the experiment only two
lots were maintained.

The gonads were sectioned and compared to the standards given by
Witschi. No case of the indirect method of development of a testis was

F. A. E. Crew 175

encountered, the gonads were either ovary or testis or else the tissues
were too undifferentiated to allow of a decision being made. In all 2850
tadpoles and frogs were examined.

But of the frogs which resulted from the union of the male with the
abnormal reproductive system and a normal female, every one of the
774 examined and found with gonads sufficiently developed was a female
normal in every respect. Those of the control lots were typically male or
female and no case of abnormality was met with among them. The re-
lative numbers of males and females varied in the different lots but
only to a slight degree, the average being 4G 7o males, 54 7o females.
The average of all the lots was 28 °/^ males and 77 "/^ females, which
was distinctly different from the figures of the parental generation
(80 °/^ males and 20 °/^ females), constituting, in fact, a very complete
swing-back.

The results of this breeding experiment go far to prove that the
male parent was a " somatic " male, a masculinized female. Though
possessing the male organisation yet in chromosome-constitution it was
a female {XX), and when mated to a normal female {XX) produced a
generation consisting entirely of females. Its chromosome-constitution
had become over-ridden by external factors.

There is considerable evidence, as Huxley recently has shown, that
chromosome-constitution may thus become over-ridden. He cites the
work of Goklschmidt who lately has bred from ZW males and from
■^-^ females of Lymantria, nnd that of Shu 11, Strasburger, and Doncaster
on Lychnis dioica.

Lillie's work on the free-martin can be interpreted as further proof
that in the female — presumably XX in chromosome-constitution —
embryos in cattle, co-twins of males, the whole organisation can be so
altered that even the gonad itself takes on the characters of testis and
the accessory sexual apparatus becomes more male than female.

Doncaster in one of his latest papers suggested that the sterile
tortoiseshell tom-cat possibly is a female chromosomally which becomes
transformed into an individual with an almost complete male
organisation.

It is generally accepted that the agent responsible for the actual
transformation of the gonads is the internal secretion of the testis — the
male hormone. It may be assumed that the external factors which are
primarily responsible for the reversal of sex act through the medium of
the ductless glands.

Lillie found that twins of cattle are derived exclusively from

176 Sex- Reversal in Frogs and Toads

separate zygotes, so far as the evidence from sixty-one cases goes. The
embryonic membrane of such twins, however, fuse in an early stage
(embryos of about 30 mm.) and the blood-vessels of the two individuals
anastomose. If one is male and the other female the reproductive
system of the latter fails to develop its usual characters, and characters
of the male appear instead to a variable extent which appears to depend
upon variations in time and degree of the vascular anastomosis. Such
individuals have long been known as free-martins. The gonad is testis-
like in form and structure owing to complete suppression of the ovarian
cortex and hypertrophy of the homologue of the seminiferous tubules.
The Miillerian ducts usually degenerate, and the Wolffian ducts may
develop into quite typical vasa efferentia ; guhernacula arise as in the
male ; but, save in very exceptional cases, the external organs of repro-
duction and the mammary gland conform to the female type. In rare
cases (about one in eight cases of two-sexed twins) the vascular anasto-
mosis fails to develop, and in such cases the female is normal. No
abnormalities of the reproductive system of the male arise in two-sexed
twins.

Sex-determination in the zygotic sense is thus seen not to be the
exclusive determiner of sex-differentiation in mammals, even in respect
to the most fundamental sex-characteristics. The possibility of complete
sexual inversion, by means of hormones of the opposite sex, and of
control of sex-determination in this sense, is thus postulated.

Chapin who made a microscopic study of the reproductive system of
foetal free-martins demonstrated that there is a fusion of the embryonic
membranes and a subsequent anastomosis of blood vessels of the cattle
twins. If one twin be male and one be female, the latter is commonly
sterile. This is the result of the introduction of the interstitial secretion
of the male into the blood of the female. It is manifested by the de-
velopment toward the male condition, of those organs in the free-martin
which are present in the indifferent stage {rete, first set of sex cords,
primary albuginea), and the absence of those organs which develop in
the normal female at sex differentiation or later (cords of Pfluger, defini-
tive albuginea, union of Miillerian ducts to form uterus).

There is much variation in the reproductive organs of the free-
martins. This is due to two variable factors : (1) the time at which the
interstitial secretion of the male embryo may first enter the circulation
of the female embryo, and (2) the amount of secretion which may be
introduced.

Willier who has more recently investigated the structure and homo-

F. A. E. Crew 177

logies of the gonads of the frco-martin shows that in th(; free-martin (a
sterile female co-twin to a normal male in cattle) an indifferent gonad
with a primary female determination, imder the influence of sex
hormones from the male twin, may develop variable degrees of trans-
formation in the male direction. So far as the structural evidence from
sixteen cases (seven foetal and nine postnatal) goes, the reproductive
glands of free-martins may be placed into three distinct groups, which
may be characterized as low, medium and high degrees of transformation
in the male direction, and constitute, therefore, a chain of connected
links between an embryonic ovary and a testis. Every organ of these
modified free-martin gonads is affected. The sexual chords exhibit a
series of gradations between medullaiy cords and seminiferous tubules
(complete except that male sex cells are lacking). The rete ovarii trans-
forms into a rete testis chiefly by developing connections {tiihnli recti)
between the rete tubules and the epididymal tubules. In the least trans-
formed gonads the epididymis is absent ; in gonads exhibiting a moderate
degree of transformation the head of the epididymis alone is present, and
in the most completely transformed gonads a complete epididymis is
developed. The distribution of the blood-vessels ranges from a typical
ovarian to a typical male arrangement. It is thus seen that the most
fundamental sex organ may be rather completely inverted by means of
hormones of the opposite sex. Sex differentiation, then, is not ex-
clusively determined by the union of the gametes.

But, as is illustrated in the cases of abnormality already describerl,
the spermatic tissues may not become expressed until the ovarian have
attained functional maturity and it has been shown that the appearance
of a small amount of spermatic tissue upon the inner border of a gonad
which previousl}' has had every character of a normal ovary, is associated
with the immediate degeneration of the ovarian tissues and a gradual
but inevitable assumption on the part of the individual of the male
organisation.

In these circumstances, if the gonal tissues exert their influence by
means of internal secretions, then that of the testis must be more
powerful by far than that of the ovary for the former wipes out the
ovarian tissues and directs the imposition of male characters.

There is, in the frog, an antagonism of sex hormones and the ovarian
is powerless in the presence of the spermatic. Quantity, apparently, is
not a factor, for a very small nodule of testicular substance can still
provide sufficient of the male hormone to wipe out the g:-eat mass of
ovarian tissue and its abundant internal secretion.

178 Sex- Reversal in Frogs and Toads

A difficulty naturally arises. If the above be true, then there is
reason to question the accepted statement that Bidder's organ is a rudi-
mentary ovary. This organ persists in the male but disappears in the
female, which suggests that it is spermatic rather than ovarian in nature,
but so deficient in interstitial tissue that the hormone of the ovary,
becoming exhibited in the absence of the hormone of the spermatic
tissues, produces its ultimate destruction. In this connection, it will be
remembered that Cerruti found spermatozoa actually within a Bidder's
organ and that Hoffmann described rudimentary seminal tubules
therein, whilst the experimental work of Harms supports the idea that
Bidder's organ is male in nature.

Harms maintains that Bidder's organ is an organ of internal secre-
tion and has conducted experiments which show that the development
of the typical male finger-pads is controlled in some part by this organ,
for, while extirpation of the testis or of Bidder's organ has no influence
upon the development of the pads, yet if, after extirpation of both testes
and Bidder's organs, a Bidder's organ is grafted in the dorsal lymph-
sacs, the typical seasonal development of the pigmented pads follows in
due course.

Goldschmidt has shown that Bidder's organ undergoes changes which
are parallel to the sexual cycle and that in the case of the male there is
increased activity and regeneration of its constituent cells, co-incident
with the formation of the spermatozoa.

Cytologically it has not been proved that the cells which constitute
Bidder's organ are ovarian, and there undoubtedly are reasons for
questioning the generally accepted opinion that this organ is a rudi-
mentary ovary.

Summary.

1. It is shown that the recorded cases of abnormality of the repro-
ductive system illustrate the process by which an individual, at one
time possessing solely the complete sex-equipment of the female, comes
to exhibit the organisation of the male. The sole female character which
remains in what otherwise appears to be a typical male may be the full
development of the Miillerian ducts, or the presence of ova amid the
spermatic tissues.

2. Such a transformed individual not only has the male organisation
but also behaves and functions as a male.

3. It so happened that a transformed individual, functioning as a
male, fertilised the eggs of a normal female. This male was one of an

1

F. A. E. Crew 179

original population including 80 7^ males. All (774) of his offspring
which attained a sufficient development to permit identification of the
sex were females. Including these 774 females, the tadpoles produced
from the original population were 23 7o males and 77 7„ females.

4. This " swing-back " is taken as evidence which supports the con-
ception that the frog has a chromosome constitution of the XV,
XX type, and that these cases of abnormality are females {XX), the
chromosome constitution of which has become over-ridden by external
factors, and which are " somatic " males or masculinised females. These,
being XX in chromosome-constitution, mating with normal females, also
XX in constitution, will produce a generation consisting entirely of
females. The presence of such "somatic " males in a population will
thoroughly disturb the sex-ratio of the succeeding generation, producing
a " swing-back " which is a certain indication of their presence and re-
productive activities.

o. The mechanism by which the individual is thus transformed is
one which acts through the internal secretions of the gonads. It is
shown that in the frog the phenomenon of sex-reversal is very similar
to the same process in cattle save that in the frog there is an antagonism
of the sex-hormones and that the ovarian is powerless in the presence
of the spermatic.

6. It is shown that there is a most intimate relation between the
primary sex-glands and the accessory sexual apparatus and the
secondary sexual characters.

7. Reasons are given for questioning the generally accepted opinion
that Bidder's organ is a rudimentary ovary.

Since the present writer first offered a, suggestion as to the possible significance
of the abnormalities in frogs, an important contribution to the literature dealing
with the subject of reversal of the sex-organisation has been made by Champy, C.
(C. R. Acad. Set. Paris, Seance du 9 Mai, 1921.)

Champy had previously observed that the annual incidence of spermatogenesis
in Tritons could be prevented by starvation, and that in males, killed in Spring
after such treatment, the testicles had become replaced by fat which contained
groups of undifferentiated gonocytes.

More recently he found that two such males when fed intensively began to
assume the external characters of the female. One was killed for examination and
the gonad was found to have been replaced by fat as in the previous cases. The
other specimen was kept alive for further observation. Its history was known ; it
had been a perfect and complete male, and had fertilised the eggs of a normal
female. Gradually it came to be exactly like a female in appearance. It was ulti-
mately killed. On examination each testis was found to have been replaced by a wide
Journ. of Gen. xi 12

180 Sex-Reversal in Frogs and/ Toads

long strip of fat within which was an elongated organ with the appearance of an
immature ovary and enclosing a typical oviduct. On histological examination the
organ proved to be an ovary consisting mainly of young oocytes.

" Knowing the previous history of the animal there can be no doubt that we
have here a case of complete sex-reversal."

Champy has not yet bred from such a feminised male.

BIBLIOGRAPHY.

1. BouLANG^, H. "Un cas d'hermaphroditisme vrai bilateral chez Rana fusca

(Thomas)." Feuille des Jeunes Naturalistes, Tom. xliv. p. 79. Paris, 1914.

2. Bourne, A. G. " On Certain Abnormalities in the Common Frog." Q. J. M. S.

Vol. XXIV. p. 83, 1884.

3. Cerruti, a. " Contribvizioni per lo studio dell organo di Bidder nei Bufonidi.

I. Di una speciale penetrazione di ovuli in ovuli nel Bufo vulg. Laur." Attr

R. Sc.Jis. mat. Vol. xii. ser. 2, No. 1, pp. 1—8. Napoli, 1903.
4. . "II. Presenzadispermii neir organo." Btdl. Sci.Nat.\d\..x\i\.\)\t.\bO—b.

Napoli, 1903.
5. . " Sopra due casi di anomalia dell' apparato reproduttore nel Biifo vulg.

Laur," Anat. Am. Bd. xxx. p. 453, 1907.

6. Chapin, C. L. "A Microscopic Study of the reproductive system of foetal

free-martins." Journ. Exfer. Zool. Vol. xxiii. p. 453, 1917.

7. Cole, F. "A case of Hermaphroditism in Rami temporaria." Anat. Anz. Bd. xi.

p. 112, 1895.

8. Crew, F. A. E. " A Description of certain Abnormalities of the Reproductive

System found in Frogs and a Suggestion as to their Possible Significance."
Proc. Roy. Phys. Soc. Edin. Vol. xx. Part 5, p. 236, 1921.

9. Cutler, D. W. and Doncaster, L. "The Sterility of the Tortoiseshell Tom-

cat." Journal of Genetics., Vol. v. p. 65, 1915.

10. Doncaster, L. " The Tortoiseshell Tom-cat, — a Suggestion." Journal of

Genetics, Vol. ix. p. 335, 1920.

11. Gerhartz, H. " Rudimentarer Hermaphroditismus bei /iawa escM^ento." 2irch.

f mikr. Anat. Bd. Lxv. p. 669, 1905.

12. GoLDSCHMiDT, R. Mechanismus u. Physiologic der Geschlechtsbestimmimg.

Berlin, 1920.

13. Harms, W. Experimentelle Untersuchungen Uber die innere Sekretion der Keini-

drilsen. Jena, 1914.

14. Hooker, D. "Der Hermaphroditismus bei Froschen." Arch, f mikr. Anat.

Bd. Lxxix. p. 181, 1912.

15. Huxley, J. S. " Note on an Alternating Preponderance of Males and Females

in Fish and its Possible Significance." Jmwnal of Genetics., Vol. x. No. 4,
1920.

16. Kent, A. F. S. "A Case of Abnormal Development of the Reproductive Organs

in the Frog." Journ. Anat. and Phys. Vol. xix. p. 129, 1885.

17. King, H. D. "Some Abnormalities in the Genital Organs oi Bufo lentiginosus

and their Probable Significance." Amer. Journ. Anat. Vol. x. p. 159, 1910.

18. Knappe, E. "Das Biddersche Organ." Morph. Jahrb. Bd. xi. pp. 489 — 548,

1886.

F. A. E. Crew 181

19. KORTSCHAGIN, A. N. "Fall einer Anoinalie der Goschlcchtsorgane Leitn

Frosche." Tagehl d. zool. Aht. d. Gesellsch. d. Natnrfr. Mod-au, No. 1, 1890.

20. Latter, 0. II. "Abnormal Reproductive Organs in Rana temporariaJ' Journ.

Anal, and Phys. Vol. xxiv. pp. 369—372, 1890.

21. LiLLiE, F. R. "The Theory of the Free-martin." Science, N.S. Vol. xmii.

p. 611, 1916.

22. . "The Free-martin; a Study of the Action of Sex-hormone.s in tho

Foetal Life of Cattle." Journ. Exper. ZooJ. Vol. xxril. p. 371, 1917.

23. Marshall, A. M. "On Certain Abnormal Conditions of the Reproductive

Organs in the Frog." Joni-n. Anat. and Phys. Vol. xviii. pp. 121—144, 1884.

24. Meisenheimer, J. " Ueber die Wirkungen von Hoden- und Ovarialsubstanz auf

die sekundaren Gesehlechtsmerkmale des Frosches." Zool. Anz. Bd. xxxviii.
1911.

25. MiTROPHANOW, p. M. "Ein Fall des Hermaphroditisnuis beini Fro.sche."

Arb. a. d. zool. Lab. d. Univ. Warschau, Bd. ii. Aviszug im Zool. Am.
Bd. XVII. p. 98, 1894.

26. Ognew, S, J. "Ein Fall von Hermaphroditismus bei Ra7ia tempo7-aria.'' Anat.

Anz. Bd. XXIX. pp. 194—203, 1906.

27. Pedaschenko, D. "Un cas d'anomalie des organes genitaux chez la Rana

temporaria." Revue Sc. Nat. Soc. Natural. Ann. i. No. 6, pp. 207 — 267.
St Petersbourg, 1890.

28. PuNNETT, R. C. " Note on a Hermaphrodite Frog." AnnaU and Mag. Nat.

Hist. Vol. VI. pp. 179—180, 1900.

29. RiDEWOOD, W. " On an Abnormal Genital System in a Male of the Common

Frog." Anat. Anz. Bd. in. pp. 333—336, 1888.

30. Smith, W. R. "A Case of Hermaphroditism in a Common Frog." Journ. Anat.

and Phys. Vol. xxiv. p. 218, 1890.

31. Spengel, J. W. "Das Urogenital System der Amphibien." Arbeit zool.-zoot.

Institut z. Wiirzburg, Bd. in. pp. 1 — 114, 1876.

32. Sumner, F. B. "Hermaphroditism in Ra}ia virescens." Anat. Anz. Bd. ix.

p. 694, 1894.

33. Sutton, J. B. "Diseases of the Reproductive Organs in Frogs, Birds and

Mammals." Journ. Anat. and Phys. Vol. xix. p. 129, 1885.

34. Sweet, Georgina. " Variation in the Anatomy of Hyla aurea.'' Proc. Roy.

Soc. Victoria, Vol. xxi. Part 1, 1908.

35. Tarnani. " Anomalie im Bau der Geschlechtsorgane des Frosches." Notiz d.

Novo-Alexandriner Landioirtsch. Institut, p. 55, 1898.

36. TiCHOMiROW, A. "Zur Frage iiber den Hermaphroditismus bei den Vogeln."

Nachrichten d. Gesellsch. d. Naturfr. Bd. Lii. 1887.

37. WiLLiER, B. H. "Structures and Homologies of the Free- martin's Gonads.''

Joimi. Exp. Zool. Vol. xxxiii. No. 1, 1921,

38. Youngman, W. "A Specimen of Rana temporaria with Abnormal Rei)roductive

Organs." Anat. Anz. Bd. xxxv. p. 301, 1910.

12—2

A STUDY OF THE SEGREGATION OF A QUANTI-
TATIVE CHARACTER IN A CROSS BETWEEN
A PURE LINE OF BEANS AND A MUTANT
FROM IT.

By I. LEITCH.

(With Four Text-figures.)

The service which Mendel and his rediscoverers rendered to the study
of heredity in demonstrating that certain qualitative characters behave,
relatively to each other, as separate entities, has been paralleled by
Professor Johannsen, Jennings and others for quantitative characters.
Until their researches were made, size, for instance, was regarded in the
light of Darwin's and Galton's work as something capable of endless
variation — the net result, in any particular case, of a summation of causes,
natural selection, fitness, and other equally indefinite concepts. What
Professor Johannsen did was to delimit certain quantitative characters;
to show that, although to some extent influenced by environment, still
they are definite characters, capable of accurate description and analysis.
What makes the problem of quantitative differences appear more difficult
than that of differences in colour and other such qualities is that they
cannot be described and classified on inspection, but the two problems
approach in cases where, for instance, cumulative colour factors operate,
and quantity of colour enters the field. There appears indeed to be no
reason to regard the problems as essentially different.

Professor Johannsen established for quantitative characters the prin-
ciple of working with homozygous, " pure," material as the basis of all
analysis. He demonstrated for his pure lines of beans and barley that
such material breeds true ; that, in a homozygote, such characters as size
and weight are as constant as gi'eenness or smoothness ; and he made
familiar the exact description and comparison of such characters by means
of their distribution curves. For instance, the sizes of the beans in two
plants, or pure lines, or aggregates of beans, will be described fully and
accurately in any given case by the distribution curves of their lengths
and breadths, and the comparison of different sizes will be a question
merely of the comparison of distribution curves. These curves have not

184 Segreffation of a Quantitative Character in Beans

in the first place any mathematical or statistical end in view, but are
only a summary expression of exact measurements, and become the
equivalents of such descriptive terms as red and white or smooth and
rough.

The constancy of quantitative characters having been shown, it
remained to prove whether or not they behave in a cross like qualitative
characters, showing the usual forms of Mendelian segregation. Crosses
made between beans of different lengths and breadths showed segregation
similar to that where several cumulative characters are in question, as
Professor Johannsen points out in his Elemente der exakten Erhlich-
keitslehre, 2nd Ed. p. 558. And further he communicates interesting
results in connection with the origin of the mutation dealt with in this
paper and of another mutation. He finds that the latter mutation,
occurring as a heterozygote, gives the original form and a new form as
segregates in the simplest Mendelian ratio, 1:2:1. [Elemente, 2nd Ed.
p. 653.]

The data to be dealt with in this paper are the results of a cross
between the pure line E and a long mutant from it, made in hope of
throwing some light on the nature of the mutation. I am indebted to
Professor Johannsen for the following account of the origin of the material
in question.

" The beans used were a variety of Fhaseolus vulgaris nana called
brown Princess beans, cultivated for commercial purposes in certain
districts in Fyn. In 1901 a number of pure lines were isolated from
bought material. Only biotypes were chosen of which the seeds did not
touch each other in the pod, but developed quite freely. And all types
where there was a tendency to abortion, that is, where the pods were
not filled with seeds, were excluded. All the biotypes were therefore
from the beginning, healthy, fertile and with well-formed seeds, suited
to be measured singly with the apparatus described in the Elemente der
exakten Erhlichkeitslehre, 2nd Ed. p. ] 3.

"As to fertilisation, many years' work with the pure lines in question
has shown that, when the beans are cultivated under a voliere of wire
netting with a mesh of about 2 cms. there is little chance of crossing.
And all the experiments made on these pure lines have clearly shown
their genotypical constancy ; all selection has been without result. Fur-
ther, in the crosses that have been made — and made only with great
difficulty — it has appeared at once that the conditions, as regards in-
heritance and variation, are quite different from the behaviour of pure
lines, as is pointed out in the Report of the Third International Con-

I. Leitch 185

ferencc in Genetics, (1906), p. 98, and as is demonstrated in the Elemente,
2nd Ed. p. 558.

" The pure line E, with which the present paper is alone concerned,
is distinguished from all others in my material by having pure white
petals and a yellow-white colour in the rest of the flower, without a trace
of violet. The other lines of Princess beans have a violet tinge, often
weak but unmistakable. And in the crosses made between E and other
lines, the F^ generation has always had this violet colour in the flower,
and the F.2 generation has split strongly and clearly giving plants with
violet and plants with white flowers in the ratio 3:1.

" Therefore line E is exceedingly well suited for further studies where
it is important to be sure that no inadvertant cross has occurred ; such
a cross would at once appear in the colour of the flowers. Line E then,
as I have described in my paper ' Om nogle Mutationer i rene Liniers '
(Biol. Arbejder tilegnede Eug. Wanuing, K^benhavn, 1911), has twice
shown a sudden change in genotypical constitution, ' mu(e)tation ' —
and that without any alteration in its white colour — and one of the new
biotypes, which from its appearance (probably by vegetative mutation)
was homozygous and as such had proved constant for a number of years,
was chosen to test further the nature of the mutation. A cross was made
between the original, unaltered, E type and the mutant in question,
mutant x original. The immediate result of the cross was three seeds
which were sown and gave in all 62 descendant seeds, which have thus
Fo embryos in F^ seed-coats."

Of these 62 seeds, two were rejected as misformed and the remaining
60 sown in six rows of 10 each, with two control rows of E and M
respectively. It was the harvest from these 60 beans that was handed
over to me in the Autumn of 1914 for analysis.

The following figures from Professor Johannsen's paper show the

amount of the difference between the original pure line E and the long

mutant, which, for convenience, we call M.

Original type ... L 12-459 i? 9-017 7 72-4

Long mutant ... 1,13-564 £9-137 7 07-4

(/ is the length-breadth index, 100 B -r- L).

Or comparing the indices through a number of years.

Year Original tyi>e Long mutant Difference

1906

72-4

67-4

50

1907

77-8

71-2

6-6

1908

72-4

66-8

5-6

1909

75-3

69-5

5-8

1910

73-3

67-1

6-2

1914

68-6

650

3-6

1917

69-2

64-3

4-9

1 86 Segrerjation of a Quantitative Character in Beans

With regard to the general conditions under which the further ex-
periments were conducted, in 1914, 1915 and 1916 the beans were grown
in the voliere in the experimental part of the Botanical Garden. The
year 1916 was in every way unfortunate. The weather was bad and the
ground " bean- tired." The results for 1916 are therefore not as good as
for the other years. For that reason, the beans were sown in 1917 and
1918 outside the voliere in the experimental garden. The ground in
which the beans were planted was, as far as possible, uniform, and they
were planted exactly equidistant from each other. More or less room to
grow has considerable influence on the size and number of the beans
produced by a plant, and to avoid disturbances from that cause, any gaps
occurring from failure to germinate were filled with beans from another
pure line with violet pigment in the flowers and stems and black seeds
of a different shape. No trace of cross-fertilisation ever appeared in
either race, inside or outside the voliere.

The beans are sown in May and ripen about the end of August, and
after harvesting are spread out to dry in a loft. They must not be dried
too quickly nor subjected to much warmth or the pods will open and
much material be lost. When the beans are thought to be dry, weighing
tests are made, a sample fresh from the pods being weighed, left for a
few days in a warm room and again weighed. When the weight is con-
stant all the beans are removed from the pods, plant by plant, and carefully
examined. Any that may be unripe, all misshapen beans and all that
are spotted with fungi, though only to the extent of slight discoloration,
are rejected. The reasons for the rejection of these first two classes of
seed are obvious. Beans that are unripe or misshapen will give measure-
ments that are abnormal and incorrect. But neither unripe nor misformed
beans are numerous, and misformed ones occur usually together in pods
that have been curved in growth by some accident of position or by
injury. The most serious ground of rejection lies in the attacks of fungi
and seeds so injured must be rejected for two reasons. In the first place,
it cannot with certainty be said that the fungus is without influence on
the size of the bean, and, in the second place, not only are such beans
useless as seed themselves, but they are liable to infect others stored in
the same glass.

Each year the original pure line and the mutant were sown as con-
trols, and in 1916, 1917 and 1918, when the number of rows had become
big, ten rows of controls were sown in all, five of E and five of M, dis-
tributed, one of each together, at equal distances over the field. The
variation in the controls will give a picture of the differences due to
inequalities in the ground.

I. Leitcii 187

In measuring the beans both length and breadth were recorded. But,
as far as the problem in hand is concerned, the length is of uujst im-
portance. The range of difference being greater, it is obviously easier to
distinguish differences in length than in breadth. Therefore, in the data
accompanying this paper, only curves for length are included. Further,
since after 1914 it was impossible to continue measuring every bean, a
random sample of 25 beans was measured from each plant.

In the analysis of the material two different methods of representation
were employed. The first was the plotting of the distribution curves for
the lengths and breadths of the beans of each row of plants. Except in
1914 each row contains plants from the seed of one parent plant only.
The curve of a row in 1915 will therefore be an expression of the geno-
typical nature of the plant from which its seed was selected. And so for
each year. The curves for the six rows in 1914 are not so distinguished,
and from them only general deductions can be drawn as to the general
nature of the 60 seeds sown. These curves of distribution were plotted
directly and not calculated to constant area ; for, plotted directly, charac-
teristic differences are more striking, such as the flatter form of typical
M curves and the typically lower fertility of the 31 type.

This method affords little help in the selection of plants for further
breeding. P'or that purpose some method that would give a compre-
hensive graphic survey of the material is desirable, and is found in the
plotting of what I call charts. On these charts length and index ( lOOL/B)
were plotted, the x axis giving length and the y index, and each plant
being represented by a point whose co-ordinates were respectively the
average length and index of its beans. Such a chart was made each year
for the control plants of E and M on ordinary mm. paper. Over it the
limits of their distributions were sketched on transparent mm. paper
and the hybrid forms plotted there. This gives an immediate picture of
the position of every single hybrid plant in relation to the two parent
lines and the forms can be classified on inspection. The method is of
course cumbrous and ill-suited for reproduction, but it proved of greater
practical value than any other.

As to the 1914 harvest then, the 60 hybrid plants, bearing seeds
containing Fo embryos in F^ seed-coats, show in their distribution curves
a range of variation from below the lowest limits of the line E to the
upper limit of 31. And this fact at once suggests a complication of the
problem. A distribution varying from the lowest limit of E to the upper
limit of 31 might have been expected in accordance with what Professor
Johannsen found in the hybridisation experiments already referred to.

188 Segrerjation of a Quantitative Character in Beans

But it must be noted that to give such transgressive distributions there
must be a positive difference in genotypical constitution between E and
M. Of course it could not be said at this early date that these indications
were reliable, but at least the indications were there and the jDroblem
had become enriched by the question of the cause of the transgressive
variation.

The chart of the hybrids bore out and emphasised the evidence of
the distribution curves, for here the measurements do not merge as they
do in the curve for a whole row, but individual differences between the
plants stand out. Some of the plants with the smallest beans lay quite
outside the lowest limits of the E biotype, and there was an absence of
forms intermediate between E and M, those which one would have ex-
pected to be most numerous. In selecting seeds from the hybrids for
sowing in 1915, beans were chosen which, from their position on the chart,
relatively to the chart of E and M, seemed likely to prove either segre-
gated E forms or segregated M forms ; in addition, two of the strikingly
small lots and one which, alone, seemed to occupy a satisfactory inter-
mediate position between E and M. These will be found in the first
genealogical table (p. 202) marked, in 1914, with letters indicating the
forms to which they were thought to belong; in all, six apparently M,
two extremely short, the one intermediate just mentioned and the rest
apparently typical E.

The harvest from these beans was as before gathered and preserved
plant by plant, and a sample of 25 beans taken at random from each
measured. The result was that in the six rows of plants whose seed had
been selected as M, the beans of each plant belonged to that type ; in
the two rows whose seed had been selected as extremely short, all the
plants bore seed whose distribution curves lay well to the left of the
control curves for E. On the charts, the M seed gave only M beans in
every case ; the extremely short gave only forms lying in a group to the
left of and higher than the E group. On the other hand, the distribu-
tion curves for the plants from the E seed showed the same transgressive
variation as the 1914 curves ; on the charts the plant averages stood out
sharply and clearly as segregates, some belonging to the extremely short
type, which we shall henceforth call X, some apparently E and some M.

The distribution curve for the intermediate — seed from plant 46, i,
in 1914; row 9 in 1915 — maintained its position, and on its chart the
group of plants lay intermediate between the E and M groups with-
out any evidence of segregation such as appeared in the groups last
mentioned.

i

I. Leitch 189

Thus a new type, X, shorter and broader than the original E type
has been established ; the segregated M remains constant and the
apparent E seed gives plants bearing seed which is, plant by plant,
either X or £" or M. Taking these segregating rows alone, seven rows
in all, and counting on the charts what forms are M and what are
either E or A'^, the rows give respectively, 13 (E or A') and 9 (M):
17 (E or X) and 5 {M): 17 (E or Z) and 7 (M): 18 (E or Z) and

2 (il/): 15 (E or Z) and 7 (M): 16 (E or Z) and 5 (M): and 10
(Eor X) and 3 (M); in all, 112 non-M and 38 il/, which is an ideal

3 : 1 ratio. At first no significance was attached to this ratio. In view
of the intermediate between E and M, row 9 in 1915, and of the new
Z type, it was to be expected that il/ would also produce new extreme
types.

To throw light on the questions which now arise the following
selection was made for further breeding ; two M types, the intermediate,
and four of the segregating types. The rows chosen were rows 12, 5, 14
and 21 (segregating), 9 (intermediate) and 18 and [20, 22] (M), and
these numbers have since been used to indicate the seven groups of
descendants. The following summary account of the further experi-
ments in each of the seven groups will be most easily understood if
reference is made while reading to the genealogical tables on pp. 202 — 4.

As group 21 offers no complication of any kind it is a suitable
starting-point. For sowing in 1916 were selected four of the extremely
short types, two apparently E and two supposed M. The short were
planted in rows 53, 48, 41 and 7 in 1916 ; the E seed in rows 12 and
17 and the 31 seed in rows 34 and 22. From the distribution curves for
these rows it appears that the small types remain constant in position
relatively to the control curves from the original lines ; that the plants
from the M seed bear beans true to the il/ controls ; and, on the chart,
the two apparent E types give some plants with Z beans, some with E
and some with il/. With regard to the short A" type there might i)ossibly
be a question as to whether row 7 is identical with the three rows 53,
48 and 41. It might be supposed that 7 was an intermediate between
Z and E. To test the point three short variates and one long were
selected from 7 and planted in rows 63 (long) and 64, 65 and 66 (short),
in 1917. The curves for these four rows are practically identical and it
is certain that no segregation takes place. Further selection of plus and
minus variates in 1918 also failed to show any difference. That is to
say, to judge from group 21, the X type shorter and broader than E,
once isolated, is fixed and constant.

190 Segregation of a Quantitative Character in Beans

To test further the constancy of the M segregated types the longest
four and the shortest (7) plant from row 22 were used as seed again in
1918 giving rows 46, 47, 48, 55 and 56. It will be seen that these give
identical curves corresponding in type to the control curves for M in
1918. Further the two groups 18 and [20, 22], sown in duplicate from
the seed of one plant in 1914, were continued in 1916, and row 20 from
group 18 in 1918. It will appear also from their curves that their seed
corresponds with the control M types and with the Ms segregated from
group 21.

Group 12 behaves in a similar manner. In 1915 three M types, two
extremely short, three apparent Es and one doubtful were selected for
seed. In 1916 the Ms bred true, the Xs bred true and the apparent Es
and the doubtful type split. Of particular interest in this group are the
curves for rows 23, 24, 25, 26, 27 and 28 in 1918, representing selected
short (23, 24, 25) and long (26, 27, 28) variates from an X group in
1916. The curves are identical, indicating again that the X type is
homozygous and that therefore selection has no effect in altering the
type.

Row 10 (1916) appeared of a rather doubtful composition. Although
in general M in type, there were two doubtful, possibly intermediate
plants, and one which produced certainly very few beans, but these so
short that the possibility at least of something unusual had to be
admitted. These were therefore sown ; rows 9 (ext. short) and 13 and
14 (intermediate), with two of the longest types from row 10, rows 10
and 11. The curves show no difference of significance; the difference
between 13 and 14, which seems the greatest, is in any case not greater
than that between 68 and 18 of the control pure Ms.

In groups 5 and 14 occurred the only serious disturbance that arose,
in that types selected as X did not in every case j)rove to be X. In
view of the subsequent behaviour of their descendants they must,
however, be regarded as very aberrant members of the intermediate
heterozygous group usually selected as apparent Es. There is always
transgressive variation, and 1916 was a very bad year. It will appear
that other types too suffered, the beans being shorter than usual ; and
as the chief cause was, no doubt, that the ground was " bean-tired," the
shifting of the types to the left had probably begun in 1915.

In group 5, then, there were selected in 1915 four supposed X plants,
four E and two intermediates. In 1916 only one of the rows from sup-
posed A'' seed bred true. The other three and the four Es showed
segregation, while the two intermediate proved to be M. It is signi-

(

I. Leitch 191

ficant that the intermediates in this case proved to be M, indicating
that the same disturbance is operating here.

In*1917 the extremes from row 3 were sown, and as they give a
typical example of the method of segregation, the chart of these plants
has been chosen to illustrate the method. The curves for row 3 and
other typical forms are also given. The constant X, row 45, was con-
tinued in the four rows 27, 28, 29 and 30 in 1917, and as 28 and 30
showed each one plant with beans of unusual length these two plants,
with two short variates, were continued in 1918. The curves for these,
31, 33 (short) and 32, 34 (long), show at once that the aberration was
without significance and that here again the X line breeds true.

The other rows segregate as before, giving X, M and a type that
again splits.

In group 14 a similar disturbance appears in that, of the two
apparently X types chosen as seed in 1915, one bred true and the
other proved to belong to the segregating type. One E type chosen
proved as before heterozygous. Three doubtful intermediate types were
selected, and, of these, two, as in gi'oup 5, proved to be M while the
third proved to be entirely different. It gave rise to a group of plants
occupying a position similar to that of the parent plant, and like the
9 group already mentioned intermediate between E and M. Continued
in 1917 this intermediate group, 6, shows a more scattered distribution
but little or no evidence of segregation. To test the point, the extreme
short variates (2) and longest (1) in row 37 were chosen as seed, the
shortest and longest, at the same time unusually narrow, from row 39
and the two longest from 35. The two short 37s give rows 35 and 36
respectively in 1918 and the long variate, row 57. Both chart and
curves show that there is no difference between them ; indeed the
curve of lengths for 57 lies, if anything, slightly to the left of those
for 35 and 36. As regards the two rows selected from 39, 18 (the
aberrant type) and 37 (the short), there is admittedly a slight differ-
ence. But the range of 18 is characteristic for the more diffuse types
in this group, and 37 corresponds closely to the other rows just dealt
with. 18 shows no trace whatever of the unusual breadth of its parent
seed in 1917. It must therefore be concluded that there is no evidence
of segregation in this group. Group 6 belongs to a non-segregating type
intermediate between the original lines E and M.

The same behaviour is found throughout the group 9. From the
1915 material were chosen seven plants for seed, including those with
the shortest and the longest seed. The general result was a group.

1 92 Seciregation of a Quantitative Character in Beans

somewhat diffuse, but still intermediate between E and M. For further
analysis were chosen two very short types, one from row 9, the other
from row 36 ; two very long types fi'om row 31, and a number of^ypical
forms between these limits, in all 18 rows. The two shortest types gave
rows 49 (from 9) and 62 (from 36) in 1917; the two longest, rows 56
and 57. In 1917 rows 49, 56 and 57 are indistinguishable and the
group of descendants is in this case very strikingly homogeneous and
again intermediate between E and M. To this there is however one
exception. The other of the very short forms, row 62, gives an isolated,
very short, group immediately distinguishable from the intermediate
group itself, and at once suggesting an X group. Continued in 1918
the 62 group, of which the three shortest and one longest plants were
sown, proved as exj)ected to give an X group, identical with the X forms
segregated from other groups. Three extremely long types were also
continued in view of the possibility of segregation taking place in that
direction also. The result was, as before, negative.

Thus, this group, with the exception of the one case in which an
X group is segregated out, behaves like the 6 group, remaining inter-
mediate between E and M. It seems legitimate to assume that the X
was produced by splitting in the same way as was found typical in the
other groups; and so, to sum up the results, the hybrids between E
and M either split into M, an entirely new type shorter and broader
than E, and an intermediate type that again splits ; or they do not split
at all. The non-segregating type arises twice, independently, and in
different years.

As to the significance of these results, first it is evident that the
theory that mutations are due to the loss of a factor or factors cannot
be applied here. From the moment the X type is established, the
theory of loss of factors is insufficient. No system of recombination of
factors, cumulative or otherwise, could explain the appearance of a new
type lying outside the limits of the original forms. It is necessary to
regard the difference between E and M, not as a case of presence and
absence of a factor or factors, but as a case of positive difference. This
does not necessitate the assumption that anything has been added.
The simplest view seems to be that a factor has been modified in
the original pure line, to give the mutation. Why not more than one
factor ? Because the fact that, as already indicated, the typical splitting
into X, intermediates, and M, gives a ratio of non-ilfs to Ms of 3 : 1, a
result confirmed in 1916 when the total number of rows splitting was
13, giving a total of 176 non-ilf plants; 63 Ms points to there being

I. Leitch 193

only one change. That, together with the faihire of selection to show
any further splitting in the X or M gi-oup, seems a sufficient indication
that only one fixctor has been altered in E to produce the mutation M.

However, this modification must not be taken in any restricted
sense. There is every reason to interpret the change in the broadest
fashion. In the first place, as will be evident from the curves of distri-
bution, and as is strikingly evident in handling the material, the M
type produces usually fewer beans than the original E type or the
X type. But there is certainly a correlation between the number of
beans produced by a plant and their size, and whether the difference
between the types be regarded as a difference in size or in fertility,
primarily, is worth considering, though the point may not be of great
importance. Further the M type is much more liable to be attacked b}^
fungus than the E and X types. No attempt has been made to deter-
mine the reason for this, but it indicates that the change from E to M
must be regarded as a change in the reaction of the whole organism.

If the change from E to M could be regarded as a sudden discrete
transformation of one factor into a modification functioning as a new
factor, the hybrids between E and M would be doubly heterozygous and
they would be expected to split into a variety of types including some
such type as X and possibly a type still longer than M. If the foctor
altered be called A and it be considered as modified to B, then the
hybrids would have the composition AaBh, and to explain what really
does happen, it would be necessary to postulate complete coupling
between A and B. In addition it would be necessary to assume that
forms containing neither factor are indistinguishable from aaBB (the
M form) ; the X form then would be the double homozygote with both
factors ; or alternately, if presence of factors be taken to mean additional
size, and B be regarded as modified into A, it must be assumed that
AABB is indistinguishable from A Abb. And even so, the difference
between the two heterozygous forms, one of which is phaenotypically
like E and the other intermediate between E and 31, remains to be
accounted for. This may prove to be a very far-fetched attempt to
reconcile the results with Mendelian results in general and until other
work on similar material has thrown more light on the subject, it is not
a very profitable discussion. The one point stands out, that the theory
of loss of factors here fails to account for the nature of the mutation.

With regard to the difference between the behaviour of the groups
of heterozygous plants in general and the two aberrant, similar groups (1
and 9, Professor Johannsen has suggested the possibility of cytological

194 Segregation of a Quantitative Character in Beans

difference, and efforts will be made to have the types examined by an
expert. Where there is anything unexplained there is always hope of
fresh discoveries.

-| — I — I — r

1 — I — r

1 — I — r

1 I r

J L

r'"-r-'"]--t- I

U' 11-' 11-1 12" 12' 12^ 12^ 13° 13' 132 13^ 14° 14' 14» 14^ 15° 15' 15' 15^ 16" 16' 16' 16^ 17"
Fig. 1. Typical Control Curves, 1917. 1 and 50 : £. 2 and 51: ill.

^

T—

"1

r"

n

r

1 I

1 1 1

1 1 1 1 1

1 1 1 1 1 1 1 1 1

1

\ ^

,6

bU

-

\

1 9

~

40

_

/i

\l5

-

30

-

/ 1

^^ \

-

20

-

/

^

/

,-/

^-

^

-

10

—

-k

^

_y

y

V

y

1 1

1 1 1

1 |\-4- 1"^ 1

1 , 1 >-4— 1__I_ 1 1

1

10^ 11° 11' 11'^ 11^ 12° 12' 12' 123 13° 13' 13' 13^ 14° 14' 14' 14^ 15° 15' 15' 15' 16° 16' 16' 16^
Fig. 3. Group 0 (1917). Segregates from 3 (1916).

1

1 > 1 1 1 1 1

1 1 1 1 1 1

...

1

1

1 1

1 1

uu

gQ

80

-

/ '.

61

-

70

-

•.62

-

60

-

-

fiO

_

-

40

_

^

-

30

-

-^ /V^5^

-

90

—

^Z /-..^

^— "''■ "^^ "^^

—

10

..I..-

■■■[•.■' 1 _i-::^f-^r . 1

•• N •

1 1 1 11 1

V —

L.

~f-

id.

-^ ,

, ,_"

11° 11' 11' 11^ 12° 12' 12' 123 ,30 131 135 133 140 141 14' 143 15° 15' 15' 15^ 16° 16' 16' 16^ 17°

Fig. 4. Group 9 (1917). Segregates from 66 (1916).

JOURNAL OF GENETICS, VOL XI. NO. 2

75

1 — I — 1 — \ — \ — \ — I — r

T — I — I — r

©

74

73

©

72

71

19 . _ 20

19 19 20 / 16

/ 20 20

/l9 19

1619 '^/ ^-. 1

,6/ © '°

70

21 / 19

69

®

68

67

V ■-'

1 23 ^'~'

\

66

\ '2'!

65

64

63

62

61

60
11

I I I

I I I I I I L

J I L

6789 123456789

2 3 4 5

Fig. 2. Rows 15, 16, 19, 20, 21, 22, 23 in 1917.

Dotted lines give limits of distributions of controls E and .V.
The .r-axis gives length : the (/-axis index = 100 length/breadth.

1 — I — r

1 I \ I — I — r

\
\

23 /

16 " /

©

23 .. 23

\ 22 22

22 \

\

\

\

\

\

\

®

\

\

\

J \ I L

I I I

75

73

72

71

70

69

68

67

66

65

64

63

62

61

60

789 ,123456789, ^1234567
14 10

The numbers refer to rows and each number represents a plant.
Circles indicate that these plants were selected for sowing in 1918.

I. Leitch

195

« 00

(M <M i-H CO lO

00 Ol

I M O T»)
1-1 O 00 »ffl iM

i-H .—I

C*l 05 05 O 00

■^ (M i-i e(5 I

rl* »o "H CO iM

t> «5 CIS 00 iH

05 CO 00 O t~

»C ?0 <M 00 iM
i-H ,-t »-( 1-1

55 1~ c- m 00

r-( i-( rH Ol 1—1

05 05 00 00 -^

tH <M I CO iH
iH CO i-l 05 to

»H U5 (M O «0
rH

CO «0 ■■# CO i-H
—I i-t

t» t- 1* ca OJ
1-1 (M

I 00 ic

I (M r-l

>0 (M
I (N 1-1

1 »o te

O 'I* IM CO CO

1-1 1-1 rH iH

O iO lO >0 00

I-l

■^ CO •«*' fH >ft
■M »-( -t(

to "* lo 00 lo

rH rH

>ra ff^ CO o 05
1-1 1-1

CO 'I* 05 i-< rH

X !0 rH

I
I

C5 — c<i to eo

1-1 IM rH

•^ ~ 00 CO -^

O •* CO C^l tH

1—1

-^< -* CO 1— I Tf

to

I— I

OS

I I
I I
I I

I I

! I

-f I -H I IC

IN I f 1-1 O

o c^i i>- OS 00

Ol t» OO U3 t»

^ -^

05 I -^ 00 iH IM

3S M CO 05 lO t>

1-1 i-l rH

lO t- C5 IM 00 -^

CO ■^ -1< IM

CO 00 IM 'O CC
rH rH

00 1(5 O Ci IM

rH 05 00 en >ffl rH

(M rH rH rH rH rH

CO 05 00 O CO ^

rH

•^ t> «0 -it c- CO

IM -^ <N CO rH to

«5 t-l r^ ,-\ n I

to 1^ I I I

^ I I I 1-^

IM C<l »o C5 e<5 to

rH M '* U5

O c; O -^ CO

CO t-l

IM iM -^ l« W

rH

QOrHiMlOlM CSiMtOCSrH
rH rH rH rH

*101 I |rH C-CJ-»t--rC>J

IrHrHC^rH C0>Oa0»O?n
rH C<1 CO -t "-H
I I "5

I 1-^

-" I I

I I I

I I I

Journ. f>f Gen. xi

13

196 Segregation of a Quantitative Character in Beans

S

1^

s

r2

g

'Jh

:; -^

o

!>5

»-■

• W

<M «5 iH i-l CO 1— I

CO O i-H <M

iM i-l CO -^ I <M ■>* C<1

(M <M 05 -^

CO W5

lO OS C~ O

r-l l> 00 •-<

1—1 I— I 00 05

t- O C^ «5 (» 00

o o 00 oi o 00

t~ 00

00 00

C- 00

OS CO
I-l

CO CO

CO I>

1-1

-^ o

1— 1 iH

05 '^tt

?o ?o

«5CO

^ i-l

(M .-1

IM CO C» O CO 05
CO CO CO -^ -rf l>.

1-1 O » r-l I-l OS
I-l I-l ,—1 ,-1 r-l I-l

CO «5 00 IM -^ -rtt

c^ ■

(N

CO CO I CO

I CO I

05 S

r

CO

■ij

s

§5

"^

I-l

?>

O)

^

rH

?>

5»

c^

o

6 ii

1^ I

■^1-1 I 'C CO

CO 1

«5 IC

CO (M I t- 00
I rH

«5 OS I 00 <N

r-l >0 (N

u? O <N

r-l O -t* IM

I CO rH
I rH rH

CD

-* W I

lO rH T)l

lO CO Ttl
f-H

CO ■* rH W5 t» CO

IN t-^

COCOODrHOOCdCO'^t-'COSO
COIMrH rHCOrHrHTt((M-)<

COOl [(MlMrHt-CO-^lOt-rHrHOOOOOlMaO
«5M ICDIO CO-^C^ rHCOrH(N-*-<^CO

rHCOCOi-IOO<NrH»ra<NQOCOcOOC<le001»OeO
■•^IW -^IC rH»O-*C<)rH(M-^i7qcq-^11C0-^

-*O5COOS00rHCOiCTjlC5IMl-*JCClffl00Ttlt~<N
-^IM M<CO C<ICO-*<MrHCOTtl(MC<lCOCOCO

«(?C-C5(M-^COrH<M05rHeOt-OCOOi<MOt-
COCO lOCO lM-^COCOi-l(MlO-^COCOn3CO

050irHl0l0OC^'llrHC0Q0»C05'#C0C<IC0-^
00COCO-*CCCO(NCO-^>O-i«l-^'*«5»OCaCOC0

COiNCOt-OOi-^INrHOOQDCOJOOaOCOCOCO
■^>Ot-(MCOCOTtl-*COeO-*>OCOCS»CrH'M(M

OaO'^»OrH00005>0-^IO(NOrH050'*<C5
C0C0C-lNrHO«:)<NS^I>-»0»0>'5C0l0rH(NrH
rH

t~rHO5C0OSC0iMC0(MrHO5O5O5rHiI<lC0C0rH
C-IXtlClrH OCOO^SOCD-^-^COt^-^ T-i r-\

■*OiMOOO'+i-H-*(NiOCOiMeOt~t»-^<QO'*

(M-^00 rHOl^-rHr-ilOCDCD<M-*CO '-i

rHlMCOlC |'M05»OC5"^<NOCOVO<Mr-l»OCC
r-ICOt- IC^ ■X'rH IMIO-*IMCO<N

rHlflirHlO |'MCOOeO»OC5<MO(MOCO>C-^
rHSVJTtl |in-^rH IM-*(Mi-H(>JIN

rH IN I I <N rH i —

rH rH rH IN I Ol CO rH
T-\ <-i I rH (N

I O O rH

\^^ I
I 'O 00 I

I I

CO IN

00 O CO
rH CO i-H
CO IN rH

COC^-^iOOSCCOOOCOINiC-^OSC^rHOOOlN
rHrH ,-{ f-^ t-i r-\ rH(NrH<Ne<l

>^ :^'

§,-

T. Leitch

197

C^ S^ I i-H

CO Tj(
Tj( C(5 (N <£>

■<# CO t» O
(M

-H «5 O OT

D- O CO iH
1-1 tH CO

O <N iM c:
<N t-( (M CO

lO -^ CO CO
so (M CO CO

.H to CO (N

lO O 00 Ci

CO 1-H CO OS

rH t» »0 00 CO

(M(M»-ICOCO»Ht~OOCOCOCO

r-( 1-1 i-H i-l 1-1

CO CO

<35 00

CO "^
OS o

tH CO
CO CO

1-1

O CO

>0 00 o

00 irt ,_( 00

CO »0 IC lO

OO t- CO U5
«0 (M ,H tH

22 CO ^ o
t» <;>i CO ^

tr CD >o CO

CO T)< lO ■*

O OS CO CO

"5 ■* lO CO

OS S CO >ft
<M Tft >C (M

d *1 Ol "^1

T-l CO T}( (M

p o ira CO

1-1 (M <M rt

-^ OS U5 t-

1—1 r-i

l?>l CO -■* CO
'-' tH ■<# S<I
<M I r-l (N

c-cooOr-iTtte^it-iM

lO Tjt rH CO 1-1 >-• 00

(M 1-1 1-1 rH

ire K5 CO Ui 00 CO lO
rH 1-t

CO CO >o

(M OS <M so

r-l tH

SO t~ SO

■* 1-

eo CO <N 1*

1-1 CO SQ T)(

o t- *» i-<

i-( ^ so

»o CO •^ so

CO <N

O ■* OS CO

OS CO
CO -*

•* OS

CO CO

00 OS

(N iH

00 •*

CO ^

so OS

CO »H

■* o

00 >« o
I-l

CD CO CO

■^^ 1-1 Tf

^ I o

I iH Tj<

CO CO »o o

OSCO--l-*'*'?'«"^
CO 1-1 1-1 1-1 CO I Tj< U5

>re

1-1 so CO 1-1

198 Segregation of a Qaantltatlve Character in Beans

^

V

o P

r-l CO ■* tr-

(N (N i-l 00

■* 00

t^ OS

CO

O iH t~
CO

t- 1-H O

O I-H <M

T)t I— I

OS CO o

00 <M

CO

CO

CO i-H (N ,-1 ,H

t~ CO

■* l>. I-H

-* 00

rH

O <M

I-H rH

00 C~ I-l

CO ■^ ■^

I-( rH

05 ■*

00 00
lO

(M iH

CO

(M

i-l CO

<M (M OS
CO « CD

CO

^

■* OS to (N CO
rH (M iH CO

t- CO rH iN ■*

IN

<M

rH T-H <M lO
IN rH CO Crt
-<^ US rH rH CO -^ »0

CO 00

rH rH t- »0

rH "^ Oi CO lO

Tjt rH OS t-

I-H Tjt O I-H CO OS

CO «5 Oa IN I-H

CO U5 IN CO O IN

lO CD IN rH I-H rH

CO CD IN rH

I-H CO 1* (N

IN CO t^

rH

CO ^ CO

•^ IN Ttl

CO IN
IN CO

I-H
»0 rH O

CO Tt* CO

IN

00 00 (N
(N

OJ IN CO

rH IN

OS t- CO

rH IN

rH O O

rH rH rH

IN t- CO

I-H I-l

00 W "*

t^ Cfl Ttl

00

I-H IN ,-H OS

ira «5 CO

rH rH

CO IN 00

I-H rH rH

(N O OS

<-< r-t r-l

rH -I* 00

CO O »o
(N

U5 IN rH

>«

-* 00 O (N

I ^ I

'ftOCDrHCDCJt-TjIOS
rHrHlN(NCO00-*Ttl

t- rH I rH

COTjllOCOC~OOOSOrH(NCOTj<

rH rH rH IN IN

CD CO W O ■* •"*

lO CO O CD IN 00

OS CO rH «5 lO "^
rH (N rH IN IN "-I

o 00 e<i t- <N »-•

I-H I-H iH "H

3 I

(N IN
rH (N

Tjl CO CO OS

lO I »0 CO

rH I IN CO

I <N

I. Leitch

199

^ r-l

O 2

o ^

d P

I— I i-H ?c c— w' I c*a >— I

I-H c<i

■-H oi •.'^ ^ 1-1 ^ LI c;

rH oc rH i-l

2''OOt--*iMCi-l
(N (M OC <M e<l (M M

OilOOOOOCMC^"^
CO i-t -qi as (M S<1 (N

c~mt-oc~<Moooo

rHeOW-^OOUSIMC-l

00 IC C^l !>• C3 L-^ O "*

e<) CO <M L'T c-i L-j 1— 1 1— (

C^C5Ot-C0t~0Ci— II

O-^OC-O'-'iOO
»OiOir5-<l<!Ml«0Ci— I'

^C50CX0OC-Xl«'
OO G^ oc 03 "^ r-H '

s; 1.0 oc i-i I c>i t~

lO « <M t- I M O

oc i-H

I oc

--I M I »0 ^

I r-H 1— t

i-H t- oc 'a

IN I-H

C>1 00 oc oc

■* o

00 -*

> I-H :r ^ -M
I S5 >-< 'M -#
I I-H C- IC Tfl

IC:»C>10OC<10ClC.-l'

.-lS<105C<l'MOCCi'-lto-#
(Mi-IIMSOOl'MC'l i-l

?CC<>INC5C~t-CCSt'— loo
i-((NCCeCcqi-l(Ni>JOCtH

QOOCt-lOtNOit-XCiTti
Cq OS in oc TJH I-H <?! "M

oo-^xoi-iwisino

»OOiCO-'*'*i-li-IOCOCi-l

OinHC-C-t-C<JC5CCXCO
lO 05 O CO oc CO I-l

r-tCimooai'>it~ciiOi-i

LO I-l lO OO <M •* 1— I

C5 oc o; c-1
M -* ■* oc

^ O -rt* ->!

O oc -M f-H
■^ 1*' lO (M

kct~c5 0oci-i-rj<ci-i

OJ I— I I— I I— I 1— I

fM ^ ^: L- I-l I CO CO <N

1-1 05 -^

-*

C<1 X c^ ts

»« -* I r-l

■^ -r rj IC -*

■^ C^ i-( t^ ■>!
-f -^ X -1" -^
1^ t^ -^ ^ oc

oc -#

^■1 '

M O "M oc 'M "^

■-I d

X o ^ c; oc oc

I-l "N

X •N X I-l o t-
C<l oc -M 'M 1-1 C^

X O i.O C5 C oc

oc -^ -* 1^1 c-i oc

C5 •* oc {O
I-l IQ I-l ^M

5 -*l O «3 O
I >-l C^l (N -g

) X I— I O OS
I OJ I-l CI ^

o t- X I* ;s 2«}
1-1 C^ I-l I-l oc C>1

I O 05 >-'? c-

■ c^ oc ^
c~ O i.-; o

I-l 1^ C^t-<NOOX

1-1 C<l ^^

i vs ■N n -^

:s ■"<< C5 X
I-l I X oc

I Tj<(NOXc;«Ci-ioc-f-^

I ^H »H T-i

I Tji c>) >0 •^ W5 CO I I >-~ ?}

I 1 I

I rH so 1* t~ (M I I I M r>

I ^ rH I O I rt I I rH ,^

III II I ! I I I I II

»ooc5C'-i!Noc-^i;s50t-xc50 i-is<joc'^xc;Oi-<-»j<m

i-(rHi-|iM3^i<J^S<lWC<lC<lI^O0C CCOOOCCOWOC'Sl-'JtT)!^**

<>1 f-H

13—3

200 Segregation of a Quantitative Character in Beans

^- Jo

=C I— I

C::^ CO

■>' -7

fS' 1—1

I rH _f. (N

CO ic »c> eo

-^ t> C5 CO

O :d -f c£

CO O lO lO

c^ »o o C5

CO .-( C<l CO

-^ t- O CO
(M rH !M (M

1 CO -^ lO
I CO

,-H T-l O 1

CO rH I

,H 00 O ^

^ ■-(

-H i-H CD ir?

C^ IC <M CO
(M ^ CO -sf

T-l 1-4 IM r-H
.-I

'5-^ i I

O ^ <M I

lO CO CO I
CO I

-* IN -^ i-H
CO

CO -H I <-< I 00 «5

to ^ CO C^ IM X !N

-« l-H i-H rH

O «5

-t( CO

S5 (M C<J C5

1-1 1—1 t- 1— I

■* O ■* OS

iH rH CO "-I

00 ^ Ci IM

t^ i-f

CO ■* CI «5

1-1 r-l t~ CO 1-1 C~

CO CO Lo as

rH i-l rH OJ

IM X (N CO

1-1 rH <M

CO O lO o

-)t CO C<l CO t- -^
rH

rH (M I ira -!j( CC

rH CO "* N ; C~

D^ 00 CO d ^ C<>
CO

O O C<l CO ^ CO
rr> 1 — I rH 1— I

00 C5
CM

lO CO

00 O 95 rH

(M

[- ^ rH C5 ifS O
rH rH rH CO -^ IM

CO "M CO -^H

(M '-'

■^ (M CO <M

rH rH

rr t- rH 00

00

(M

c^

CO

CO

CI

c-

o

t-

o

rH

CO

t~

CO
rH

03

CO

CI

05

lO

C5

CO

CI

7-i

C5
rH

05

c

rH

CO

C5

00

r-i

CI

CI
CI

o

in

05

1-i

-*<

CO
CI

rH
CI

CO

CO
CO

CO
CO

rH
CO

Oi

CO
CQ

CO

CI

00

CO

rH

<CJ

CI

o

CO

CO

CO

<-<

»«

■CI

CI

t-

CI

CO

CO

CO

rH

o

CI

rH

co

CJ

o

CO

o

rH

00

T-i

CO

CO

^

c-

o

t-

o

1-t

CO

CO

-H

Ut)

00

CO

1

^

CO

T-{

m

•^

1

CO ,_| rH rH

rHrJflOCOt^OOOSOrHCl C5O>C>C0t^C0r--

COCOCOCOCOCOCO-^-*-* lOCDCOCO'OrHCO-

I. Leitch

201

III!

I i I

I I

%

^

o ;::;

M j IM rH -H 1-1
I i-H O QC) j I

-H nS 1* O C^l rH

-t* o X o CO c<j ,

■M (M t~ rO W 05 -*<
>-< f^l to ^ TO

« O '-; c; to (M 00

■M ^ «3 CO '^l SV1

£2 ec o o to O t~

C<I T)< O -<^ 00 M ^

O C5 GO to C5 C- !>.
-r M M CO TQ CO rH

o t^ CO o to 00 o

"5 -i* '^ Ifj C<1 Cvl ^

■— < to "^ LC> O CC to
"^ CC ^ CC C^ C"! C<l

-;^ °° ::i "=> ^' ^ -"

^1 1-H OJ ^ C<1 OQ T-H

~« 1— I O to c^ — H ~5
^ >-( r-l Tl

C^ CO CO X 'M U-^ -t<

•-H c-i I CO

■^ CO C>1 CO 1-1 o

CO-^rH^N |tOTj(C0«t~iqt0c>l-0'M>O i
I i-< I

^■^c^icotor^totoQOic-rnooi-ioic^-^to i

roic'MOi-i-ij<otoc^irtto-+<cocoi,':i^C5 •

c^to-*30'i;^ic5i>cot>o>ooi,otot>co i

^^'^25S'-'^2*="'-''-''^=''oo(M<?'o I

'M •MOl -*r-li-ICgC<lf-llNwS'-lSI

SS"S2SS'-''"'^otsi.'3tococ5ooco I

>2tOtOtO»OXt0^5>0-HC;Oi-li-l-M-f<QOtO
-? 1-1 r-l CI 1-1 ■M ■?! TJ< 00 CO C<) 1-1 IM Ol CO C^ >M

c<ii-ico.-icg(M'M~;^^^, ^^;;j,i;^^^-^

CO i-( rt 1-1 CI ^ ;M CO CO :»1 'M C5 r-H CO C^l '^ -^ ■>]

eOC^JO-.^COOOCOt^t^'M'MtOOOJ^COO
(Mi-ICMrHIMCOIMCOTHOJCQi-lrt-^fSMTiicO

25r!r:;rr!"*ri'^5^'*c50i-Oi-itDcococ5i>

M IM OJ OJ '^ 1-1 C<l (M rH 1-1 sq i-( (M S CO CO U5

S2~— 2?'^'-^ — ~3CC0-t<c0C5O5Q0t-

t~^'-SO-t<00-r<0'MX— 'tOO^tOOOOW
^ 1—1 COi-li-ITJi— I 1—1 1—1 CQi-ltO

— ::^Ot- |i.~t-,^:otOi-(Xt-icOCO-HCiCJ

i-li-l Ir-lrti-H ,-, ,^ ^

^^'MlCCO l-Tj<LOr-i-f<-t<^ I |COC<I-*OJfO
I '-I ^ II CO

C- -^ CO

C--1*

I I

o ^

I I

i>ntot-aocic<icor)iocoe-ooC50'

-" I

CO 1^
«5 i-H CO

•M C; i-(

t> lo c;

I CO 1-1

>C 10) •*
o C-. >-■: <

CO 00 ■* '

1—1 C^ CO to

1 to 1-1 rH

I o c^ to o
I rj i-i w iM

CO

(M I ■M rt

-^ I

r I

I I I I I

>; IS

202 Segregation of a Quantitative Character in Beans

Numbers indicate rows in the various years ; letters, plants. In connection with plants
the italic letters show that they were selected as A", E, M, etc. ; in connection with rows
they indicate the behaviour of the row as a whole.

The italic ^figures denote the numbers of nou-il/ and BI plants.

43

•44

Genealogical Table of Ey.M.
45 46

47

48

1914

c g J
M E X

1 f h

M M X

k

M

E int. E

f h

M M

1915

3 7 4

M 1319 X

15 19 6 8

M

flO)

116)

12

14 9 17 21 18

M X 1715 M 1717 1812 1517 int. 16/5 UjS M

[201

122J

M

Groups : 12 5

,non-il/j> 112

Total of -

Ms 38

14

21 18

f20[
1 22)

1915

1916

Group 21.
21

A' A' 31 E E M

7 53 48 41 34 12 17 22

A' A X X M X X M

E E

M M

1917

1918

63

I I
b h

64

a m

- +

65

I — — 1
I I
.i e
- +

66

m b

- +

g

c AllM

3 4 49 50 51 52 53 54 46 47 48 55 56

No difference All M

I. Leitch

203

1915

Growp 12.
12

f
M

h
M

M

1916

1917

■

i

.1

1

0

/

16
A'
E

M

21
M

2

6
V

32
X
E

.1/

37
X
E
M

44

M

49
X

i
M

1

1

i
M

1

c
X

1

1

a:

1

P
X

1

q
1

a

+
1

e

+
1

g
1

J
? int.

1

1

1

8
M

I
X

4
A'

1
X

1

6

1

7
21/

1

9

1
10

1 1
11 12

No difference

1
13

1
14

1918

23 24 25 26 27
No difference

28

Group

5.

5

1915

1

a b

d

e

i

i

1

m

u

V

'

\: E

1

E

1

X

1

int.

1

E

1

1

int.

1

X

A

1

1 1
3 8

1
13

1
18

1
23

1

30

1
35

1 1
40 45

1
50

Si

0. X
E
M

X
E

M

1

Sp.

M

X
E
M

X
E
M

M

Y

X
E

1916

1
1

1

1

1

c

1
1

e

R

h

c

g

i q

1
I P

1
d

e

g

h

X

\

E

1

X

1

E

1

E

1

M E

1 1

X

1

1

5 i1/

1

X

1

X

1

A

1

A

1

1
15

i'«

1
19

1
20

1
21

1
22

23

1
24

1
25

1
26

1
27

1
28

1
29

1
30

A

A
E
M

A

A
E
M

A
E
i¥

.1/

A
E
M

A

A
E
M

.1/

A

A

i?

A

A
1?

1917

1

a

1
c

K

1
c

d

e

g

i

1

0

1
f

E

1

M

M

1

A

1

X

1

E

1

A

1

1

A

1

2

1

1
39

1
40

1
41

1
44

1
38

1
45

1
31

1
32

1
33

1

34

1918

A
E
M

M

M

A
E
M

A

A
E
M

A

A

A

A

204 Segreyatloii of a Qaatititative Character m Beans

Group 14.

1915

14

]
int.

I

int.

int. E X

6

1

33

1
11

1
25

i
38

1
54

N.B. int.

M

X

M

X

X

E

E

M

M

'

1916

1

1

1

1
a

d

I

k

1

1

n

g

m

1

b
X

1

1
A'

1

1

n
X

1

1
31

1
34 3

5

1
36

1
37

1
38

1
39

1
40

1
41

1
42

All intermediate

1

A-

X

X

1917

L-

1
1

1 —

"1
1

1
c

1

1
e

c

1
m

d

c

, +
1

+

T

i

+
1

2

1

1

1918

1
59

e'o

1

35

i
36

1
57

1

18

1
37

All

in

termediate

1915

1916

Group 9.
9

4

int.

9
int.

19 27
int. int.

31

int.

36

int.
i?

I

I

51

int.

typical I
h p

I int.
b d e

J m p r

g k 1
+ + int.

I I I

c d e j
int. int. int. -

I I i I

43 44 45 46 47 48 49 52 53 54 55 56 57 58 59 60 61 62
int. inter mediate int. int. X

1917

1918

58
int.

62 61
int. int.

19 20 21 22
X X X X

i

#

CAMBRIDGE UNIVERSITY PRESS

Devonian Floras. A Study of the Origin of Cormophyta.
By E. A. Newell Arber, MA., Sc.D. With a Preface by
D. H. Scott, F.R.S. Demy 8vo. With a frontispiece and 47
text-figures. 17s 6d net.

Water Plants. A Study of Aquatic Angiosperms. By
Agnes Arber, D.Sc, F.L.S., Fellow of Ne^vnham College, Cam-
bridge. Royal 8vo. With a frontispiece and 171 text-figures.
31s 6d net.

The Origin of Man and of His Superstitions. By

Carveth Read, M.A., Lecturer on Comparative Psychology
(formerly Grote Professor of Philosophy) in University College,
London. Demy 8vo. 18s net.

The Influence of Man on Animal Life in Scotland.

By J. Ritchie, M.A., D.Sc, F.R.S.E. Royal 8vo. With 91 illustra-
tions, including many original photographs and drawings, and
8 maps. 28s net.

The Concept of Nature. Tarner Lectures delivered in

Trinity College, Cambridge, November, 1919. By A. N. White-
head, Sc.D., F.R.S. Demy 8vo. 14s net.

An Enquiry concerning the Principles of Natural
Knowledge. By A. N. Whitehead, Sc.D., F.R.S.
Demy 8vo. 14s net.

An Introduction to the Study of Cytology. By L.

Doncaster, Sc.D., F.R.S., late Derby Professor of Zoology in the
University of Liverpool. Demy 8vo. With 24 plates and 31
text-figures. 21s net.

The Determination of Sex. By L. Doncaster, Sc.D.

Demy 8vo. With 23 plates. 12s net.

Mendel's Principles of Heredity. ^By W. Bateson,

M.A., F.R.S., V.M.H., Director of the John Innes Horticultural
Institution. Third impression with additions. Royal 8vo. With
3 portraits, 6 coloured plates and 38 figures. 15s net.

The Methods and Scope of Genetics. An Inaugural
Lecture delivered 23 October 1908 by W. Bateson, M.A., F.R.S.

Crown 8vo. 2s net.

Growth in Length: Embryological Essays. By

Richard Assheton, M.A., Sc.D., F.R.S., Trinity College, Cam-
bridge. Demy 8vo. With 42 illustrations. 2s 6d net.

Cambridge University Press

Fetter Lane, London, E.G. 4 : 0. F. Clay, Manager

CONTENTS

All Bights reserved

PAOX

Frances Pitt. Notes on the Genetic Behaviour of certain characters
in the Polecat, Ferret, and in Polecat-Ferret Hybrids. (Wjth one
Text-figure and Plates XV and XVI) 99,

C. H. OsTENFELD. Some Experiments on the Origin of New Forms
in the genus Hieramim sub-genus Archieracium. (With Plates
XVII and XVIII) 117

H. Onslow. The Inheritance of Wing-Colour in Lepidoptera. V. Me-
lanism in Abraxas grossulariata (var. varleyata, Porritt). (With
six Text-figures and Plate XIX) 123

F. A. E. Crew, Sex-Reversal in Frogs and Toads. A review of the
recorded cases of abnormality of the reproductive system and an
account of a breeding experiment. (With twenty -three Text-
figures) 141

I. Leitch. a study of the Segregation of a Quantitative Character in
a cross between a pure line of Beans and a Mutant from it. (With
four Text-figures) 183

The Journal of Genetics is a periodical for the publication of records of original
research in Heredity, Variation and allied subjects. The Journal will also, from time to
time, contain articles summarising the existing state of knowledge in the various branches
of Genetics, but reviews and abstracts of work published elsewhere will not, as a rule, be
included. It will be issued in parts as material accumulates, and a volume will consist of
three such parts.

Papers for publication may be sent either to Mr Bateson, Manor House, Merton,
S.W, 19, or to Professor Punnett, Whittingehame Lodge, Cambridge. Other com-
munications should be addressed to the University Press, Cambridge.

Papers forwarded to the Editors for publication are understood to be offered to the
Journal of Genetics alone, unless the contrary is stated.

Contributors will receive 50 separate copies of their papers free, A further 50 may be
had at cost price : these should be ordered when the final proof is returned.

The Subscription price for a volume is 425. net (post free) payable in advance ; separate
parts 15s, net each. Subscriptions may be sent to any bookseller, or to Mr C. F. Clay,
Manager, Cambridge University Press, Fetter Lane, London, E,C, 4.

Volumes I — X (1910 — 1920) are now I'eady, The price of each volume, in four parts,
stiff paper covers is 42s. net. Single parts can be supplied at 12s. 6d. net each. Quotations
can be given for bound copies of these volumes, as well as for supplying binding cases and
for binding subscribers' sets.

The Cambridge- University Press has appointed the University o-f Chicago Press agents
for the sale of the Journal of Genetics in the United States of America, and has authorised
them to charge the following prices : — Annual subscription, $ 8,50 net ; single copies,
$3,00 net each,

SPECIAL NOTICE. To effect economies in binding and postage a volume will, in
future, be issued in 3 parts instead of in 4. The size of the volume will however remain
as before.

CAMBRIDGE;: PKINTED BY J. B. PEACE, M.A., AT THE UNIVERSFXY PRESS

Vol. II, No. 3

December, 192 1

JOURNAL OF GENETICS

EDITED BY

W. BATESON, MA., F.R.S.

DIRECTOR OF THE JOHN INNES HORTICULTURAL INSTITUTION
AND

R. C. PUNNETT, MA., F.R.S.

ARTHUR BALFOUR PROFESSOR OF GENETICS IN THE UNIVERSITY OF CAMBRIDGE

CAMBRIDGE UNIVERSITY PRESS
C. F. CLAY, Manager
LONDON : FETTER LANE, E.C. 4
also

H. K. LEWIS & CO., LTD., 1 36, GOWER STREET, LONDON, W.C. I

WHELDON & WESLEY, LIMITED, 28, ESSEX STREET, LONDON, W.C. 2

CHICAGO : THE UNIVERSITY OF CHICAGO PRESS

(Agent for the United States and Canada)

BOMBAY, CALCUTTA, MADRAS : MACMILLAN Sc CO., LTD.

TOKYO : THE MARUZEN-KABUSHIKI-KAISHA

Price Fifteen Shillings net

Issued 30 December 192 1 ]

THE UNIVERSITY OF CHICAGO PRESS

The Origin of the Earth. (The University of Chicago
Science Series.) By Thomas C. Chamberlin, Head of the Depart-
ment of Geology in the University of Chicago, xii + 272 pages.
12mo. Cloth, lis 6d net.

This book, by one of the leading geologists of the world, sets forth the disclosures
that led to the rejection, one after another, of the older views of the origin of oiir
planet, the futile attempts then made to amend these or to build others upon the
same foundations, the final rejection of all tliese, and the construction of a radically-
new view based on a new dynamic foundation. The later chapters of the book treat
of the early stages of the earth and of the way in which its leading processes took
their start from their cosmogonic antecedents, these being held to be essential
factors in the genesis of the planet.

Heredity and Eugenics. By John M. Coulter, William

E. Castle, Edward M. East, William L. Tower, and Charles

B. Davenport. 312 pages. 8vo. Cloth. 19s 6d net.

Heredity anjl Eugenics is a book intended to present the recent developments of

knowledge in reference to evolution, heredity, and eugenics, and related subjects.

Five of the leading investigators in this field have collaborated in its production,

and the result is a work which will appeal to the general public, because it presents

the latest conclusions in a popular and ektremely interesting manner. *

Senescence and Rejuvenescence. By Charles Manning

Child, Associate Professor of Zoology in the University of Chicago,
xii + 482 pages. Svo. Cloth. 26s net.

The author of this volume, after some fifteen years of experimental investigation
of the nature and origin of the organic individual, has established certain facts which
afford a more adequate foundation for the general consideration and interpretation of
the age changes in the organic world than we have hitherto possessed.

Certain experimental methods have made it possible not only to follow the
physiological age changes in some of the lower animals, but to learn something of
their nature. The most important result of this investigation is the demonstration
of the occurrence of rejuvenescence quite independently of sexual reproduction.
The book differs from most previous studies of senescence in that it attempts to
show that in the organic world in general rejuvenescence is just as fundamental and
important a process as senescence.

Individuality in Organisms.- (The University of Chicago ,
Science Series.) By Charles Manning Child, x + 212 pages.
Small 12mo. Cloth. 10s net.
Professor Child's work is an attempt to state, and to present evidence in support
of, a conception of the nature of organic individuality which the author has developed
as a result of fifteen years. of investigation of the processes of reproduction and
development in the lower animals. In these forms organic individuality appears in
relatively simple terms, and it is here that we must look for the key to the problem
of individuality in the higher animals and man. The book appeals not only to the
physiologist and to the botanist, bat also to the neurologist, to the psychologist, and
even to the sociologist.

The Biology of Twins. (The University of Chicago Science

Series.) By Horatio -H. Newman, Associate Professor of Zoology

in the University of Chicago, x + 186 pages. Small 12mo. Cloth,

10s net.

In this volume many questions relative to twinning have been answered partly

through a study of data on human twins themselves and partly through a much

more detailed study of what has proven to be the most remarkable type of twinning

known— that in the armadillos of the genus Dasypus.

Sold in the British Empire (except Canada) by the
Cambridge University Press, Fetter Lane, London, E.C. 4

Volume XI DECEMBER, 1921 No. 3

DWARF FORMS IN BARLEY. nrw yoh«

ik/i anjcal
By BUNGO MIYAZAWA. uakubn

(With Plate XX.)

With the object of barley breeding a number of /", plants between
Goldenmelon^ and the Japanese variety called Sekitori were grown in
1914 at the Agricultural Experiment Station, Kanagawa-Ken, .Japan.
One of these F^ plants was crossed back by Goldenmelon in that year,
and a dwarf plant appeared in 1915 among the 96 offspring resulting
from this back-cross. As the hereditary behaviour oi-this dwarf may be
of some interest I am going to describe here briefly the results of my
investigations. The characters of this plant, together with those of its
parents, are as follows :

Goldetimeloii

Sekitm-i

Dwarf plant

Height of plant

tall

short

less than Sekitori

Bows in ear

two

six

two

Length of awns

long

short

long

Time of ripening

late

early

later than in Goldenmelon

Although this dwarf plant produced shoots more abundantly than
we have ever seen in any variety, over half of the shoots thus developed
failed to produce ears ; and moreover they were smaller than either of
the two parents in various respects, as length and width of K-af, size of
kernels, etc.

In November 1915 about 40 seeds of this dwarf plant were sown,
almost all of which came to germination, but towards Januar}^ or February
1916 some died, and only 24 completed their growth. Of the latter three
bypes are easily distinguished with respect, to the height of plant,
length of ear and awn, etc., each of which contains in its turn normal
and dwarf forms, so that we have altogether three kinds of noi-mals and
three kinds of dwarfs. Below I will classify all these simply into two
classes, viz. normal and dwarf. According to this classification the 24
plants above cited consist of 18 dwarf and 6 normal individuals, whence

1 Though this variety is here called by this name, it seems that it differs from real
Goldenvielon.

Joum. of Gen. xi 1^

206 Divarf Forms in Barley

we might perhaps be led to the false conclusion that we were dealing
here with the 3 : 1 segregation. In November 1916 seeds taken from
9 dwarf and 6 normal plants were sown ; in 1917 all dwarfs were observed
to undergo segregation, the offspring from 9 dwarf plants containing
altogether 292 dwarfs and 161 normals, i.e. 64*5 and 8.5'5 per .cent, re-
spectively, whilst all normal plants were found to breed true to their
type.

From the experiments above stated we see firstly that the normal
type is recessive to the dwarf, and secondly that all dwarf plants are of
heterozygous constitution. Were the segregation observed in 1916 of
the 3 : 1 type, only six out of nine families of dwarfs should be hetero-
zygous. Moreover, as the ratio of dwarfs and normals in 1917 is approxi-
mately equal to 2 : 1 (= 292:161), it is clear that we are not dealing
here with a typical 3 : 1 segregation.

The experiments of the iwo following years, viz. 1918 and 1919, fully
confirm what was said above. Thus all the offspring of dwarf plants
underwent segregation into dwarfs and normals in the ratio indicated

in the following table :

Dwarfs

Normals

Totals

Kesults in 1918
1919

599

748

304
375

903
1128

Totals ...

1347

679

2026

Expected

1350-66

675-33

—

From the above results we see that the ratio of dwarfs and normals
is approximately 2:1. On the contrary all offspring derived from normal
plants produced only normal plants.

Furthermore, various crosses were made between the offspring, and
also between them and either one of the two original parents, or other
normal varieties. The details of these crosses will be omitted here; it
will suffice to say that the results corresponded exactly to expectation.

Ph. De Vilmorin^ has studied the inheritance of dwarf forms in wheat.
He obtained two dwarfs against one normal and came to the probable
conclusion that such a result might be caused by the fact that any zygote
homozygous in respect to dwarfness is either not produced, or, if produced,
dies very early.

Now I have entertained the view that our case, which is very similar
to that described by Vilmorin, might be explained by either one of two
alternatives below stated, viz. (1) seeds containing zygotes homozygous
in respect to dwarfness do not germinate at all, or (2) they die off soon

1 "Sur une Race de ble naiu infixable.'' Journal of Genetics, Vol. iii. 1913, pp. 67—76.

B. MiYAZAWA 207

after their germination. To decide between these two alternatives, ex-
periments were devised to test the germinating power of the seeds : they
were sown in a Peti'i dish containing sand, as well as in the field ; the
result was that almost all seeds were perfectly viable, because the rate
of their germination was 90 per cent, or even more. The first alternative
was thus proved to be untenable. Now to pass to the second : since
some young seedlings might die fi-om the severe cold of January and
February, some seeds derived fi-om dwarf and normal plants were .sown
in pots late January 1919 and placed in a cold frame to protect them
against severe cold. A certain number of seedlings were transplanted
into the field March 1st, because the frost in this month is generally only
slight in the vicinity of our Experiment Station, and various plants,
especially barley, begin then to grow vigorously, .so that there is no
more danger of their suffering from severe cold. In late April I dis-
covered among these plants a quite new dwarf fm-m, very different from
the dwarf described above. This new form produces a great many young
shoots ; thus in some cases 152 shoots were counted in a stock; its growth
is very slow, so that even in late April, when the ordinary dwarf plants
have attained their proper height and are alread}' ready to produce their
ear, we could hardly see in the new form any indication of the production
of the ears. Its height too was so low as to be easily mistaken for some
other graminaceous plant (see PI. XX. fig. 1), Even in late May or early
June none of these plants had produced any ear. All were seriously
attacked hy Erysiphe graminis and their growth became graflually poorer.
By the middle of June almost all of them died, whereas the other two
types had completed their growth. This new dwarf type which is unable
to produce an earl shall call sterile-dwarf m contradistinction to ordiiiaiy
dwarf form,s.

The segi'egation observed in 1919 was as follows:

sterile-dwarf

Dwarf

Xornial

Trital

Actual

71

172

80

323

Expected...

80- 7o

161:50

so-7r,

—

We may therefore conclude as follows: even in the experiments of
1916, 1917, 1918, and 1919 a certain number of sterile-dwarf plants must
have appeared, but all of them died early under severe cold on aocoiiut
of their weak constitution, so that they escaped our observation. In the
experiment in which the seeds were sown in January 1919 they were
however protected against it and saved, so that tliey came under our
observation.

14—2

208 Dwarf Forms in Barley

The above conclusion was perfectly confirmed by the following breeding
experiment : a certain number of seeds of ordinary dwarf plants were
sown in early spring 1919 instead of the autumn of the preceding j^ear
as it is usually the case in Japan. By this method I was enabled to
observe the segregation into three forms just discussed very easily, and
the numerical results were as follows :

sterile-dwarf

Dwarf

Normal

Total

Actual ...

156

340

188

684

Expected

171

342

171

—

The ratio of the three types is approximately 1:2:1. The number
of sterile-dwarf plants, however, is in both cases somewhat smaller than
might be expected theoretically; this may be ascribed to the fact that they
die more easily than the others on account of their weak constitution.

Thus we reach the conclusion that the dwarf plant which was found
at the beginning of my experiments was heterozygous with respect to the
characters in question, and that it was intermediate externally, i.e. smaller
than normal, but taller than sterile-dwarf In other words, if we denote
the allelomorph for dwarfness by D and its absence by d, we have sterile-
dwarf = DD, normal = dd, and ordinary dwarf = Dd. The seed which first
gave rise to the dwarf plant Dd may perhaps have arisen from dd by
mutation. Cases similar to ours have been observed in respect of dwarf
forms in plants. This is however the first case recorded in which the
sterile-dwarf has appeared in addition to the ordinary dwarf,

EXPLANATION OF PLATE XX.

Fig. 1. On the left two sterile-dwarf plants ; one plant in the middle is normal and two

plants at right are ordinary dwarf types.
Fig. 2. In the middle is a normal plant and on either side are dwarf plants of the same

type. The characters of these plants are very different from those shown in Fig. 1,

and this is one of the three types mentioned on page 205.

JOURNAL OF GENETICS, VOL. XL NO. 3

PLATE XX

NOTE ON THE DETECTION OF SEGREGATION BY
EXAMINATION OF THE POLLEN OF RICE.

By F. R. PARNELL.

(With Plate XXI.)
Glutinous Rice.

For several years a study has been in progress on the inheritance of
the (so-called) glutinous character typical of a small group of cultivated
varieties of rice. These varieties differ from the ordinary starchy type
in that the grain becomes gelatinous when cooked in water, so nmch so
that if boiled in the ordinary way the result is a sticky mass resem-
bling glue.

Glutinous grains can be distinguished by their appearance since the
endosperm is dead white, giving a fracture like porcelain, whereas in
starchy gi'ains it is always more or less translucent.

Microscopic examination of sections of the endosperm shows no
marked difference between the two types, the cells in both cases being
filled with the highly compound starch grains characteristic of rice.
The action of iodine, however, brings out a very sharp distinction — the
starch gi-ains of ordinary starchy varieties take on the usual deep blue
colour, whereas those of glutinous varieties become reddish in dilute
iodine, passing through wine colour to dark brown as the strength of
the solution is increased. The same difference is seen when grains are
broken across and dipped into iodine solution.

The production of a reddish colour with iodine is characteristic of
amylodextrine, a rare form of starch that occurs in mace, and it is pre-
sumably the presence of this form of starch that gives to glutinous rices
their special character.

Inheritance.

In inheritance the glutinous character behaves as a simple recessive
to starchy but, siijce double fertilization takes place, the endosperm is a
fertilization product and the usual complication of results is obtained.

Thus when flowers of a glutinous plant are fertilized with starchy
pollen they produce grains with starchy endosperm. An F^ plant, selfed,

210 Detection of Secfregation of the Pollen of Rice

bears a mixture of starchy and glutinous grains. These give rise in F.,
to three groups of plants, bearing respectively — all starchy grains, all
glutinous, and a mixture of starchy and glutinous as in t\. Those bear-
ing all starchy or all glutinous grains breed true to these characters in
Fs, whereas those which bear the mixture give the three groups again
•mF,.

The mixture of grains on an F^ plant should give a simple 3 : 1 ratio
of starchy to f/lutinout>. This ratio is approached fairly nearly in the
numbers recorded, though there is always a slight excess of starchy. For
nine plants, each giving about the same ratio, the following total figures
were obtained :

Starcliy

Glutinous

grains

grains

Total of it plants

5,292

1,587

Expectation 3:1

5,15'J

: 1,7^U

In F.J the three groups of plants would be expected to give a 1:2:1
ratio of starchy : miwture : glutinous. In the families so far raised there
is a very considerable departure from this ratio as shown by the following
figures :

All starchy

Mixturu

All glutinous

Total of 81 families

y,2ii

13,729

5,021

Expectation 1 : 2 : 1

6,990 :

13,980

: 6,990

It appears most probable that the single factor explanation of the
difference between starchy and glutinous is correct but that some dis-
turbing influence affects the ratios in F~^. There are a number of possi-
bilities in this direction and some of these are being investigated.
Certain preliminary results suggest that differential germination and
dying-off may be responsible.

Pollen Binwrphisin.

In searching for some explanation of the F.j ratios an examination
was made of the pollen on F^ plants. With a view to distinguishing the
two genetic types the pollen was treated with iodine. The result was
most satisfactory — two distinct types became evident, one giving the
dark blue reaction of ordinary starch and the other the reddish reaction
of amylodextrine. With iodine solution of the right strength a very
sharp distinction can be obtained since the starchy grains become dark
blue whilst the others remain only yellowish.

Pollen from the two pure parents was then examined in iodine
solution. The starchy parent gave only the blue type of pollen ; the
glutinous parent gave the reddish type.

F. R. Parnell 211

Plate XXI shows photographs of whole anthers of the two parents
and F^. These were mounted in chloral hydrate solution containing
iodine, which shows up the pollen grains very well. Part of an F^
anther is also shown more highly magnified, as also some free F-^ pollen
mounted in plain iodine solution.

There can be no doubt that the two types recognizable in F^ pollen
represent the two ' genetic types produced by segregation. It thus
becomes possible to keep some track of segregation and to determine
the relative proportions of the two types of gamete in different plants,
flowers, anthers, etc.

Examination shows that the two types occur mixed throughout all
parts of each pollen sac. Counts were made of the number of pollen
grains of each type in various anthers, the latter being teased out in a
drop of iodine solution in such a manner that every developed grain
could be distinguished. There was some variation in the proportions of
the two types and the ratio was rather more uniform for anthers in the
same flower than for anthers from different flowers of the same plant.
In eighteen anthers, from three plants, the starchy pollen varied from
43-2% to 51-87^, the average being 48-17^.

An attempt was made to determine at what stage in spermato-
genesis segregation occurred. When this work was started the main
flowering was finished and only very poor material was available. It
appears, however, that there is little hope of success in this direction
owing to the absence of starch in the early stages. Pollen mother cells
in various stages of division were examined but no trace of starch could
be detected. The pollen grains themselves show no starch until they
are well developed and beginning to lay down reserve materials.

A further examination will be made of better material from a crop
that is now being grown. It is hoped also, if time permits, to look for a
similar indication of segregation in oogenesis.

A paper has recently been seen in which Bateson ' refers to the work
of Renner^ who detected dimorphism, due to genetic segregation, in the
pollen of Oenothera. The author has seen only a short abstract of this
work stating that the differences noted were in respect to size of pollen
grains and shape of their included starch grains.

It is probable that a systematic search in other plants would reveal
many cases of similar dimorphism in pollen, e.g. in the form of starch,

1 Bateson, W., " Genetic Segregation." American Naturalist, Vol. lv. No. 636,
January — February, 1921.

- Renner, 0., Ber. Deutsch. Bot. Gcsell. Bd. xxxvii. (1919), No. 2, p. 129.

212 Detection of Sefiregation of the Pollen of Rice

its presence and absence, or the presence and absence of other substances
recognizable by microchemical tests. This line of work appears to offer
distinct possibilities from the point of view of determining the stage at
which segregation takes place, especially if such dimorphism can be found
in a species in which the pollen grains remain united in tetrads.

DESCRIPTION OF PLATE XXI.

Iodine Reaction of Starch in Pollen Grains.

Fig. 1, Anther of s((rr(;/H/ type, pollen all dark.

Fig. 2. Anther of glutinous type, pollen all light.

Fig. 3. Anther of -P'l , pollen mixture of dark and light.

Fig. 4. Part of Fj anther more highly magnified.

Fig. 5. Free pollen of Fi showing two types.

JOURNAL OF GENETICS, VOL XI. NO. 3

PLATE XXI

Fig. 1.

Fig. 2.

l-'iK. n.

V-

•

Mh\

•

%A^

^?*s

Fig. 4.

FiK.

I

NOTES ON THE CYTOLOGY AND GENETICS
OF THE GENUS FUCHSIA.

By RUDOLPH BEER, B.Sc, F.L.S.

(With Plates XXH— XXIV.)

In contrast to the extensive literature which has grown up rouivl
the cytology and genetics of Oenothera very little work has been done
upon these subjects in the case of the Fuchsias. In 1850 Wiinmel(l7)
called attention to the irregularities in the number and size of the
pollen grains produced from the mother-cell of Fuchsia and in 1886
Wille (16) made further observations upon the same subject. He counted
five to fourteen microspores arising from a single mother-cell. He ex-
plained the occurrence of these supernumerary pollen grains by assuming
that " In dem Falle, wo bei Fuchsia sp. sieben Zellen entstanden waren,
konnte ich nicht darliber im Zweifel sein, das diese daher kam, dass
drei der Zellkerne der Tetrade noch sich einmal getheilt hatten, ehe die
Cellulosequerwande ausgebildet waren, wahrend der Vierte ungetheilt
blieb"(16 p. 61). Where five or six microspores were produced he
believed the explanation to be similar. In the case of the higher
numbers (e.g. 14) he was unable to follow the cell-divisions, and was
uncertain whether the additional pollen grains are due to secondary
divisions of one pollen mother-cell alone or whether they are to be
derived from two or more primitive mother-cells which have not become
separated from one another in the usual manner.

In 1907 the present writer (1) published a short note upon "The
Supernumerary Pollen Grains of Fuchsia and their mode of develop-
ment." It was found that the abnormal numbers of pollen grains which
develop from the mother-cells is due to the irregular distribution of the
chromosomes during the first meiotic division and that no evidence
could be found to support Wille's explanation of the phenomenon.

In 1912 Bonnet (3) published an account of the tapetal cells of
certain Angiosperms and included those of Fuchsia among the number.

Beyond this I can find no literature dealing with the cytology of
Fuchsias.

214 Cytology and Genetics of the Genus Fuchsia

Some time ago I continued my studies of the cytological features
which are connected with the development of supernumerary pollen
grains in certain species and hybrids of Fuchsia and concurrently
I began a series of observations upon the genetics of the genus.
Circumstances have intervened and prevented the completion of the
work, but it was thought that it might be useful to future workers
if a brief statement were made of such results as were obtained, and of
the lines of work which appeared to promise interesting results.

A. Cytology of Fuchsias.

I will first describe the facts which are to be seen in a hybrid form
known as " Alice Hoffman."

The early stages of the first meiotic division in the pollen sacs take
place in a normal manner as will be seen from PI. XXII, figs. 1 — 8.
The spindle which develops is apolar and the chromosomes become
arranged regularly at its equator at the conclusion of the prophase.

During the anaphase, however, the chromosomes move very irre-
gularly towards the spindle poles and some, either singly or in groups,
lag behind the rest, and often become entirely cut off from the two main
chromosome groups. This irregular distribution of the chromosomes is
shown in figs. 9 and 10. At the conclusion of the division these
scattered and separated chromosomes behave variously. In the majority
of -cases distinct nuclei are developed in association with them. In some
instances only a single chromosome may become organised into a small
nucleus, but more commonly small groups of two or more chromosomes
are associated together in the development of a nucleus. The size of the
nucleus which is formed depends upon the number of chromosomes
which enter into it. Examples of such nuclei can be seen in figs. 11
and 12. In several cases the chromosomes were so scattered at the
beginning of the telophase that they did not lead to the formation of
a number of separate nuclei, but they all became included within a single,
large, irregular nucleus. In fig. 12 one small nucleus is seen to have
developed round a chromosome which had become widely separated
from its fellows, whilst all the remaining chromosomes were included in
the large irregular nuclear body which occupies the centre of the
mother-cell. Whilst it is more usual to find nuclei organised round the
scattered chromosomes there are other instances in which these bodies
degenerate without giving rise to a nucleus. In fig. 18 the chromosomes
excluded from the two principal groups are probably about to undergo
degeneration. All the nuclei, both small and large alike, enter upon the

Rudolph Beer 215

second meiotic division in the usual manner. Spindles, varying in size
with the number of chromosomes, are developed in association with each
nucleus, and the course of mitosis proceeds in ([uite a normal manner.
No irregularities in the distribution of the daughter-chn^nosomes arc to
be seen in this division.

PI. XXII, fig. 17 gives a representation of the second meiotic division.
Single chromosomes each associated with a small spindle can be seen in
fig. 14. In PI. XXIII, fig. 16 we have a case which probably represents
the second division of one of the large, irregular nuclei referred to
above and represented in fig. 12.

The nuclear divisions which have been described in the foregoing
account are followed by the division of the cells so that each nucleus
becomes associated with a distinct cell.

PI. XXIII, fig. 19 represents a mother-cell which has divided up in
this manner to form the initials of a number of pollen grains.

In fig. 18 ten pollen grains will be seen to have developed fi'om
a single mother-cell.

It is readily apparent that the size of the pollen cell is dependent
upon the size of the nucleus with which it is associated and this, in turn,
depends upon the number of chromosomes which enter into its com-
position. The details of this relationship will not, however, be dealt with
in the present paper.

From what has been said above it will be seen that no evidence was
found of the existence of secondary divisions of the cells such as Wille
described, nor do the present observations give any support to this
author's suggestion that a fusion (or non-separation) of jjrimitive mother-
cells might occur in those cases in which the supernumerstry microspores
were very numerous.

In addition to the variety "Alice Hoifman " the meiotic divisions
of Fucltsia globosa and Fuchsia corallina were studied in detail. In both
these plants more than the normal numbei* of four microspores are
produced fi'om each pollen mother-cell. It was found that the distribution
of the chromosomes during the first meiotic division is quite similar to
that taking place in " Alice Hoff'man " and that single chromosomes, or
small groups of them, become separated from the rest, and usually give
rise to small nuclei. Here also some of the isolated chromosomes may
fail to organise imclei, but undergo degeneration instead. It will be un-
necessary to describe the details of meiosis in these two forms, but it
will be sufficient to call attention to PI. XXIII, figs. 20 — 24 which
represent the principal facts of interest.

216 Cytology and Genetics of the Germs Fuchsia

The pollen grains of Fuchsias follow the general Onagraceous type.
Their membrane consists of an exospore, a mesospore, and an endospore,
and is furnished with two or more interstitial bodies. The Fuchsias
have been divided into two groups according to whether their pollen
grains possess two or three interstitial bodies. Fuchsia glohosa is a
typical example of the group which f»ossesses three interstitial bodies
upon each pollen grain, whilst Fuchsia procunibens has pollen grains with
only two interstitial bodies.

Whilst these numbers arc characteristic of the majority of fertile
pollen grains in each case, yet they are not constant in those instances
in which irregular pollen development takes place. In these cases the
number of interstitial bodies appears to vary with the size of the pollen
grain, and in some of the smaller grains only a single interstitial body is
formed. Fig. 27 shows such a small, supernumerary pollen grain with
only a single interstitial body. Although not shown in this figure, these
small pollen grains, apart from the number of interstitial bodies they
possess, have membranes which are identical in structure and chemical
composition with those of the larger grains.

We have already seen that the distribution of the chromosomes to
the pollen grains is an irregular one, and that the small grains receive
only a small proportion of their normal number of chromosomes.

Notwithstanding this all the pollen grains develop walls which are
characteristic of the genus both in structure and chemical composition.
These facts have an interesting bearing upon the theory of the localisa-
tion of generic and specific characters in particular chromosomes, since
the chromosomes which any particular pollen grain receives is perfectly
haphazard in the present instance. The explanation is probably similar
to the one which has been suggested in the case of the development of
certain animal eggs in which the cytoplasm becomes set to a definite
line of development at an early stage. We may probably assume that
the cytoplasm of the pollen mother-cell has already been set, through
the influence of the still undivided nucleus, to a definite course of
development, and that it already has the mechanism implanted in it for
the formation of pollen membranes of a definite structural and chemical
constitution. It is a matter of secondary importance for carrying out the
work which is allotted to it at an early stage how the chromosomes
subsequently become distributed at the meiotic division, or how it
becomes divided up at the conclusion of that division.

From the foregoing account it will be seen that the development of
abnormal numbers of pollen grains from the mother-cells depends upon

Rudolph Beer 217

the existence of irregularities in the distribution of the chromosomes
during the uieiotic divisions.

Observations have been made to determine the extent of the
occuiTence of such irregularities of pollen development in a number of
species and hybrids of Fuchsia with the following results :

A. Forms with irregular pollen development or all steinle pollen.
Fuchsia glohosa.

F. corallina.

F. excorticata.

F. parviflora (no pollen development ; anthers collapsed).

F. Cottinghami (no pollen development; anthers collapsed).

F. macrostemma, var. conica.

F. macrostemma, var. discolor.

F. simplicicaulis (most marked early in the season).

F. arborescens.

F. Riccartoni.

F. virgata x F.fulgens (3iV).

F. virgata x F. fulgens (3 P).

B. Forms luith regular pollen development
Fuchsia fulgens.

F. procumbens.

F. gracilis.

F. rejlexa. '

F. virgata.

F. myrtifolia.

F. alpestris.

F. pumila.

F. corymbijlora.

F. venu^ta.

F. rosea.

F. pumila x F. alpestris (2 N ).

F. pumila x F. alpestris {2 F).

F. globosa x F. gi^acilis (11 ^4).

F. globosa x F. gracilis (11 B).

F. globosa x F. Riccartoni (10^4).

F. fulgens x F. virgata.

F. corymbijlora x " Ballet Girl."

The iiTegularities in the distribution of the chromosomes in the
Fuchsias frequently leads to the sterility of the pollen grains. Tischler

218 Ciftology and Genetics of the Genus Fuchsia

(15, p. 108) has pointed out with justice that such an unequal partition
of the chromatin need not necessarily lead to the sterility of the resulting
pollen grains. At the same time one cannot avoid the conclusion that
the fact of the existence of such an abnormal distribution is an earlj-
indication of a derangement in the mechanism of the cell and, in the
Fuchsias at least, this may become intensified in the later stages of
development so that many of tlie pollen grains which result become
sterile. This sterility of the pollen grains has frequently (see 2) been
regarded as the peculiar attribute of hybrids, but the list of Fuchsias
which has been given above does not support this view.

Thus F. arhoresceris, which is a pure species, produces a large pro-
portion of sterile pollen, whilst a definite cross between F. ptimila x
F. alpestris exhibits quite regular pollen development, and nearly all
the grains which are pi-oduced are fertile. Other instances of the same
fact will be seen in the list.

A study of the species and varieties of Fuchsia shows quite definitely
that the hybrid character is not the only determining factor in the pro-
duction of sterile pollen grains.

There are a number of other instances already known in the literature
which all point in the same direction. Thus Gates and Goodspeed(7)
have recently described a number of interesting cases of pollen sterilit}^
in plants which are undoubtedly not hydrids. A striking instance of
this is to be found in Scoliopus Bigelovii in which, with all possibility of
crossing excluded and in their native habitat, from 3 per cent, to 32 per
cent, bad pollen is produced, whilst in individual anthers the observed
amount of bad pollen exceeded 45 per cent. "This in itself is a sufficient
refutation of the hypothesis that bad pollen is necessarily a sign of
hybridity." "Pollen sterility is rather a physiological condition which
occurs in all degrees of intensity and may be duo to a variety of causes"
(7, pp. 3 and 4).

Dorsey(4) has also concluded that in the grapes hybridity is not
necessarily the cause of sterility since both sterile and fertile hybrids
occur among cultivated varieties.

Earlier references pointing to the same conclusion may be found
in the literature. Thus Lidforss (9 and 10) found that the hybrid
Rnbus caesiiis ^ x Rubus acmninatus $ had a higher percentage of good
pollen than one of its parents (R. acmnivatus). Hildebrand(8) recorded
the fact that whilst the hybrid Chaniaedorea Ernesti Augasti $ x Ch.
Scliiedeana ^ produced mostly bad pollen, the reciprocal cross gave rise
to individuals in which the pollen appeared normal.

Rudolph Beer 219

These examples might be multiplied but sufficient has been said to
show that, whilst hybridisation maij lead to pollen sterility, this is neither
the sole nor ultimate cause of this phenomenon. Any factor which upsets
the harmonious interaction normally taking place between nucleus and
cytoplasm may lead to the sterility of the pollen grains.

I may mention here several cases which have come to my notice.

During the summer months of 1906 and 1907 I made numerous col-
lections of the flower buds of Trago])ogon pratensis in order to study the
development of the pollen. In all these collections, without exception,
the development of the pollen grains took place in a perfectly normal
manner and four pollen grains were invariably formed from each
mother-cell.

Several plants were, however, grown late in the season of 1907 and
anthers were examined in '6 °/^ NaCl solution upon 20th December.
There had been a cool, damp autumn and the thermometer had on
several occasions sunk to 25" F. It was found that under these conditions
the development of pollen had become irregular and that numerous
supernumerary pollen grains had been formed. In one case nine grains
had been produced from one mother-cell (Fig. 28) and in otheivs <mly two
pollen grains were formed from the mother-cell. In some instances only
a single pollen grain was produced from the mother-cell. The variation
in the size of the pollen grains Avas very great : thus in one instance two
neighbouring pollen grains within one mother-cell measured respectivelj''
28fjL and 12 /a.

Some similar observations which I made upon Oenothera biennis have
already been mentioned by Prof Farmer and Miss Digb3'(5, p. 200) in
a footnote to their paper "On the cytological features exhibited by certain
varietal and hybrid ferns." Here the late season, with cold nights, was
also effective in producing abnormalities in the pollen development of a
plant w-hich usually exhibits perfecth' normal pollen formation.

Another interesting instance of irregular pollen development has
come under ni}' notice. It is that of Geranium ihericum Cav. A plant
of this species has been growing in my garden for many years and during
this time it has constantly failed to develop fertile pollen. I have
examined a number of other plants growing at Catford and at Kew and
in all these cases the stamens never developed any fertile pollen grains.

A study of the pollen development showed that the distribution
of chromosomes was very irregular during meiosis, and that a varjnng
number of pollen grains were initiated in each mother-cell.

Fig. 25 shows a case in which eight cells had been formed from one

220 Cytology and Genetics of the Genus Fuchsia

mother-cell. The later stages of pollen development were very abnormal,
and, whilst the materials for the formation of a pollen membrane were
manufactured by the cytoplasm, these were only rarely laid down as
a coherent wall.

In Fig. 26 two pollen grains are represented at a later stage of
development. Under a very slight pressure upon the cover-glass, the
cytoplasmic body, bi-nucleate in one case and tetra-nucleate in the other,
emerged from the imperfectly developed membrane in the manner shown
in the sketch.

Through the kindness of the authorities at the Herbarium at Kew
I was able to examine several flowers of 0. ihericuvi Cav. which had
been brought from their native habitat in Armenia. In the stamens of
all of these perfectly good pollen had developed.

Whilst working at the John Innes Horticultural Institution I ex-
amined the flowers of a plant of G. ibericum growing in the garden of
the Institution and to my surprise I found that this developed perfectly
fertile pollen grains like those found in flowers from Armenia.

The Merton plant regularly set seed and I raised a number of seed-
lings from it. These I planted in my garden side by side with the sterile
specimen and these have regularly, ^^ear after year, produced fertile
pollen and set tlieir seed, whilst the sterile plant continued to form
stamens containing only abortive pollen. The fertile and sterile plants,
apart from the difference in pollen development, were identical in all
respects both in their vegetative and floral characters.

These observations seem to indicate that a plant, fertile in its own
homes, when brought under the changed conditions of a different country,
may produce offspring some of which are identical with the parent form
whilst others are sterile mutants.

The facts which have been recorded here do not support the view
that irregular meiosis and non-fertile pollen are necessarily an indication
of the hybrid character of the plant producing them. Cases have been
quoted in which undoubted hybrids (e.g. F. puniila x F. alpestris) de-
veloped fertile pollen, and showed no irregularities during the meiotic
divisions. On the other hand it has been shown that plants which we
have every reason to regard as good species, such as Tragopogon -pratensis,
may exhibit all the irregularities in pollen development which have b}^
some been regarded as the attribute of hybrids alone. In a third case,
that of Geranium ibericum, sterile mutants (as regards the pollen grains)
appear to have been developed side by side with offspring which have
remained unaltered under the influence of a changed environment.

Rudolph Beer 221

B. Genetics of Fuchsia.

In this section a number of lines of investigation were commenced
and a brief account of these, so far as they have at present gone, is
given here.

(1) "Faux Iiyhiides" in Fuchsias.

Certain hybrids were described in 1894 by Millardet(13 and 18 a)
between various species of the genus Vitis in which the maternal
characters were represented to the total exclusion of all those derived
from the male parent. To these hybrids Millardet gave the name of
" Faux hybrides." The genus Vitis falls into two sections — Euvitis and
Muscadinia — and it was found that it was only when species belonging
to different sections of the genus were crossed that " faux hybrides "
appeared. Thus when varieties of Vitis vinifera (belonging to the section
Euvitis) were pollinated by Vitis rotundifolia Scuppei^nong (belonging
to the section Muscadinia) offspring were obtained which in all respects
resembled the female parent, Vitis vinifera. If these seedlings were in
their turn again crossed with Vitis rotundifolia Scupper riong there was
still no appearance of any of the rotundifolia characters in the following
generations, but all the seedlings entirely resembled Vitis vinifera.

Gard(6) found in 1903 that the resemblance between the offspring
of these pollinations and their maternal parent extended even to the
anatomical features. Millardet also observed that the pollination of
varieties of Vitis vinifera with the pollen of Ampelopsis hederacea also
yielded " faux hybrides " in which no single character of Ampelojms
occurred.

Another interesting case of the existence of hybrids in which the
characters of one parent are alone represented is presented by the genus
Fragaria. In this instance it is usually the male parent which pre-
dominates although the reverse is said also to occur in other crosses
within the genus (Millardet 13 a).

Solms-Laubach (14) for example found that when Fragaria vir-
giniana was pollinated with pollen from F. elatior the offspring exactly
resembled the male parent {F. elatior) in all their characters. " Faux
hybrides " of apparently a similar nature, but in which the maternal
characters predominate, have been observed in the genus Ruhus by both
Millardet and Lidforss(13a, 9 and 10).

Among the species and varieties of Fuchsia several cases have been
recorded in which the female parent apparently predominates over the
male. Lowe (11) mentions that Fuchsia fidgens crossed by the variety

Journ. of Gen. xi ^^

222 Cytology and Genetics of the Genus Fuckski

" Semiramide •' and the reciprocal cross yielded seedlings which in both
instances resembled the female. Meehan (12) recorded the occurrence
of offspring resulting from the pollination of Fuchsia arhorescens with
pollen of a garden variety of Fuchsia in which the characters were
entirely those of the female parent (F. arhorescens). I commenced a
series of observations upon this subject but these have up to the present
been only partially completed. I will, however, give a brief account of
these observations so far as they have gone.

After due attention to the careful castration of the flowers of a plant
of Fuchsia fulgens these were pollinated with pollen from F. virgata.
Twenty-eight seedlings were obtained and these entirely resembled the
female parent {F. fulgens) and no trace of F. virgata could be detected
in any of them.

Subsequently a further cross was made between Fuchsia fulgens ( $ )
and F. virgata ((/'). These gave seedlings which, with one doubtfu
exception, all resembled F. fulgens in every respect.

These F. fnlgens-Mke offspring were selfed and, of the 50 seedlings
which were grown on, 44 plants flowered and were the exact replica of
F. fulgens both in their flowers and their vegetative characters, whilst
the six remaining plants did not flower before my observations had to be
broken off, but they resembled F. fulgens in all their vegetative characters.
Another cross was made between F. fulgens {%) and the garden variety
" Ballet Girl " (cT ). Twelve seedlings were obtained of which five flowered
before the observations were interrupted. These all resembled F. fulgens
and showed no trace of " Ballet Girl " in any of their characters. In all
cases every precaution had been taken to remove the stamens from the
$ flowers at an early stage of their development and before there was
any possibility of their dehiscence.

The reciprocal cross between F. virgata { % ) and F. fulgens {(^) was
made and in this case the offspring showed chaiacters derived from
both parents. The size and the form of the leaves were intermediate
between those of the two parents. The red pigmentation, which is
characteristic of F. fulgens but absent in F. virgata, was very strongly
marked in some of the hybrids (especially in the plant 3 0) and com-
pletely wanting in others (e.g. 3 5). In other cases it was entirely
limited to the veins, which were bright red, whilst the general tissues
of the leaf were pure green (e.g. 3 F). The flowers were in all cases
closer to F. virgata than F. fulgens in their general form and colour,
but in many instances showed traces of their F. fulgens ancestry in the
slightly increased length of the calyx-tube.

Rudolph Beer 223

The crowded arrangement of the flowers in a terminal inflorescence
was in some cases very like that occurring in F. fulgens and altogether
dissimilar to the loose, paired arrangement of the blooms in F. virgaUi.
This feature was particularly well developed in the hybrid 3 P and the
tact that the individual blooms were all very like those of F. virgata
made the appearance all the more striking. These hybrids showed a
very wide variation in size, some being stronger growing than either
of the parent forms whilst others were very diminutive. PI. XXIV
represents the two extremes in size among these seedlings. The taller
plant measured 130 cm in height, whilst the dwarf specimen was only
8 cm. high.

The late E. J. Allard, Superintendent of the gardens of the John
Innes Horticultural Institution, made a cross between Fuchsia corym-
biflora ( $ ) and Fuchsia hybrid (" Ballet Girl ") (^ ). The flowers of F.
corymhiflora were castrated at an early stage and precautions were
taken to prevent foreign pollen reaching the plant.

A number of seedlings were obtained which resembled the female
parent in all respects and showed no trace whatsoever of the male
parent " Ballet Girl." I made further crosses between these seedlings
and " Ballet Girl " and obtained three fruits in one case and two from
another cross, but unfortunately the work was interrupted before the
seedlings from these fruits could be raised.

Meehan recorded in 1891 the fjict that Fuchsia arhorescens{%)
crossed with pollen of a garden variety of Fuchsia yielded offspring
which solely resembled F. arborescens to the total exclusion of the
characters of the male parent.

I made a large number of crosses between F. arborescens and F. vir-
gata and between F. arborescens and Fuchsia hybrid (" Ballet Girl "),
but none of the fruits set, probably owing to the lateness in the season
when F. arborescens first came into flower that year.

From what has been said above it will be seen that there are
certainly two instances of " faux hybrides " in the Fuchsias, viz., the
cases of F. fulgens and F. corymbiflora. A third instance — F, arbores-
cens— has been recorded, but this requires to be repeated with all the
precautions which we now know to be necessary to prevent foreign
pollination.

The underlying cause of the appearance of these " faux hybrides "
in Fuchsias can only be determined by a careful cytological study of
the individual cases. Several explanations are, however, possible. It may
be that, as in Caelebogyne ilicifolia, Euphorbia duJcis, Nothoscorchnn

224 Cytology and Genetiesi of the Genus Fuchsia

fragrans, etc., fertilisation of the egg-cell takes place and this leads to
the development of an adventitious embryo, formed vegetatively, with-
out the egg-cell itself contributing to the new generation^

Or, as in the case of Opuntia vulgaris, the embryo may arise vegeta-
tively as the direct result of the stimulus given by the penetration of the
pollen tube into the tissues of the stigma and style although fertilisation
of the egg-cell is not necessarily effected at alP.

A third possibility is that fertilisation of the egg-cell takes place
normally, but the female nucleus alone is able to impress itself upon the
cytoplasm of the egg whilst the male nucleus remains completely with-
out influence.

Such a case is known to zoologists through the work of Godlewski
upon the fertilisation of the eggs of Echinus with the sperm of the
Crinoid Antedon. Here he obtained larvae which showed no trace of
Crinoid character".

It is to be hoped that future investigation may discover which of
these possible explanations of the phenomenon applies in the case of
the Fuchsias,

(2) Double Fuchsia, "Ballet Oirl."

A plant of this was selfed and yielded a quantity of seed. From a
sowing of this six seedlings were raised and grown to the flowering stage.
Of these plants all six were identical and the exact likeness of the original
parent. Although I have not been able to ascertain the history of "Ballet
Girl" it cannot be doubted that it is a hybrid and it is rather unexpected
to find no trace of segregation amongst its offspring.

(3) F. globosa (green) x F. gracilis (variegated).

Fuchsia, globosa with green foliage was crossed with Fuchsia gracilis
possessing variegated leaves.

Thirteen seedlings were obtained shownig the following characters: —

8 seedlings with green cotyledons.
3 seedlings with yellow cotyledons.
2 seedlings with white cotyledons.

' See Strasburger, E., "Ueber Polyembryonie." Jen. Zeitschr. f. Naturwiss. Bd. xii.
p. 662, 1878.

- See Ganong, W. F., "On polyembryony and its morphology in Opuntia vulgaris."
Bot. Gaz. Vol. XXV. p. 224, 1898.

* See Godlewski, E., " Untersuchungen iiber die Bastaidierung der Echinoideu und
Crinoiden Familien." Arch. Entwick. mech. Bd. xx. p. 579, 1906.

Rudolph Beer 225

(4) In addition to the above several other crosses were made but an
insufficient number of seedlings had been obtained to make them worth
while describing at any length.

(a) Crosses were made between a single form with njsy-pink petals
and the double Howered "Ballet Girl" with white petals. Offspring pro-
duced flowers which had rosy-pink petals and there were numerous
grades among them between completely single flowers and completely
double ones.

(6) Fuchsia procumbens (^)x F. refiexa ((/ ). Only a single seedling
was produced from this cross. The upper branches of the plant were
erect like those of F. refiexa whilst the lower ones were trailing like
those of the female parent. The flowers were small, with pale pink
sepals and a corolla reduced to a small deep pink scale. The stamens
were very rudimentary and completely sterile. The pistil possessed a
four-lobed stigma ; the style being pink and the stigma a lighter shade
of the same colour.

(c) A number of crosses were made between F. Gottinghami ( $ ) and
F. refiexa {(^); and between F. parvifiora(^ ) and F. refiexa {^). In
both cases the female parent possessed completely sterile anthers whilst
the male parent {F. refiexa) produced good, fertile pollen. The object of
the cross was to test the inheritance of the fertile and sterile character
of the pollen in the two parents.

Results were not carried far enough to deal with the point efficiently
but one seedling was obtained from crossing F. Cottinghami with F. refiexa
in which the pollen was perfectly fertile and resembled that of F. refiexa
in possessing two interstitial bodies.

Further crosses were made between this seedling (No. 7)((/) and
F. Cottinghami { $ ). In the offspring of this cross either all the stamens
were sterile or in some cases (e.g. No. f^) a certain amount of fertile
pollen with two interstitial bodies was produced amongst a quantity of
grains which were empty and bad.

The work for this paper was carried out at the John Innes Horti-
cultural Institution and I wish to express here my most sincere thanks
to Mr Bateson for the facilities which were attorded me aiid for the
interest he has taken in the work.

226 Cytologn and Genetics of the Genus Fuchsia

EXPLANATION OF PLATES.

For Figs. 18, 25, 26, 27, and 28, an one-eighth inch objective and No. 2 compensating
eye-piece were used; for all other Figs, a Zeiss 2 mm. lens of 1-40 N.A. and an 18 com-
pensating eye-piece were employed.

PLATE XXII.

Figs. 1 — 13. Fuchsia hybrid. Stages of first meiotic division in pollen-mother-cells.
X 2700.

Figs. 14 and 17. Fuchsia hybrid. Stages of second meiotic division in pollen-mother-
cells. X 2700.

PLATE XXIII.

Figs. 15 and 16. Fuchsia hybrid. Stages of second meiotic division in pollen-mother-
cells. X 2700.

Fig. 18. Fuchsia hybrid. Ten young pollen grains all derived from one pollen-mother-cell.

Fig. 19. Fuchsia hybrid. Supernumerary cells all derived from one pollen-mother-cell.
X 2700.

Figs. 20 and 21. Fuchsia rjlobosa. Telophase of first meiotic division of pollen-mother-
cells. X 2700.

Fig. 22. Fuchsia ylobosa. Second meiotic division ; early anaphase, x 2700.

Figs. 23 and 24. Fuchsia corallina. First meiotic division of pollen-mother-cells, x 2700.

Fig. 25. Geranium ibericum. Eight young pollen grains all derived from one pollen-
mother-cell.

Fig. 26. Geranium ibericum. Binucleate and tetranucleate protoplast escaping from in-
complete wall on slight pressure.

Fig. 27. Fuchsia hybrid. Pollen grain with only a single interstitial body.

Fig. 28. Trayopoyon pratensis. Nine pollen grains all derived from one pollen-mother-cell
as a result of low temperature.

PLATE XXIV.

Fuchsia virgata x F. fulgens. Two plants (3 R and 8 T) of Fj.

LITERATURE CITED.

1. Beer, R. "The Supernumerary Pollen -grains of Fuchsia." Ann. But. Yol. xxi.

p. 305, 1907.

2. Blackburn, K. B. and Harrison, J. W. "The Status of the British Rose

Forms as determined by their cytological Behaviour." A7in. Bot. Vol. xxxv.
p. 159, 1921.
.3. Bonnet, J. "Recherches sur revolution des cellules-noun-icieres du pollen chez
les Angiospermes." Archiv f. Zellforschimg, Bd. vii. 1912.

4. DoRSEY, ]\I. J. " Pollen Development in the Grape with special reference to

Sterility." University of Minnesota Agric. Expt. Stat. Bull. 144, 1914.

5. FARMEit, J. B. and Digby, L. " On the cytological Features exhibited by certain

varietal and hybrid Ferns." Ann. Bot. Vol. xxiv. p. 200, 1910.

6. Gard, M. "Etudes anatomiques sur les Vignes et leurs hybrides artificiels."

Act. Soc. Linn. Bordeaux., Ser. 6, Tom. viii. pp. 185 — 319, 1903.

7. Gates, R. R. and Goodspeed, T. H. " Pollen Sterility in relation to Crossing."

Science, N. S. Vol. xliii. pp. 859—861, 1916.

8. Hildebrand, Fr. " Ueber einige Pflanzenbastardierungen." Jen. Zeitsch. /.

Xaturwiss. Bd. xxiii. pp. 413- 548, 1889.

JOURNAL OF GENETICS, VOL XI. NO. 3

PLATE XXII

.^'

^

^?^ \ c^

10

tNUFf

«

13

12

I

"' 4 I

14

4

:^^-

JOURNAL OF GENETICS, VOL XI. NO. 3

PLATE XXII

c:

>

24

^■^b

28

\

23

4'

t

T'

JOURNAL OF GENETICS, VOL XI, NO. 3

PLATE XXIV

Fuchsia virgata x F. fulgens. Two plants (37? and 3 T) of Fi.

Rudolph Beer 227

!J. LiDFORSS, B. "Studier ofver artbildningen inom sliiktct Rulni.s." Arkio f. Hot.
13d. IV. p. 41, 1905.

10. . "Studior ofver artljildniiigon inom sliiktct Riibu-s." A, -hi,' Dot. Bd vi.

1907.

11. LowK, E. J. " Rep. Sci. Committee R. Hoi-t. Soo." Gard. Chron. 3 Ser. Vol •vxii.

p. 538, 1890.

12. Meehan, T. "On the varying Characters of Hybrids." (iard. Chron. 3 .Scr.

Vol. X. p. 109, 1891.

13. MiLLARDET, A. "Notc .sur la fausse hybridation chcz Ics AmpeHdces." Rev. Vitic.

Tom. XVI. pp. 677—680, 1901.
13a. — — . "Note sur I'hybridation sans croi.sement on faux hybi-idatiou." Mem.
SoG. Sci. phys. nat. Bordeaux, 4 Ser. Tom. iv. pp. 347 — 372, 1894.

14. Solms-Laubach, H. Graf zu. "Ueberunsere Erdbeeren und ihrc {Jeschichte."

Bot. Zeit. Bd. Lxv. 1 Abt. pp. 45—76, 1902.

15. TisCHLER, G. "Zellstudieu an Sterilen Bastardpflanzen." Arch. Zellforsch.

Bd. I. pp. 33-151, 1908.

16. WiLLE, N. " Uebcr die Entwickelungsgeschichte der PoUenkorner der Aiigio-

spermeu." Christiania, 1886.

17. WiMMEL, T. "Zur Entwickelungsgeschichte dcs Pollens." Bot. Zeit. Bd. viii.

1850.

LINKAGE IN GAMMARUS CHEVREUXL

By J. S. HUXLEY,

Neiij CuUege, Oa-ford.

The work of Allen and Sexton ('17) shcjwed that Mendelian inherit-
ance occurred in Gavirnaras chevreaxi. Three mutations in eye-colour
appeared and proved to be due to recessive changes in three Mendelian
genes. These factors may be designated respectively B (black) and h (red);
(J (colour) and c (absence of colour) ; and W (white) and w (no-white).
However, some of the ratios obtained from crosses where two or three
factor-pairs were involved departed considerably from expectation. It
seemed that they might be due to linkage of factors B and C. This
hypothesis could not be tested from the data in Allen and Sexton's
paper, since the make-up of the heterozygotes involved was not recorded.
Accordingly it was decided to test for linkage by breeding. Through
the kindness of Mrs Sexton, I was provided with a number of almost

(B G W\
' ' j and

' ' ^] with triple recessives

I — ^ — ) . Unfortunately, only a very few triple recessives were available,

so that the more desirable back-cross could not be undertaken at once.
A number of broods, however, have been raised from the triple hetero-
zygotes mated mter se, and figures obtained which warrant the assertion
that linkage exists between B and C. I have therefore thought it best
to publish at once the data on which this assertion is based, while con-
tinuing and amplifying the experiments in various directions.

If no linkage exists among the factors and no differential viability
among the classes of offspring, the mating I have mentioned should give
the following classes and proportions of offspring in F^ :

Black, B, C,W 27

Red, h, C, W 9

Albino, JS (6), f, W 12

Black, No- white, B, C, iv ... 9

Red, No- white, b, C, w ... 3

Colourless, B (b), c, w ... 4

Total , 64

230 Linkage in Grammarus chevreiixi

There appears from Allen and Sexton's work to be very little diifer-
ence in viability between the different classes. The recessives are perhaps
a trifle less hardy, but the differences from expectation thus produced in
the ratios are negligible. As we shall see, this factor can probably be
disregarded in the experiments here recorded, since the double recessive
classes, which should be the least viable of all classes, are not below
expectation.

There is no possibility of confusing any classes save blacks and reds
{BOW and hCW). Occasionally the heterozygous blacks show a distinctly
russet or dark crimson tint. This, however, is in almost all cases wholly
distinct from the light scarlet of the reds. Very rarely specimens are
met with among the coloured-with-white class which are apparently in-
termediate ; whenever this is so the fact has been noted. The doubtful
cases almost all cropped up at the outset, and disappeared with practice.
In the coloured no-white classes {B, G, lu and b, G, w), however, I have
never met with heterozygotes which could not be at once classed with-
out hesitation; the colour stands out much more clearly when not against
the chalky background of the white pigment. If there is linkage between
B and G (and only between B and G) the ratio of black-with-white
(B, G, W) to red-with-white {b. G, W) should be the same as that of black
no-white (B, G, w) to red no-white {b, G, lu). Since the former ratio is
actually found to be very nearly equal to the latter, and since in this
latter ratio there is no possible source of error from confusion of classes,
we may assume that any error thus produced in the former ratio is
negligible, an assumption further justified by the rarity of doubtful
cases.

Some broods were kept at room -temperature (10''-15°C.), others in
an incubator at 25°-26° C. Otherwise the treatment was the same for
all. The figures indicate that linkage (as in Drosophila — see Plough, '17)
is less intense at the higher temperature. There are further some slightly
aberrant ratios in the no-whit& (ww) classes. With these subjects I do
not propose to deal until I have further data ; here I shall only discuss
the question of linkage between B and G.

If there is linkage between B and G of such strength that p repre-
sents the linkage and 1 -p the crossover- value in the female, while q
and 1 — 5 are the corresponding values in the male, we should expect
the following ratios (see Haldane, '19):

Black -I- Black No-white {B, G, W -{- B, G, tv] 2 + i^g

: Red -I- Red No-white [b, G, W+b, G, w] 1 - P*/

: Albino -f Colourless [B (b), c, W + B (b), c, w]...l.

J. S. HrrxLEY 231

The following table gives the data. All broods in which there was
any doubt over the eye-colour of any individual have been excluded from
the main table, but are given at the close for the sake of completeness.

TABLE I.

Red

No-wliite Albino Colourk-ss

Category n, C, W B,C,v> b,C,W b, C, w B {b), <; W li(h),<:,v; ToUls

At room temperature (74
broods)

First broods after trans-
ference from room to
25° C. (15 broods) 149 29 62 210

C. Subsequent broods at
25° C. (41 broods)

I). Total of A , B, and C (133
broods)

232

59

291

42

20

62

115

50

165

389

129

518

426

137

1129

642

2011

E. Total of (£>) with all
doubtful broods added
(13 more broods) 1293 353 563 2209

F. Expectation for (/)) if no 848-4 282-8 •282-8 94-3 377-1 125-7
linkage present ' . ' ' r—-^ ^ , •'

1131-2 377-1 502-8 •2011-1

It will -be seen that, while Albino + Colourless are close to expecta-
tion, the ratio of the W to the w classes, as I mentioned above, shows
some peculiarities, the classes containing imu usually appearing slightly
in excess, especially at the higher temperature. The reason for this is
not yet apparent ; but whatever the cause, the deviations do not support
the idea of linkage of W with the other factors.

When we come to the ratio of blacks to reds, however, we find an
excess of blacks and a deficiency of reds, both in the with-white and no-
white classes. The ratios of black to red previously obtained by Allen
and Sexton {loc. cit, p. 343), when only the factor-pair B, b was involved,
conformed very closely to the expected 3:1. Differential mortality may*
also for this reason be excluded. Since in my stock B came in with 0,
and b with c, we should expect to find this excess of blacks and deficiency
of reds (a con-esponding excess of reds and deficiency of blacks being
carried by the albino and colourless classes).

Assuming that there is linkage between B and C, then let r be the
geometric mean of the linkage values p and q. We then find that the
value of r, as calculated from the totals in (D), is given as follows :

2 +pq 1190 „ , . , , — ^g_.

, -- = -r:^ irom which r — \ pa = bZoo ;
I — pq 303

and 100 (1 — r) . (= mean crossover-value) = 37"45 + 0'69 7c-

This gives a gametic ratio of almost exactly oBC : SBc : 'ibC : obc.

232 Linkage in Gammarus chevreuxi

The figures for A (broods at room temperature only) are :
r = '6397 ; mean crossover- value = 36'03 + 0'89 °/^.

Those for G (broods at 25° C. only) are :

r = "5608 ; mean crossover- value = 43'92 + 1"28.

It will be seen that the difference between the crossover-values of A
and C is 7*89%. This is more than 6 times the larger probable error,
and is therefore significant.

Group B represents the total of the broods first produced after trans-
ference from room temperature to 25° C. The eggs were already fertilized
and in the pouch before transference, so that the high linkage value

r = "7149 ; mean crossover- value = 28*51 + 1*49 °/.^,
must be regarded as an accident.

The few animals obtained from the back-cross -,' — '- x ,'^ —

0, C, 10 0, c, w

corroborate these findings. They are as follows :

B, C, W B, C, w h, C, W h, G. w B (b), c, W B{b), r, w Total

(6 broods) 13 14 2 10 32 27

27 12 49 88

Expectation p : 1 — p '• 1

These show the same deficiency of both classes of reds. (They also
show the excess of individuals containing ww.)

If we neglect the fact that the c, W and c, w classes are not exactly
half the total, we find that^ = f^ = '692, with crossover-value SO'S ± 4-87^,
which, with the few broods available, is a sufficiently close approximation.
It should in this connection be mentioned that different pairs vary con-
siderably in the strength of linkage which they exhibit. It should be
remarked that since the crossover-values for the back-cross are of the
*same order as those for the heterozygotes mated inter se, linkage must
be of about the same intensity in the two sexes. In this Gammarus
differs from Insects, and resembles Mammals.

The abnormal ratios in Allen and Sexton's paper may all be explained
on the assumption of linkage of an intensity close to that here found.
E.g. their No. 6, p. 344, 2nd mating (linkage is not involved in the first
mating of this heading).

If the composition were -j^ — x j^ — , then we should expect the ratio

s, black : 1 — s, red : 1 , albino, where s = either p or q.

The figures were BG, 31 ; 60, 8 ; B{h) c, 34; which gives s = '7949,
with c.o.v. = 20-51°/ .

J. S. Huxley 233

B C b C
No. 7, p. 345. If the cross were y — x j — , we should expect

1 + .9, black : 2 — s, red : 1, albino.

The figures were BG, 235; 66', 169; B (b) c, 144; which gives
.9 = -7450, with CO. V. = 25-5 %.

b C b C
(d), p. 347. If the cross were ^— x y^ , we should expect

2 — s (black + black no-white) : 1 + .9 (red + red no-white)

: 1 (albino -f- colourless).

The figures were B, C, 162; b, C, 202; B (b), c, 127; which gives
.9= -6646, with c.o.v. = 33-54 %.

The linkage-values for the two latter sets of crosses are in close
agreement with my figures; those for the first cross are a sufficient
approximation in view of the small numbers.

The double heterozygotes (No. 4, p. 344) and triple heterozygotes
[(e), p. 347] crossed inter se do not show linkage ratios, presumably
because the make-up was not always the same, B having sometimes
entered with C, at other times with c.

The work is being continued. But the fact that this is, I believe,
the first case of linkage recorded for Crustacea wan-ants its publication
at the present stage.

In conclusion, I would like to thank Mrs Sexton of the M. B. A.,
Plymouth, for her kindness in providing me with material, and in placing
her experience of the species at my disposal, my wife for help in caring
for the animals, and Mr J. B. S. Haldane of New College, Oxford, for
criticism and advice. The abnormal ratios in Allen and Sexton's paper
had, I found, led Mr Haldane independently of myself to suppose that
linkage might be present, but he was unable to undertake the experi-
mental proof of the hypothesis.

LITERATURE.

Allen and Sexton, '17. "The loss of the ej'e-pigraent in Gammarus Chevreu.vi."

J. Mar. Biol. Ass. Plyiiiouth,\o\. xi. 1917, 273.
Haldane, J. B. S., '19. " The probable errors of calculated linkage values, &c."

Jour a. Genetics, Vol. vilL 1919, 291.
Plough, H. H., '17. "The effect of temperature on crossing over." J. E.vp. ZooL

Vol. XXIV. 1917.

GENETIC STUDIES IN POULTRY.

IV. ON THE BARRED PLUMAGE OF CERTAIN BREEDS.

By R. C. PUNNETT, F.R.S., and M. S. PEASE, M.A.

(With Two Text-figures.)

The type of plumage known as barred, in which bands of different
colour are arranged alternately along the feather at right angles to
its main axis (cf Fig. 1, p. 239), is characteristic of several breeds of
poultry. It is to be found in its most marked form among Plymouth
Rocks, Pencilled Hamburghs and Campines. Probably also the feather
patterns designated as laced, pencilled\ and .spangled are intimately
connected with the barred. We have long felt that a genetic analysis
of barring in Poultry, with its peculiar rhythmical alternation of pig-
ment deposition in the feather, might help to throw some light upon
the problems of growth, and with this end in view we have recently
commenced a series of breeding experiments. Though the work is as
yet but in its initial stages, the present season has yielded a result of
sufficient interest to place on record. Before doing so however we would
dwell briefly upon what is known as to the genetic behaviour of barring
in poultry.

In the present state of our knowledge it would appear that the
barring found in Plymouth Rocks, Scots Greys and Dumpies is geneti-
cally distinct from that found in the barred breeds of Hamburghs, and
in Campines. In the Rock group the birds are of the same fundamental
colour throughout, and the barring effect is produced through full and
faint development of this fundamental colour in alternate bands. In
the Barred Rock itself, for example, the darker bands appear as grey-
black, while in the lighter bands the grey colour is very faint, so that
these bands appear almost, though not quite white.

1 The use of this term in the Fancy is apt to confuse the uninstructed. The "Pencilled"
Hamburgh is a barred bird. We propose, for purposes of this investigation, to restrict the
term "pencilled" to the characteristic pattern of the feathers in the " Partridge " varieties
of Cochins, Wyandottes, etc. The connection between pencilled and barred is a close one,
for feathers that would pass for barred, as well as the distinctive pencilled feathers, and
intermediate types, are to be found on the same bird.

236 Genetic Studies in Poultry

Thanks to the work of Spillman, Morgan, Pearl and others in America
we realise that the barring depends upon a definite sex-linked factor
acting upon a black basis. The factor would appear to be of an inhibitory
nature, preventing the deposition of black pigment almost completely
in the lighter bands of the feathers, slightly so in the darker ones. The
action of the barring factor is also evident uj)on a brown basis, though
not so clearly. Some years ago Ave made a cross between Plymouth
Rock % and Brown Leghorn ^ . The ^i (/</ were barred like the Rock,
while the F^ %% were full black or nearly so. In the F^ generation
appeared more or less typical barreds and blacks of both sexes, together
with browns of various kinds (including silver-greys^), both barred and
unbarred. In these barred browns the barring was not nearly so sharp
as in the barreds belonging to the black class. For the present we are
disposed to consider that the barring factor acts more intensely upon
the black pigment than upon the golden pigment, if indeed it acts upon
the latter at all. On this view the blurred nature of the barred browns
would be due to the inhibition of only the black pigment scattered
about among the gold in the mixture which goes to make up the brown
bird. We realise that this can be no more than a plausible suggestion in
our present state of ignorance of the pigments of the fowl's plumage.
The point however deserves further investigation.

As shewn by its behaviour towards black, the barring of the Ham-
burgh-Campine group is evidently different in its nature, though the
only direct evidence we know of is provided by some experiments carried
out a few years ago at Cambridge^. In those experiments a cross was
made between a Gold Pencilled Hamburgh ^ and a Black Langshan %.
The ^1 birds were black (with some gold in the hackles of the cocks),
and in F.^ some barred birds reappeai-ed. The barring of the Hamburgh
is evidently recessive to self colour in this case. But it does not neces-
sarily disappear completely on crossing. This year we mated a Gold
Campine $ with a Barred Rock ^ which was heterozygous in baning.
The cross resulted in approximately equal numbers of barred birds very
like Plymouth Rocks in plumage, and of birds wholly or predominantly
black. Certain of these last however shewed coarse silver barring in some
feathers, notably on the breast. This barring we suppose to have been

' Evidently the silver factor was carried by the Pl3'mouth Rock ,^ . Of some birds we
bred this year, ex Gold Campine $ x Plymouth Eock ^ (heterozygous in barring), the
non-barred ones which were not full black all showed some silver markings. Here again
the Rock evidently carried silver. It seems not improbable that the silver factor may be
an ingredient essential to the barred pattern of the Rock Group.

'^ Cf. Punnett and Bailey, Journal of Genetics, 1918 and 1920.

K. C. PuNNETT AND M. S. Pease 237

derived from the Campine, and it suggests that barring is not always
completely recessive to self colour. We should add that the data we
have accumulated afford no evidence for supposing that the Hamburgh-
Campine barring is in any way sex-linked.

We must suppose therefore that there are at any rate two forms of
barring in poultry, of which one is sex-linked and dominant to self
colour, while the other is more or less recessive to self colour, and does
not exhibit sex-linkage. The position appeared to us so curious that
we have started a series of experiments to try and determine whether
there is any relation between these two groups of barring, and if so,
what that relation is. For, on the face of it, it seems unlikely that a
peculiar and rhythmical pattern such as barring should be brought
about in two entirely independent ways. It is obvious that the most
direct way of attacking the problem is to cross birds of the Rock type
with those of the Hamburgh-Campine type, and to caiTy out an ordinary
genetic analysis. This we have already commenced, but it must be
several years before we can hope to complete our analysis. At the same
time we have been investigating the inter-relation of the different forms
of barring in the Campine itself, and it is with a result obtained here
that the present note is concerned.

The barred breeds of both Hamburghs and Campines are well known
in their gold and silver varieties, and a description of these will be found
in any work dealing with the breeds of poultry. In the former the alter-
nating bars are black and gold; in the latter black and white. Recently,
through the enterprise of the Rev. E. Lewis Jones, the recognised autho-
rity on Campines, a third barred form has been introduced. This is the
"Chamois" Campine, in which the alternating bars are white and gold^
The relation between the Gold and the Silver Campine is that which
exists between the corresponding gold and silver forms of other varieties.
Silver is dominant to gold, and is sex-linked. This we proved some
years ago for the gold and silver-pencilled Hamburghs, and our more
recent experiments shew it to be true also for the Campines. The white
bars in the Silver Campine correspond with the gold bars in the Gold
Campine, and the black bars in the one to the black bars in the other.
The point then arises as to the correspondence of the white and gold
bars of the Chamois with the black and gold of the gold, and the black
and white of the silver. In order to decide this we made the following
crosses during the past season.

^ Through the kindness of Mr Lewis Jones we obtained a few of these birds for experi-
ment, and for this we wish to express our sincere thanks.

Journ. of Gen. xi X6

238 Genetic Studies in Poultry

(1) Chamois (/ x Brown Leghorn $. The seven chicks hatched from
this mating were all light in down with a slight huffish tinge. The
result accords with the view that Chamois acts as a dominant white,
and that the cock used was homozygous in this respect. Of the seven
chicks one only was reared, and proved to be a white cockerel with a good
deal of gold irregularly distributed over its plumage. Here and there
slight indications of barring were present.

(2) Chamois % x Gold Campine cf. Of 21 chicks produced 14 proved
to be chamois, and the remaining seven gold barred. Both sexes appeared
in each colour class. From this, and from other evidence, we regard
chamois as dominant to gold barred, the chamois $ used here being
heterozygous.

(3) Chamois $ x Silver Campine (/. In this experiment the same
chamois ^ was used as in (2). As to be expected she gave chicks with
light and with coloured downs in approximately equal numbers, viz.
17 light and 15 dark. Of the chicks with dark down those reared (1 </
and 3 $ $ ) all developed into silver barred birds. Four also of the chicks
with light down were reared. All proved to be males, and all developed
into white birds.

(4) Chamois cf x Silver Campine $. The cock here was the same
bird as that used in (1), and it was to be expected therefore that all of
the chicks would be light in the down. Such was the case. But although
the downs were all light they fell into two distinct classes, viz. those
which were light with a slight huffish tinge, and those which were
mottled with buff. Of the 24 chicks hatched 12 belonged to each class.
Three of those with buff mottled down^ were reared and turned out to
be all pullets, while the eight with light down reared all developed into
white cockerels.

These last two experiments shew that it is possible to produce white
birds from a cross between two coloured varieties, each of which breeds
true to its own colour. Corresponding feathers from a Silver Campine (a),
a Chamois (c), and a white (6) resulting from a cross between the two are
shewn in Fig. 1.

We offer the following interpretation. The silver differs from the
gold in the possession of a factor which inhibits the production of the
gold pigment without affecting the black. This factor is sex-linked in
its mode of transmission. The Chamois differs from the gold in the
possession of a factor which inhibits the production of the dark melanic
pigment, but appears to have little or no influence on the gold pigment.
^ This is the down characteristic of the Chamois Campine.

R. C. PUNNETT AND M. S. PeASE

•239

When both of these inhibitors are present they lead to the formation of
a white bird. Such a condition arises when a silver ^ is mated with a
chamois % . Assuming both parents to be homozygous it is clear that
all of the offspring would be heterozygous for each of the inhibitors,
and would consequently all be white. We have not actually realised a
cross in which all the offspring were white. In Exp. (3) about half of
the offspring were coloured, but this, as we have already pointed out,
we regard as due to the accident of our chamois $ having been hetero-
zygous for the inhibitor of melanic pigment. The chamois </ used in
Exp. (4) was however homozygous, and all of the progeny lacked melanic

a b c

Fig. 1.

pigment except for occasional ticks. But owing to the fact that the
silver $ is heterozygous for the gold inhibitor the <f(f alone from this
cross are white, the pullets being chamois.

Here we may mention a point of some interest which has arisen in
connection with our whites. Though at first fully white, except for a
few dark ticks, the}^ shew indications of very faint barring as they gi'ow
older. We have observed this in all of the whites we have bred — which,
of course, were all heterozygous — though the barring is rather more
distinct in some than in others. Mr Lewis Jones tells us that he has
also observed this " ghost " barring in birds similarly bred by him, and
he kindly sent us several to compare with our own. Whether the "ghost"
barring is peculiar to the later feathers, or whether the feather is at
first white and subsequently developes the barring on exposure to the

16-3

240

Genetic Studies in Poultry

air, is a point we have not yet determined. We are inclined to attribute
the appearance of the ghost barring to the bird's being heterozygous
for the inhibitor of melanic pigment. It is certain that the ghost bars
occupy the position of the black bars in a gold or a silver barred bird.
For a white bird sometimes puts up a feather which is silver barred,
and occasionally the silver barring is confined to one side of the rachis.
In such cases the ghost bars are continuous with the black bars across
the feather.

Whether the ghost barring is a heterozygous character we hope to
determine next season by breeding from these heterozygous whites, and
in order to make our view clear we may indicate what we expect to be
the composition of the F., generation. Denoting the factor for the in-
hibitor of gold pigment by S, and that for the inhibitor of melanic
pigment by X, we suppose the genetic constitution of the F^ birds to
be SsXx. The gametes of the F^ (/ should therefore be SX, Sx, sX, sx.
Since S is sex-linked the female-producing gametes of the F^ % will be
sX and sx, while the male-producing will be SX and Sx. The nature
of the F2. generation to be expected is indicated in Fig. 2.

males

sx

SX

SX

SX

sx

Sx

sX

sx

Full white

Ghost-barred

Full white

Ghost-barred

Sx

Sx

Sx

Sx

sx

Sx

sX

sx

Ghost-barred

Silver-barred

Ghost-barred

Silver-barred

sX

sX

sX

sX

SX

Sx

sX

sx

Full white

Ghost-barred

Chamois

Chamois

sx

sx

sx

sx

SX

Sx

sX

sx

Ghost-barred

Silver-barred

Chamois

Gold-barred

females

Fig. 2.
Five different kinds of birds are to be looked for on this hypothesis,
though, as shewn in the appended table, the distribution of these be-
tween the sexes should be very different.

Males

Females

Totals

Full white ...

2

1

3

Ghost-barred

4

2

6

Chamois

—

3

3

Silver-barred

2

1

3

Gold-barred ...

—

1

1

Further work however must shew whether this is the correct inter-
pretation.

A STUDY OF THE VARIATION IN SEEDLINGS
OF THE WILD HOP {HUMULUS LUPULUS L.).

By E. S. SALMON and H. WORMALD.

{Research Department, S.E. Agnc. College, Wye, Kent.)

(With Plate XXV.)

In 1913 it was decided to investigate whether seedlings of the wild
hop showed any resistance to the attack of the mildew Sphaerotheca
Humuli (DC.) Burr., and if so, whether this resistance was correlated
with any morphological character.

It was necessary therefore to obtain seed of the wild hop. H. Lupuhis
is indigenous in Southern, Northern and Mid-Europe and in Northern
and Mid- Asia \ Hooker (Students' Flora) and Babington {Mamial of
British Botany) describe the plant as a true native in the South of
England. The so-called " wild hop," however, found growing commonly
in hedges in counties such as Kent and Sussex where commercial hop-
growing is practised,is either an "escape" and belongs to some commercial
cultivated variety, or possibly may be a " cross " between the wild hop
and some cultivated variety. Similarly in those countries on the Con-
tinent where hops are cultivated, e.g. Germany, France, Russia, etc., the
same uncertainty may exist as to the true status of a hop growing
apparently wild^ At the commencement of our investigations we be-
lieved it to be a fact that the hop had never been cultivated in Italy,
and we procured, through the courtesy of Prof. P. A. Saccardo, seed of
the wild hop, collected by him and labelled " Vittorio (Treviso), ad sepes,
omnino sponte, Oct. 1913." Subsequently a doubt arose in our minds as
to the certainty of this seed having originated solely from the wild plant.

1 The wild hop of North America is, we consider, a distinct species, viz. H. Americanus
Nutt. (1)*.

2 In an article by J. Schmidt entitled " The occurrence of the wild hop in Denmark "
(C. r. trav. Labor. Carlsberg, Vol. xi. 314 — 329 (1917)) the author states : By " wild " is
here understood ' ' not cultivated." The forms studied by Schmidt cannot be regarded
therefore as the undoubted offspring of the truly wild species.

* These numbers refer to the bibliography given at p. 267.

242 Variation in Wild Hop

In 1915 we were informed by Dr M. Corvi, of the R. Istituto Superiore
Agrario of Perugia, that in the years 1860-70 the hop was cultivated
near Bologna "with good results," and that experiments with its culture
were being conducted in Umbria. Further, we found that although all
the authors of Italian " Floras " give the hop the status of a wild plant,
one of them, Parlatore {Flora Italiana, iv. 303 (1867)) remarks: "II
Luppolo e coltivato in alcune parti d' Italia, come a Forli ed altrove.
Tali coltivazioni sono pero ancora molto ristrette tra noi, facendosi
venire il Luppolo stesso dalla Germania." The following information,
however, supplied by Prof P. A. Saccardo in April 1916, makes it
practically certain that the parent plant of the seedlings described below
was the wild species. " Apres investigations chez mes collegues Mittirolo,
Peglion, Beguinot, etc. je peux vous assurer, Vhouhlon n'a ete jamais
cultive dans la province de Treviso d'ou je vous I'ai expddie et ou il se
trouve sauvage partout abondamment, surtout sur les haies. M. le prof.
Peglion qui va publier dans ce moment un Monografia del luppolo in
Italia m'ecrit de Bologna que en toute Italie on cultive : 3 hectares de
houblon dans la plaine d'Orvieto et dans I'haute-plaine de Alfina
(Ombria), | hectare a Pedavena (Feltre). Comme vous le voyez vous
pouvez etre sur, tres sur, que les exemples que je vous ai expedies
sont absolument spontanes et sauvages."

The seedling plants have been raised and grown in the following
manner. The bulk of the seed from Italy was sown in seed-boxes in the
greenhouse in 1914 and 1915. Some of the seed germinated the same
year it was sown and some in the second and even the third year after
sowing. Portions of the same seed (collected in 1913) were retained and
sown in 1916 and also in 1917 ; germination was poor and only a few
seedlings were raised from these later sowings. As soon as each seedling
was old enough to transplant, it was placed in a pot and kept in the
(unheated and well-ventilated) greenhouse during its first year and
sometimes during its second year. The 1- and 2-year old seedlings
reached a height of from three to five feet. During the winter they
were planted out, in blocks of 50 or more, distributed through the
Experimental Hop-garden of 2f acres at Wye College, Kent. The
seedlings were planted 3 ft. 6 in. apart in the row, and the rows were
6 ft. 6 in. apart. The plants have been given the treatment usual in a
commercial hop-garden; the underground rhizome is "cut" or "dressed",
each winter — an operation which removes the living, swollen bases
(" strap cuts ") of the annual shoots, so that the next year's shoots arise
from buds lower down on the rhizome ; each spring the superfluous

E. S. Salmon and H. Wormald 243

stems (" bines ") are pulled off from the rootstock, and the remaining
2 — 3 stems' are " trained " up the " strings " (of coir yarn), for the first
4J feet vertically to a lower wire, then at an angle of approximately 45°
to a top wire 11 ft. 6 in. from the ground supported on poles-. The soil
was uniformly and fairly heavily manured and good cultivation was
given to the whole hop-garden.

The hop plant, Humidus Lupulus L. is a dioecious perennial with a
persistent underground rootstock from which develop long annual climb-
ing shoots. The main axis of a shoot (or " bine ") bears decussate leaves
from the axils of which develop the primary bi'anches (" laterals "). The
lower laterals are usually barren but the rest form much-branched
paniculate inflorescences. The rootstock of the seedling plant during its
first year is very small and the shoots comparatively weak and the plant
does not attain to full vigour for several years. Observations regarding
morphological characters were therefore not taken until the plants were
in their third year and even then some characters, particularly those
depending upon vigour, e.g. length of the laterals, could not be regarded
as constant.

As the plants approached an age when the various organs assumed
the characters shown by the mature plant it was seen that there were
marked differences in certain morphological and physiological characters.
That some at least of these differences were not mere fluctuations due
to environmental conditions was shown by keeping the plants under
observation for several successive years and noting which characters
remained constant (or fluctuated within comparatively narrow limits)
for the individual plants.

The plants were first critically examined in 1916 and the characters
noted of those plants which were old enough for the purposed Many of
the plants have thus been under observation during five seasons, the
rest during from two to four seasons.

The variations which have been observed are concerned with the
following points :

1 In a few cases the seedlings were planted 6 feet 6 inches apart in the row, and 6 stems
trained up the strings.

- The system of wire work used in the Hop-garden is that known as the "Butcher." It
has been described by A. Amos in the Journ. Board Agric. YoL xvi. p. 890 (1910), where a
detailed account is given of the English method of cultivating hops.

3 The variation shown by a number of male hop plants growing in the Experimental
Hop-garden at Wye College and known, or believed, to be derived from seeds of cultivated
hops, has already been recorded (2). The methods of observation and description adopted
in that instance have been applied (modified and extended where necessary) in describing
the plants noted in the present paper.

244 Variation in Wild Hop

(1 ) Time of flowering,

(2) Colour of the "bine."

(3) Relative number of glands on the leaves.

(4) Shape of the leaves.

(5) Colour of the leaves.

(6) Length of the primary branches (laterals).

(7) Characters associated with the reproductive organs, i.e. (a) in
the cT plant, the number of glands on the perianth-lobes, on
the anthers and on the receptacle ; (6) in the % plant, the
characters of the strobile (hop-" cone"), viz. size, shape, aroma,
"condition," and colour of the main axis ("strig").

(8) Degree of susceptibility to the mildew Sphaerotheca Humuli,
under greenhouse conditions and also in the open.

(1) Time of Flowering.

The great majority of the plants were much later in coming into
flower than the cultivated varieties. It was found however that they
showed considerable variation amongst themselves with respect to their
time of flowering and particularly was this noticeable in the case of the
cf plants; the earliest of these, as a rule, came into flower four weeks or
more before the latest and this difference is maintained from year to year.
By the " time of flowering " is here meant the first day on which flowers
actually opened to show the stamens, in the case of the (^ plants, and the
date of the first appearance of a well developed "brush" (condition of
the young " cone " when the stigmas emerge from between the bracte-
oles) in the $ plants.

The fluctuation from year to year shown by certain male plants
during from three to five consecutive seasons is shown in the Table
below :

Time of Flowering of certain Male Plants.

Very early

Reference

number

of seedling

Z 17
Z42
0G4

1916

1917
Aug. 1
July 30

1918
July 18
July 24
July 21

1919
July 31
July 29
July 22

1920
July 17
July 19
July 15

Early

Z55

OY23

BB35

Aug. 9

Aug. 1

Aug. 7
Aug. 4
Aug. 7

Aug. 4
Aug. 1
Aug. 3

Aug. 10
Aug. 15
Aug. 7

Midseason

0A4

OA30

OD27

Aug. 14
Aug. 18

Aug. 20
Aug. 19
Aug. 24

Aug. 15
Aug. 19
Aug. 15

Aug. 7
Aug. 14
Aug. 14

Aug. 18
Aug. 20
Aug. 18

Late
»»

0 15
BB23
II 41

—

—

Aug. 31
Aug. 29
Aug. 30

Aug. 28
Aug. 27
Auk. 28

Sept. 8
Sept. 7
Sept. 4

E. S. Salmon and H. Wormald 245

It will be seen that, in general, the period covered by the " time of
flowering" of the plants collectively is from the end of July to the
beginning of September. The Table shows too that the seasonal fluctua-
tions exhibited by the individual seedlings do not disguise the fact that
certain plants can be characterized as "early," "midseason" or "late"
forms. These terms do not denote distinct divisions but they may be used
for convenience in placing representative types which are connected by
other forms, the whole constituting a series in gradual sequence from
the typically " early " to the typically " late " plants. The three plants
at the head of the Table, however, appear to form a distinct class as
compared with the rest of the plants which came under our purview ;
not only have they been consistently early flowering, but, as will be
shown later, they have other characters in common.

The influence of the weather on the time of flowering is brought out
in the dates under 1920. The abnormally warm weather during the
early part of the growing season induced the " very early " plants to
come into flower even earlier than usual, but, owing to comparatively
low temperatures which subsequently obtained, the flowering of the rest
was delayed and this made them later than in previous years.

The female plants were collectively later than the male plants, the
earliest of them, even in favourable years, did not show the " brush "
before the middle of August. It was found too that the range of varia-
tion in their time of flowering was less notwithstanding the fact that
the female plants which came under observation were more numerous
than the male plants^; nevertheless plants could be distinguished which
were invariably earlier or later than the average as shown in the Table

below :

Time of Flowering of certain Female Plants^.

Early
»»

1)

Reference

number

of seedling

OD20
OY47
FF16

1916
Aug. 21

1917
Aug. 23

1918
Aug. 21
Aug. 20
Aug. 22

1919
Aug. 18
Aug. 18
Aug. 20

1920
Aug. 21
Aug. 30
Aug. 30

MidseasoQ

>>

Z19
Z47
Z 58

Aug. 31
Aug. 28
Sept. 4

Aug. 29
Aug. 29
Aug. 29

Aug. 29
Aug. 27
Aug. 22

Aug. 22
Aug. 23
Aug. 23

Sept. 5
Sept. 4
Sept. 5

Late

HH24
OB 24
OG39

Sept. 7

Sept. 3

Sept. 9
Sept. 1
Sept. 6

Sept. 7
Sept. 4
Sept. 5

Sept. 15
Sept. 16
Sept. 20

1 The number of the $ seedlings was 238, of the ^ 157.

' It is to be noted, particularly with regard to the ^ plants, that the terms "early,"
'■ midseason " and "late " as used in this paper are employed for comparing the wild hop
seedling plants among themselves and do not indicate their time of flowering relative to
the varieties of the hop under cultivation.

246 Variation in Wild Hop

The whole period throughout which plants in flower could be found
extended from the middle of July to the middle of September. For
comparison with this, the time of flowering of the wild hop in Italy, as
given by various Italian "Floras,'" may be noted: Comolli {Flora
Comense, YU. 216 (1857)) "June and July"; Moricand {FL Veneta, i.
425) "July"; Naccari (Fl. Veneta, v. 66 (1828)) "July and August";
Parlatore (Fl. Italiana, IV. 303 (1867)) "June to September." It appears
probable from the statement of the last-named author that early- and
late-flowering forms occur wild in Italy.

(2) Colour of the Bines.

The colour of the bines (axes of the annual shoots growing from the
perennial rootstock) varies from pale green to dark purplish red. The
intermediate forms usually show a distinct mottling consisting of small
red and green blotches, and even in the extreme forms faint indications
of this mottling can often be detected. Thus in the plants with dark
red bines close observation will show spots which are less dark than the
prevailing colour while the green bines may show traces of pale reddish
spots. The ridges on the mottled bines are often dark red giving the
appearance of longitudinal dark stripes.

The difference in colour between the extreme types (green bines and
red bines), and even between these and the strictly intermediate (mottled)
forms, is very pronounced and as there is comparatively slight fluctuation
from year to year in the colour of the bines of individual plants this
character is one which strikingly illustrates the variation which obtains
in the seedling hop plants.

The method of recording these colour differences was as follows :

0 pale green.

G (r) pale green with faint reddish spots.

Gr green predominating but reddish spots evident.

g and r distinct red and green blotches giving a mottled appearance.

Rg red predominating, greenish spots present.

R (g) dark red with greenish spots hardly distinguishable.

R dark red.

The records of typical plants are given below :

Plants with green bines appear to be of comparatively rare occurrence
among seedlings of the wild hop. Of the plants examined (395 in number)
only three, 0G4, Z 17, Z42 (the three plants which proved to be "very
early " in flowering) can be strictly placed in this division, with the pos-

E. S. Salmon and H. Wormald

247

Classified as

Reference

Male

yreen, mottled

number

or

or red bines

of seedling

Female

1916

1917

1918

1919

1920

Green

0G4

^

—

G

G

G(r)

G(r)

>i

Z17

C^

—

G (r)

G(r)

G(r)

G(r)

"

Z42

<?

—

Gr

G{r)

G(r)

Gr

Mottled

0A13

r?

_

Gr

g and r

g and r

g and r

>i

OA 4

<s

g and /•

(7 and r

g and r

g and r

y and r

>»

OG28

?

—

—

II and /•

g and r

g and r

„

Z47

?

g and r

g and r

Rg

g and r

^ and r

Eed

Z14

.^

Rg

Rg

Rg

Rg

Rg

Z24

?

Rg

Rg

Rg

Rg

Rg

OA28

c?

R (//)

R(9)

Rg

R{9)

Rg

0A16

?

iJ.9

Rg

R(g)

R(g)

R{9)

Rg

Z49

c?

R(9)

R{9)

Rig)

R(9)

0A19

?

R io)

R{9)

R{g)

R{g)

R(g)

OF 28

?

R

R

R

sible exception of one plant, II 35 ( ? ) which was under observation for
two seasons only when it became diseased and had to be destroyed; the
records for II 35 during those two years were G?- (1918), Gr (1919).
Another plant which approaches the "green" bine forms is W 49 ((/),
which from 1918 to 1920 was recorded as g and r (1918), Gr (1919),
g and r (1920), and which therefore on the whole finds its place with the
" mottled " plants.

Plants with red bines are by far the most numerous; in typical
examples the colour is recorded as consistently Rg, R (g) or R, or as
fluctuating from Rg to RigY. One plant, OF 28, is an extreme type
which produces very dark, purplish red bines distinguishing it from all
the other plants in the garden.

The colour of the petioles is correlated with that of the bines ; the
upper surface is usually reddish and darker than the lower side, the red
colour being the more deeply defined on the plants with red bines.
Summarized, the distribution of the red colour with respect to the colour
of the bine is as follows :

Colour of petiole

Colour of bine

■ ^
Upper surface

Lower surface

G and G (r)

Green, but rather darker than
lower surface

Green

Gt

Green, tinged with red

Green

g and r (mottled)

Rg

R{g)a,ndR

Eed, tinged with green

Eed

Eed

Green
Green

Green at distal end shading
off to red at proximal end

1 Parlatore (Flora Italiana, Vol, iv. 303 (1867)) writes, in his description of //. lupulus
<'I1 fusto ^...verdognolo o in parte rossiccio." It would appear therefore that forms with
dark red bines were not obsei-ved by him.

248 Variation in Wild Hop

(3) Relative Number of Glands on the Leaves.

The under surface of the leaves bears yellow glands easily visible
with a lens. It was found that the glands were more densely crowded
on some plants than on others. In order to ascertain whether this
character was of systematic value an estimate of the relative number of
glands present on the leaves was necessary. For comparative results it
was considered sufficient to record the number of glands found on a
certain portion of each leaf examined, the area and situation of that
portion being the same for all leaves. Countings were therefore made
of the number of glands within a \ inch circle on that part of the leaf
midway between the base of the lamina and one of the two sinuses at
the base of the terminal lobe of the leaf; two countings were thus pos-
sible on each leaf, one on each side of the midribs The countings were
taken during from two to five seasons, several leaves being taken for
each plant each year. The leaves were taken about half-way between
the lower wire and the top wire, the examination being made as the
plants came into flower.

Although there was considerable fluctuation in the number of glands
occurring on leaves of the same plant taken in successive years and even
on leaves taken from the same bine, nevertheless evidence was soon
obtained that the leaves of some plants were more glandular than those
of others. In 1916 and 1917 four or eight countings were made each
year, according to the number of suitable leaves present on the plant at
the time the examination was made; in subsequent years (1918-20) the
greater number of plants which came under observation did not allow of
so many actual countings but a sufficient number of leaves was examined
to enable a general average or range of fluctuation to be estimated.

It was found that in some plants the number was invariably below
the average, in others invariably much higher, while there were many
transitional forms between the extremes. Examples are given in the
following table, the figures, as explained above, showing the number of
glands in a ^ inch circle.

It will be observed that, of the three plants to which references have
been made previously as being " very early " and having " green " bines,
two, viz. Z 17 and Z 42, are included among those plants in which the
leaf glands may be described as " many"; actual countings in the case

1 For rapidity and convenience a thin piece of sheet metal, perforated at the centre
with a circular hole \ inch in diameter, was placed on the leaf in the place indicated, and
by means of a lens all the glands seen within the circle were counted.

E. S. Salmon and H. Wormald

249

of the third plant, OG 4, have not been taken but it has been noted that
the number of glands on its leaves are above the average and in 1920
as many as 40 were counted to the \ inch circle. There seems to be no
doubt therefore that OG 4 should be included with Z 17 and Z 42 among
those plants with " many " leaf glands.

Many

Glands few, mode- Reference
rate or many in number

number of seedling

OB 17

OD27
Z22
Z17
Z42

Male
or 1916 1917

Female (actual countings) (actual countings)

^ 43; 47; 50; 43 57; 45; 43; 56

58; 46; 27; 37 51; 49; 49; 55

J. _ 59; 57; 49; 43

37; 42 ; 37 ; 44
Q 25 ; 26 ; 29 ; 30 43 ; 42 ; 35 ; 41

22; 26; 36; 35
^ — 36; 30; 36; 29

28; 39; 24; 27
^ _ 55; 59; 64; 62

44; 39; 22; 47

1918 1919 1920

about 40 20—45 35—60

about 40 2-2-31 27—39

about 25 30—50 about 35

about 30 about 30 about 30

about 40 about 30 about 30

Moderate

.. OA39

^

16;

20;

29;

31

19;

20;

26;

18

about 20

16—25

—

13;

17;

15;

17

11;

10;

20;

16

Z14

^

12;

18;

26;

21

17;

11;

10;

15

about 20

about 20

19—35

16;

13;

14;

16

18;

11;

20;

21

»

.. OA59

?

16;

25;

18;

13

18;
23;

14;
17;

23;
16;

14
20

about 15

15—20

15—21

Few

Z43
0A6

Z58
Z49

Z27

OA28

2; 3; 2; 7

4; 6; 2; 9

7; 6; 9; 10

10; 12; 9; 14

6; 8; 12; 9

2; 7; 4; 5

7; 4; 8; 12

9; 13; 4; 2

7

about 5

3— 6

7— 9

10

about 5

3— 4

6— 9

4

about 5

1— 5

2— 6

12

about 7

6— 8

3— 8

6

4

about 8

3—10

4—15

5

about 4

As in the case of the terms used to denote the " time of flowering "
the terms " many," " moderate " and " few " used in the above sense are
applicable only to the seedlings of the wild hop ; the cultivated varieties
of hop and their seedlings are generally far more glandular than these
wild hop seedlings, plants with leaves having glands averaging over 60
in number to the \ inch circle being not infrequent in the former.

(4) The Shape of the Leaves.

The leaves of the mature plant are usually 3- or 5-lobed : those
towards the apex of the stem being trilobed or entire. Occasionally
however plants are met with which have a tendency to produce leaves
with 7 or 9 lobes by the sub-division of the terminal lobe into three
lobes, together with the division, in the 9-lobed leaves, of each basal
lobe into two. The most striking case is that of a $ plant (Ref No. 275)

250 Variatio7i in Wild Hop

which has only been under critical examination for two years but in both
these years leaves with 7 lobes and others with 9 lobes were present.

This tendency was shown by two plants Avhen grown under green-
house conditions; these were the two "very early" </ plants Z 17 and
Z 42. Cuttings from these two plants had been planted in pots and
placed in the greenhouse to test their power of resistance to mildew
under those conditions. Hop plants when grown in the small (about
5 in. diameter) pots used for these greenhouse trials generally give rise
to small unlobed or 3-lobed leaves. In March 1920 four pots of Z 17
and five of Z 42 were producing young bines which at that time were
about six inches high and the following notes were taken :

Z 17. In three pots the bines had 9-lobed leaves, i.e., of the five
primary lobes the terminal was again 3-lobed and the basal lobes each
2-lobed. In the fourth was one bine only ; this bore 9-lobed leaves below
then 7- and o-lobed forms and 3-lobed leaves above.

Z 42. In two pots the lower leaves were 5-lobed and the upper,
3-lobed; in another the leaves were 7-lobed (i.e. terminal lobe 3-lobed) ;
in the fourth the leaves were all 5-lobed, while in the fifth the lower
leaves were almost entire and the rest 3-lobed.

The tendency to produce leaves more divided than usual was thus
very distinctly shown by Z 17, less distinctly by Z 42, and this was
borne out by observations in the open, the original plants being care-
fully examined in the summer of 1920 when it was found that the leaves
of Z 17 were 5 — 9 lobed while on Z 42 a few leaves with 7 lobes could
be found and none with 9 lobes ^

These 7- and 9-lobed leaves bear a close resemblance to those of a
hop cultivated in Oregon known as the Oregon Cluster which appears
to belong to the species Humidus Aviericanus Nuttall (see (1)).

One plant, OD 24, shows variation in the other direction, the leaves
being 3-lobed only ; the leaves of this plant have other characteristic
features — they are dark green with a rather glossy surface and the lobes
are narrower and more acutely pointed than in the usual type.

(5) Colour of the Leaves.

The range of colour shown by the leaves is not great but certain
seedlings could be distinguished by having leaves evidently paler or
darker than the majority. This character was not taken into account

1 In Braungart's (3) work (p. 149), we find the statement : "In Weihenstephan war am
Zaune des Hopfenvarietatengarten ein Wildhopfen, der ofter in der unteren Eegion
7 lappige Blatter batte. "

E. S. Salmon and H. Wormald 251

until 1919 when three plants, viz. J 11 ( ? ), W 49 {^), OB 4 («/'), were
observed with leaves which, compared with those of the plants around
could be described as "light green"; all three in 1920 again had light
green leaves. It would appear therefore that the pale colour is charac-
teristic of these plants and is to be looked uj)on as a distinct variation
from the normal type^ The three very early ^ plants, Z17, Z42 and
OG 4 have also a tendency to produce leaves of a paler green than is
usually met with. Several plants had been noted in 1919 as having
" rather light green " leaves and the same description was applicable in
1920 to some of them ; others however showed no distinctive paler colour
in 1920 and in these cases the change was probably merely of the nature
of seasonal fluctuation.

BB 16 ( $ ) is a plant which had leaves of a " dark green " colour as
compared with surrounding plants, in 1919, and the same character re-
appeared in 1920. OD 24 ($ ) is another plant with dark green leaves.

(6) The Length of the Laterals.

In the great majority of cases the laterals (primary branches of the
main stems) reach a length of from four to five feet. In some instances
however plants which have consistently produced laterals of a much
shorter length have been noted. An extreme case is O A 5 (cT ) a plant
which has been under observation for five seasons but has never pro-
duced laterals exceeding one foot in length ; this plant has also short
bines which usually do not reach the top wire (see above, p. 243), while
the plants with long laterals have bines which readily reach the top
wire and grow along it or fall over to produce a dense " head " of leaves
and inflorescences. Other plants with laterals shorter than the average
are

W36 ((/), maximum length of laterals about 2 ft. 6 inches,
Z 58 ( $ ), OD 24 ( $ ), OC 11 {^) maximum length of laterals about 8 ft.

Plants of this type, when compared with the usual type, show a lack
of vigour, the bines being shorter and thinner. Whether this is due
to a pathological condition or whether the plants possess a factor for
" shortness " is not known. This shortness of bines and laterals would
seem however to be an inherent character of some hop plants. Thus in
the hop-garden at Wye College there is a cT hop plant Ref. no. J 36
(seedling of a cultivated vai-iety) which was planted out in 1911 and has

1 Varieties (^ and $) oi H. Lupulus possessing "golden" (yellow-green) leaves are
known (7).

252 Variation in Wild Hoi)

been observed through nine seasons; during this period the laterals have
never exceeded a length of fifteen inches and the plant shows no evidence
of ill-health.

(7) Characters associated with the Reproductive Organs in the ^ plants.

In the (/• plants, as seen by the naked eye, no variation has been
observed in the general appearance of the flowers. When examined with
a lens however glands are to be observed on the perianth-lobes, on the
anthers (along the furrow)^ and, in some plants, on the receptacle. There
is evidence of variation in the glandulation of these organs.

Glandvlation of the perianth lobes. The number of glands on the
perianth lobes varies with the insertion of the lobes on the short axis
bearing them. Of the five lobes forming the perianth of a flower, the
two lower ones (the outer ones of the bud) almost invariably bear very
few or no glands ; the inner lobes are generally far more glandular than
the outer ones and the innermost lobe may bear glands to the number
of twenty or even more. As a criterion of the variation in the glandula-
tion of the perianth lobes therefore only the innermost one is taken into
consideration and the numbers given below have that reference.

The number of glands on the innermost perianth lobe is generally
from 10 to 15. In some plants the number was found, during five years,
not to exceed 10 ; in others the number has generally been about 20,
while in one instance 30 glands were found on a single lobe. When the
number is consistently below 10 the plant may be described as having
" few perianth glands " ; when the number is usually twenty or more
the plant is classified as having " many perianth glands."

The following Table shows the maximum number of glands found
on the perianth lobes of certain plants during five years (three years in
one case).

Maximum number of glands recorded for hmermost peria7ith lobe.

Reference
number
of seedling 1916 1917 1918 1919 1920

Plants with

'many" perianth glands

0A61

20

28

18

21

22

..

>> 5>

OE23

—

~

30

26

24

Plants with

' few " perianth glands

Z 27

9

8

6

8

10

9 ,

>i )>

0A2

5

4

12

9

8

»i

0A4

4

6

8

7

8

,}

9 J } )

OA28

2

6

8

9

6

J>

0C6

8

7

6

8

7

1 Glands on the perianth lobes and anthers of the ^ hop plant have been recorded by
Parlatore {Flora Italiana, Vol. iv. p. 304) and figured by Braungart(3) (p. 206).

E. S. Salmon and H. Wormald 253

The Glands of the Stamens. The glands of the stamens are situated
in the furrow on the dorsal surface of the anthers. They usually form
a single row along the furrow but when numerous some are out of line
with the rest. The number usually varies from 1 to 6. A few extreme
types that have been met with are suggestive that some plants have a
tendency to produce more, others less, glands than the average. At the
two extremes are HHll and BB 35 of which the figures recorded as
showing their fluctuation during three successive years are as follows,
the first figure in each case being the lowest number observed, the
second the highest :

Reference number .

of plant 1918 1919 1920

Plant with "many" anther glands HHll 8—13 6—11 4—10

"few" „ BB35 0 or 1 (rarely 2) 0—2 0—2

Others in which the difference is less pronounced, observations having
been made for 3, 4, or 5 years, are :

Plants with a tendency to
produce many anther glands

Reference number

of plant 1916

1917

1918

1919

1920

I OA 15 3—12

2—10

6—11

2—8

3— 8

OA 43 4—10

4— 8

6—11

3—7

—

OD27 —

4—10

5—14

3—8

5— 8

Z 15 —

6—10

5—13

4—8

3—10

Z 51 3—7

3— 9

2— 7

4-7

4— 9

[ OB 14 1— 6

0— 3

0— 3

0—3

—

0D7 0—3

1— 4

1— 3

1—2

—

Z 31 0—4

0— 4

0— 3

0—6

0— 4

Z 33 1—4

1— 5

0— 3

0—5

—

1/55 —

—

0— 3

0—2

0— 3

Plants with a tendency to
produce few anther glands

Glands on the receptacle. During an examination (see (2)) of male
hop plants grown from seeds collected from cultivated varieties of hops
it was noticed that in most cases glands were present on the inner
surface of the receptacle round the insertion of the stamens. In the
seedlings from the wild hop however glands are usually absent from the
receptacle which is frequently more or less tumid forming a " disc " ;
thus in nearly every case the receptacle glands were recorded as 0 or
0(1), rarely 0(1 — 2), indicating that occasionally one gland, rarely two,
could be found and it was not always certain that in these cases the
glands had not fallen into the receptacle from the anthers. The only
plant which has shown any consistent deviation from the usual type
with respect to the receptacle glands is HH 35 of which the records for
the three years during which it has been examined are 0 — 4, 4 — 10
usually (but occasionally 0 — 3), 2 — 9\

1 a plant which has shown extraordinary fluctuation from year to year in the number
of glands present on the receptacle is OA47. In 1916 and 1917 the flowers of this plant
had comparatively a large number of glands on the receptacle. During these two years
Journ. of Gen. xi ^'

254 VmHatlon in Wild Hop

Characters associated with the Reproductive Organs in the $ plants.

No differences, apart from the time of flowering, were detected in
the pistillate spikes of the female plants but as the "hops" reached
maturity variation was observed with respect to (1) size and shape of the
hops, (2) " condition " (richness in lupulin glands), (3) aroma, (4) colour
of "strig" (axis of the strobile).

Size and Shape of the Hops. As already pointed out these seedlings
of the wild hop plant are on the whole much later in flowering than the
commercial varieties grown in this country ; in consequence the hops of
many of the plants never grow out to their full size ; their lateness,
again, makes those plants which are susceptible to mildew liable to in-
fection when in the "burr" or flowering stage so that in many cases
the plant has to be recorded year after year as "crop destroyed by
mildew." Thus the characters of the hops of some plants could not be
taken at all and those of other plants could only be recorded in one or
two seasons. A good proportion however yielded hops which were avail-
able for examination in from three to five seasons and the evidence
obtained indicates that some plants have a tendency to produce, in suc-
cessive years, hops larger than the average and a few produce relatively
very broad hops.

As the hops even on one bine vary considerably in size it was decided
to take for measurement only the largest well-developed ones. It was
found that in the majority of the plants the well-developed hops varied
in size from 1"0 x 0*6 inches to 1"3 x 0"8 inches and were generally ovoid
to ovoid cylindrical in shape. The Table below shows the fluctuation

Length of well-developed hops in inches.

Reference number

Date of flowering

of seedling

191G

1917

191S

1919

1920

in 1919

Plan!

is with small
hops

1

Z20
Z22

10

1-2
1-2

M
1-1

1-2
11

0-9
1-2

Aug.
Aug.

22
31

Z25

11

1-2

1-2

1-2

1-1

Aug.

19

FF7

—

1-4

1-7

1-2

Aug.

30

Plan!

ts with hops

OC39

—

1-6

1-5

1-6

—

Aug.

30

of a

larger type

"

0D17

1-3

—

1-5

1-4

1-4

Aug.

20

i

BB31

—

—

1-4

1-7

1-3

Aug.

27

the vigour of the plant was apparently normal, except that in 1917 the laterals were shorter
than in the previous year ; notes show that the plant in 1918 was "rather weak," in 1919
it was "weak, with small leaves," while in 1920 it reached only to about half its usual
height and bore no flowers. The glands on the other organs however did not show a
corresponding diminution in number, in fact those on the leaves showed an increase after
the first year. The glands on the receptacle of this plant were recorded as follows :

1916 1917 1918 1919 1920

2 — 12 1 — B 0 — 4 0(1) no flowers present.

E. S. Salmon and H. Wormald 255

in the length of the hops of certain plants which have consistently
produced hops larger than the average and of others which have
always developed small hops. The date on which the plants came
into flower (stigmas showing) in 1919 is also given to show that the
plants with small hops were, on the average, not later than those which
produced the larger hops.

With regard to the shape of the hops three plants have been noted
as producing hops relatively broader than the average ; the well-de-
veloped hops on these are about i inch broader than hops of the same
length on other plants, and may be described as broadly ovoid. This
difference is sufficient to make the individual hops appear almost sphe-
rical and when the crops are seen in mass the plants have a characteristic
appearance.

The fluctuation in the size of the hops from year to year in the same
plant shows in most cases such a wide range compared with the differ-
ences found between individual plants that the value of " size of hops "
as a systematic character is doubtful and will require further observa-
tions to determine the stability or otherwise of this character in the type
plants quoted in the table.

On the other hand the " broadly ovoid " type of hop appears to be a
distinct variation from the ordinary type. Of the three plants mentioned
as showing this character the dimensions of the hops were taken only
during 1919 and 1920 though the hops of all three had been previously
denoted as "broadly ovoid." The recorded dimensions of the hops of
these three plants are

Reference number

of seedling 1919 1920

OA59 1-5 X 1-1 1-2 X 0-9

Z22 11 X 0-9 1-2 X 0-9

OK 55 1-3 X 1-0 1-2 X 0-9

" Condition." The resins to which hops owe their preservative and
bittering properties are secreted within the lupulin glands which develop
chiefly at the basal portion of the bracteoles of the cones. If a ripe hop
is torn or cut longitudinally or transversely the glands are seen as a golden
yellow powder around the axis of the cone. According to the relative
number of glands present as indicated to the naked eye by the intensity
of the golden yellow colour the " condition " was recorded as poor, fail-,
good or very good. In most of the plants the "condition" fluctuated from
year to year between fair and good. A few plants however were con-

17—2

I

256

Variation in Wild Hop

sistently good to very good during the years they were under examination
while others were only recorded as poor or fair^

Generally speaking it was found that those plants which have good
condition have more glands on the leaves than those which have poor or
fair condition; this is shown in the following table:

Reference
number

Leaf

Leaf

— "■ ^

Leaf

i

Leaf

^

Leaf

of seedling

glands

glands

Condition

glands

Condition

glands

Condition

glands

Condition

OA34

11—20

—

V. good

V. good

about 50

good

21—38

good

OA29

15—27

25—47

V. good

about 25

V. good

about 30

good

about 30

good

ZiJO

20—35

19—32

V. good

about 35

V. good

25—30

V. good

28—37

good

88

—

—

—

—

V. good

about 20

good

27—39

good

0D19

6—18

.

about 15

fair

5—9

fair

about 6

fair

OK 55

—

—

—

about 15

—

about 10

poor to
fair

—

fair

BB5

—

—

—

about 5

__

1—3

fair

1—2

poor

Z43

2—7

3—7

fair

about 5

3—6

fair

7—9

fair

OA59

13—25

14—23

poor

about 15

—

15—20

fair

15—21

fair

By Aroma is here meant the smell given off by the fresh ripe hop
when bruised or broken; it is due to the essential oils contained in the
lupulin glands. In many cases the hops never became ripe enough to
give a distinctive aroma, while in others the aroma was not strong enough
to be distinctive even in the ripe hops. A few seedlings however gave year
after year an aroma comparable with that of good commercial varieties,
while in other seedlings the aroma was very strong and unpleasant, almost
nan seating ^

The records of a few of the typical plants are given for comparison.

Reference number

of seedling 1917

OD 19 very pleasant
OA 34 pleasant

OB 9 pleasant

1918

1919

1920

pleasant
very pleasant
pleasant
pleasant

pleasant
pleasant
pleasant
pleasant

very pleasant
pleasant
pleasant
pleasant

BB16
OD20
OF 19
FF35
AA5

very unpleasant very unpleasant very unpleasant

strong, unpleasant very unpleasant very unpleasant

very unjileasant very unpleasant very unpleasant

very unpleasant very unpleasant unpleasant

unpleasant very unpleasant very unpleasant

In the above plants there was no doubt as to the "pleasantness" or

"unpleasantness" of the aroma; in others however the aroma was less

distinctive and it was often difficult to decide into which category a

particular aroma should be placed, especially as in some cases the scent

would at first seem pleasant but this after a few seconds was replaced

by one which could only be called unpleasant.

^ Braungart (3), p. 209, observes that the "condition" ("Sekretgehalt ") varies with
different plants of the wild hop.

- No pronounced garlic-like smell was met with in the green hop.

E. S. Salmon and H. Wormald 257

There appears to be no correlation of " condition " and aroma; thus

88 aud OA34 have condition good to very good, aroma pleasant to very pleasant
OD 19 has ,, fair , ,, pleasant to very pleasant

AAo „ ,, good , ,, unpleasant to very unpleasant

OF 19 ,, ,, fair , ,, very unpleasant

Colour of the Strir/. A range of colour similar to that of the bines
also obtains in the strigs of the hop cones. "Strig" is the name com-
monly used to denote the axis of the hop cone ; on the swollen nodes of
this axis are inserted the flowers with their accompanying bracts and
bracteoles. The nodes of the strig are usually of a darker colour than
the internodes. In most of these plants the internodes are reddish with
the nodes a dark purplish red. In extreme cases the strig is pale green
with usually a faint reddish tinge at the nodes, or the strig is dark red
throughout. The colour in individual plants fluctuates to some extent
even in the hops of the same year ; certain plants however produce green
strigs (green throughout or green with reddish nodes) year after year
while others consistently produce red strigs (red internodes and dark red
nodes or strig red throughout).

It was expected that the colour of the strig would be correlated with
that of the stem. Of those plants with red bines most had red strigs
with dark red nodes, a few had strigs dark red throughout while in some
the strigs were green with pale reddish nodes. Only one of the $ plants
had green bines ; unfortunately this plant became diseased and had to
be destroyed before the hops matured and so the colour of the strig was
not determined. Those plants of which the bines were intermediate in
colour between the green and the red plants had strigs which were almost
invariably recorded as green (or faintly reddish) with reddish nodes. It
would appear therefore that there is a tendency for the plants with red
bines to produce red strigs and those in which the red colour of the bines
is less pronounced to produce green or pale reddish strigs. As is well
known the green-bined cultivated varieties of hop have green strigs.

One plant, Z .58, has strigs which have a polished (vernicose) appear-
ance ; this character has appeared, though not so pronounced, in two
other plants.

(8) Degree of susceptibility to the mildeiu Sphaerotheca
Hunmli {DC.) Burr.

Accounts have already been published ((4) to (7)) of the differences
shown by individual seedlings as regards susceptibility to the attacks of
the Hop Mildew. A resume of the chief facts observed will be given, and

258 Variation in Wild Hop

then the question considered as to whether any correhition with mor-
phological characters exists — a point not hitherto dealt with.

The method of testing the individual seedlings for susceptibility to
mildew was as follows. The seedlings when 1- or 2-years old, were ex-
posed to infection in the greenhouse by being constantly inoculated with
conidia under conditions very favourable to the growth and spread of
the mildew. During the following winter these seedlings were planted
out in the Experimental Hop-garden, where natural inoculation by the
mildew was relied upon — a method found to be quite satisfactory. Owing
to the late-flowering habit of the seedlings it was found that October
was the best month for examination as to the incidence of mildew in the
open. In the case of the $ plant, the production at the end of August
and during September of the female inflorescence and young developing
hops provided the best possible infectible material. With the (/ plant, .
the infectible material in the autumn consisted for the most part of the
young leaves of the axillary side shoots which developed from the lower
portion of the main stem. Mildew was prevalent generally in the Hop-
garden in the autumn of each season from 1916 to 1920, and particularly
severe outbreaks occurred in 1916, 1919 and 1920. From 1914 onwards
480 seedlings were kept under observation; first as seedlings in the
greenhouse and then in the Hop-garden where they grew on to mature
plants.

From the data contained in the Table on p. 259, it will be seen that
different seedlings — exposed in the open to identical conditions of soil
and weather^ — showed very different reactions to the attacks of the
mildew.

In the case of the most susceptible $ plants (indicated by the sign
*S'^ in the above Table) the greater proportion of the " hops " (cones)
were deformed by the mildew year after year, or not infrequently the
whole crop of " hops " was entirely destroyed, the female inflorescences
being attacked when quite young and their further development per-
manently arrested, the inflorescences becoming converted into white
"knob"-like growths (see PI. XXV, fig. 1). With the most susceptible cf
plants, the mildew sometimes invaded the branches of the inflorescence.

On the other hand, complete or almost complete resistance to mildew
was shown in the open by certain individual seedlings, both cf and ? .

Where, in the Hop-garden, such resistant seedlings grew side by side
with the very susceptible plants, the contrast, season after season, was

^ The seedlings being all "trained" the same way ensured similarity of conditions as
regards illumination.

E. S. Salmon and H. Wormald 259

Table shuwlmj the degree of m^ceptibility to mildeiv in the yrwuhouse (G),
arul in the llop-garden (H).

Reference
number

191«

1017

1918

1919

1920

of seedling

Sex

H

G

II

G

H

G

11

G

u

G

V91

?

—

0

—

0

S'

0

0

0

0

V92

$

—

0

— ■

0

S'

0

0

0

0

V93

$

—

0

—

0

,S''

S'

0

0

0

Zl

s

—

—

,S'-

—

,s:--^

,S'^

,S''

Z2

^

—

—

0

0

0

0

0

0

0

0

Z14

c?

—

—

0

—

S'

0

0

0

.S"'

0

Z15

0

—

—

.s-^

—

S'

hS

S"

i^'

hf^

hs

Z22

?

—

—

0

—

6'-'

"0

S'

0

,Si or S'^

0

Z23

(?

—

—

s-

—

0

hs

,S''

hs

Z24

?

.s--'

—

.S'»

—

S'i

s

.S'

's

,s»t

Z25

?

0

—

0

—

0

0

0»*

0

,S'i

0

Z26

?

,S^

—

s^

,s-»

,S'-

s

■S=' t

s*

Z38

?

—

—

S'

_

,S'^

0

s

Z39

?

—

— •

S-*

s

S'i

s

S'

s*

Z41

?

—

—

s->

—

s-

—

s->

.s-^t

s*

Z42

(J

—

—

0

—

0

0

0

0

Z43

$

—

s-

s-

.s-1

.S'-

OA25

?

—

—

S'

—

S-'

—

S'

,S''

s

OA26

?

—

—

s-

—

SH

—

.S'^t

.s*

-S'' t

s*

OA33

^

—

—

0

—

0

hS

,S2

hs

.S''

is-

0

OA34

?

—

—

s-

S-'

,S'i

0

.SI

OA35

?

—

S-i

,S'2t

,S'' t

s

s»t

s

OA49

5

—

—

0

0

0

0

0

S'

0

OB 34

?

—

—

0

—

.51

0

0

0

0

0C6

c?

—

—

0

—

0

h^

,si

IS

hS

hS

0D16

$

—

—

S3

—

sn

. —

s->

s

s»t

s

0D17

?

—

—

,si

.s^

—

*•'

i.S

,S''

kS

0D18

9

—

—

,S'3

—

S^'t

.

S--

S-'i

s*

0D19

9

-_

—

0

0

0

s^

0

,S'i

0

OK 38'

V

s^

—

s-

0

S-i

0

.S'

0

s^

0

OR 391

c?

-S''

—

-S''

0

0

0

0

0

0

0

.S^ = mere trace of mildew, .S-=fair amount of mildew; ,S'-*=:plant very mildewed.
S't indicates that the crop of hops was destroj'ed by mildew. .V indicates that the plant
under greenhouse conditions was normally susceptible to mildew, and .S'* that extreme
susceptibility was shown. 0 = no mildew present. — indicates that no record was
available.

Where running numbers occur, e.g., V91, V92, &c., the respective seedlings were
growing next to each other in the same row and 3 ft. 6 in. apart.

** A few small patches of mildew appeared on August 7, on a few of the young leaves,
after a spell of abnormally cold, dull weather. The mildew patches soon died away and in
October the plant was entirely free from mildew.

1 OR 38 and OR 39 were both immune in the greenhouse in 1914.

must marked. Thus the $ seedling OD 19 remained entirely free from
mildew through the seasons 1917 and 1918 and showed only the merest
trace of mildew in 1919 and 1920, while the $ seedling OD 18 planted
next to it and growing so close that the lateral branches of each became
intertwined, was so susceptible to mildew that each year the hops and
the leaves were infected to the degree of S'-^ and in 1918 and 1920 the
crop of hops was completely destroyed by mildew. A photograph taken

260 Variation in Wild Hop

in 1920 is given in PI. XXV, fig. 2 showing a " lateral " of OD 18 twin-
ing round a lateral of OD 19 ; OD 18 is so virulently infected with
mildew that no "hops" have been able to develop, all the inflorescences
having been turned into little white " knobs," while the hops of OD 19
have resisted all attacks and developed normally, quite free from a trace
of mildew.

Complete immunity to mildew was shown by 27 seedlings (13 (/>
14 $ ) when grown in the greenhouse. When cuttings were taken from
these seedlings after they had been growing for several years (in
one case 5 years) in the (manured) hop garden, they showed the same
complete immunity in the greenhouse. On the other hand, cuttings taken
from other seedlings which showed varying degrees of susceptibility in
the Hop-garden were susceptible when grown in the greenhouse. In no
case has any seedling which has shown immunity in the greenhouse
shown susceptibility there when tested again in other seasons.

Seedlings immune in the greenhouse show usually some degree of
susceptibility when tested in the open, the degree of susceptibility being
usually only very slight, but in rare cases reaching to S- or even to S'^.
(See Table, p. 259.)

Certain individual seedlings (7 in number, 4 </ and 3 ?) showed
the phenomenon ((6) (7)) of "semi-immunity" when grown in the
greenhouse ; these seedlings exhibited in the open also a certain degree
of resistance.

A numerical analysis of the 480 seedlings based on the relative
susceptibility to mildew in the open gives us the following figures :

Number of

Degree of susceptibility

seedlings

Percentage

S'^ (including S'^ f) ''^c'' year(

96

20-00

(11<?; 85?) s

,S2 to S3 (sometimes S^t) \

115

23-96

{35cJ;80$) i

^2 /

(54c?; 26$) i

80

16-67

S^ to S- \

(41(?; 48?) 1

89

18-54

(19cJ; 12?) i

31

6-46

0—^1 \

(9(?;9?) i

18

3-75

(all c?) i

6

1-25

0— ,S3 1

(9cJ; 18?) \

27

5-63

(all?) i

3

•60

,Si to S^ \

(5(?; 10?) i

15

3-12

Totals 480 100

E. S. Salmon and H. Wormald 261

Of the 480 seedlings raised, 189 plants (39 '38 /J proved to be (/ ,
and 291 (GO-62 7J ?.

A separate analysis of the records of the two sexes gives us the
following figures :

Analysis o/* 189 (^ seedlirujs of the " wild hop."

Number of

Degree of susceptibility

seedlings

Percentage

S»

11

5-82

,9-— .S«

35

18-52

S2

541

28-57

,S''— S-

41

21-69

.S'l

19-

10-05

0—S^

9^'

4-76

0

6*

3-17

0—^2

95

4-76

S^-S^

5

2-65

Totals 189 100

Analysis o/' 291 $ seedlings 0/ the " xvild hop."

Number of

Degree of susceptibility

seedlings

Tercentagt

<S^ (including

S-'t) ...

85

29-21

S-—S'^ (sometimes S'^ -t")

80

27-49

s-^

26

8-93

s^—s-' ...

48

16-49

S^

12

4-12

0—S^ ...

9

3-09

O—S"- ...

18

6-19

O—S'^ ...

3

1-03

SI— S--! ...

10

3-44

Totals 291 100

The percentage figures in the two tables given above indicate a
higher degree of susceptibility in the $ seedlings than in the </. It
would be unsafe, however, to consider this proved, owing to the circum-
stance that the observations as to the incidence of the mildew were
made at a time peculiarly favourable for attacks on the $ plant. The (/
and the ? plants would need to be tested at a time when each provided
the same amount of infectible material before any inference could safely
be drawn as to their relative susceptibility.

Since excellent material to test the susceptibility was always present
in the case of the $ plant, it appears safe to use the analysis of the
records of the 291 ? seedlings on which to base an estimate of the

1 Of these, 1 seedling was semi-immune in the greenhouse.

- Of these, 3 seedlings were immune in the greenhouse.

•^ Of these, 5 seedlings were immune, and 1 semi-immune, in the greenhouse.

■• Of these, 5 seedlings were immune in the greenhouse.

5 Of these, 2 seedlings were semi-immune in the greenhouse.

262 Variation hi Wild Hop

percentage of seedlings occurring in the different grades of suscepti-
bility.

In the highest grade of susceptibility, in which the plant was each
season mildewed to the extent of S'^, there are 85 seedlings. Our records
show that of these 85 seedlings, no less than 34 had the crop of " hops "
entirely destroyed by mildew in one or more seasons ^

In the next class, the incidence of mildew varied from S^ to *S^* in
different seasons. This class comprises 80 seedlings ; of these, 21
seedlings were attacked in some season to the extent that the entire
crop was destroyed".

If we add together the numbers 34 and 21, we get 55 as the number
of seedlings which have had their entire crop of hops destroyed in some
season or seasons. This number represents 18*90 °/^ of the total number
of the ? seedlings.

We may consider the class of 85 seedlings attacked ;S^" each season —
representing 29"21 °/^ of the $ seedlings — as exhibiting excessive
susceptibility to mildew, probably exceeding that shown by any com-
mercial variety of hop cultivated to-day.

To the class of very susceptible seedlings may safely be added the
80 seedlings of the S'^ — S^ class. This gives us 165 seedlings (or 5670 °/„
of the total number) belonging to a grade of susceptibility in which
virulent attacks of mildew are common.

In the class S^, consisting of seedlings which have shown a medium
attack of mildew each season, we have 26 plants, or 8'93 °/„

In the next class, where the incidence of mildew varies in different
seasons from S^ — S^ — that is, from a trace of mildew to a medium
attack — we have 48 seedlings, or 16'49'^/„. In this class an appreciable
degree of resistance to the mildew begins to be shown by some of the
seedlings, but it is in the succeeding classes that this phenomenon is
most clearly seen.

In the class S^, i.e. those seedlings which have shown only a trace of
mildew each season, we have 12 plants. Of these, 5 cannot be considered
to have been sufficiently tested, leaving us with 7 seedlings which have
proved to be persistently resistant in the open. Of these 7 seedlings, 4 are
immune and 1 semi-immune, under greenhouse conditions.

' The actual records for these 31 seedlings are as follows, where the integer gives the
number of seedlings, the numerator of the fraction the number of times the entire crop
was destroyed and the denominator the total number of seasons during which observations
were taken: 2, |; 5, t; 7, J ; 2, §; 7, J; 11, ^.

2 In all cases S^ f was reached only once by these seedlings, the actual records being
as follows : 1, i; 9, J ; 6, A; 5, J.

E. S. Salmon and H. Wormald 263

In the class 0 — *S'^ wo have 9 seedlings, of which 1 has been in-
sufficiently tested. Of the remaining 8, 6 are immune, and 1 semi-immune,
under greenhouse conditions.

In the class 0 — *S-, in which there are 18 seedlings, 2 plants (both
of which show immunity in the greenhouse) have proved decidedly re-
sistant in the open, the grade of infection only very rarely I'eaching to S".
The other seedlings in this class, while certainly showing some resistance
to the mildew, are liable to fiiirly serious attacks in some seasons.
Among these seedlings are 2 which are immune, and 1 which is semi-
immune, under greenhouse conditions.

The 13 seedlings in the classes 0 — S'-^ and *S'' — S'^ do not constitute
in either case a proper group. Of the 3 seedlings showing 0 — *Si', one
only probably possesses a slight degree of resistance. Of the 10 seedlings
showing S^ — S^ in different seasons, 9 almost certainly belong to the *S'^
or S'^ class, while the remaining seedling, OR 88, which is immune in
the greenhouse, is exceptional in showing in the open as high a grade of
susceptibility as *S'^ in some seasons.

The degree of susceptibility to mildew shown by the seedlings has
no correlation, so far as we can see, with other characters. Thus among
the resistant plants are some with green bines others with red bines,
some are early, others late, some have many glands, others few, and
similarly with the very susceptible plants. Again, plants Avhich are in-
distinguishable in other ways show different powers of resistance towards
the fungus as shown by the following records :

Description of plants

^ very early, green bines, many leaf ]Z17 0 — — — — O — O
glands, few perianth glands, rather ■Z42 0 — 0 0 — O — O
pale leaves J 0G4 — — S- — S'' — S' —

$ late, red bines, few leaf glands, ) ^ . An n 0 0 0 0 S^

condition fair, aroma pleasant,^ ^-r. ^ „;j ,^3 03 ct.

strig red

5midseason, red bines, few leaf
glands, condition poor to fair, strig
red (aroma generally more pleasant
in OD 19 than in OD 16)

$ late, red and green (mottled) bines, '\
many leaf glands, condition fair I Z 22 O — S^- 0 S" O S^ or S- O
to good, aroma pleasant, strig red- [ OG28 — — -S* — S'' — .S*t —
dish with darker nodes j

As regards 'the characters taken, then — which comprise the most
important morphological ones — for many (if not all) of the " immune "
forms there appear to be susceptible forms which are similar morpho-
logically. It is to be noted that similar forms in other species of plants

OD 19 o — o 0 .51 o ,51 o
0D16 S-^ — S-^-f — S^ S .S-t S

264 Variation in Wild IIojj

have been met with, both in nature and also among the hybrid offspring
produced in certain breeding experiments.

One of us has pointed out (8) that within the species Broinus mollis
two "races" exist, morphologically identical but one susceptible, and the
other immune, to certain forms of the mildew Erysiphe Grayninis DC.

Vavilov (9) states that Triticum dicoccum comprises, with respect
to E. Graminis, both susceptible and immune races'.

Biffen (10), commenting on the characters of the hybrid plants raised
by crossing " Michigan Bronze " (a wheat susceptible to the rust Puc-
cinia glumarum) with " American Club " (immune), says : " The two
parents differ from one another in other characters besides the immunity
and liability to the attacks of rust, and it may be noted that, as one
would expect, individuals similar to the immune parent morphologically
but susceptible to rust were found, and also the rust-proof form of the
susceptible parent^."

General Considerations.

The various characters reviewed above would appear to have varietal
significance, for not only do they persist from year to year in the same
individual but all the evidence to hand suggests that many — if not all —
of the characters mentioned would reappear on propagating the plants
vegetatively ^ Thus two plants have been raised by "cuttings" from a
$ seedling (OR 38) and one plant from a cf seedling (Z 11); it was found
that the plants raised by cuttings were similar to the original plants in
time of flowering, colour of bine, glandulation of leaves and J' flowers,
" condition," aroma and colour of the " strig " of the hop-cone, and degree

1 Ou the general question, Vavilov (9), p. 52, remarks: "the physiological individuali-
sation of race is very often accompanied by morphological characters which, however, are
externally not very conspicuous. In these cases the peculiarity in behaviour towards fungi
obliges the observer to pay more attention to this or that previously unsuspected race, and
in the end he usually succeeds in finding in it some other confirmatory dilierences."

■•^ In another case, it was found that no correlation existed between the leaf-characters
of the parent varieties of wheat used in "crossing" and susceptibility to rust — "the
immunity [in the F2 generation of the hybrids] simply depended on the luck of the shuffle"
(Biffen (11)).

^ This has been shown ((2), p. 177), by systematic records to be the case with the
corresponding characters of (J plants raised from seeds taken from cultivated varieties. It
is also to be noted that the characteristic features of certain recognised commercial
varieties (which are propagated vegetatively by cuttings) are of the same order as some of
the characters discussed in this paper, e.g., early flowering ("Early Bird" variety), pale
green bine ("Canterbury Whitebine "), very pleasant aroma (" Goldings "), large size of
hop " cone" and strong aroma ("Fuggles").

E. S. Salmon and H. Wormald

265

of susceptibility to mildew. Again, a large number of " cuttings " have
been taken from a large number of seedlings some of which were immune
from, and the rest susceptible to, mildew in the greenhouse ; the dis-
tinctive characters of each seedling remained constant'.

A certain range of variation is found in each series of characters, the
extreme forms in some series being widely different, as in the green-bine
and the red-bine forms, and in the resistant and the very susceptible (to
mildew) forms. There is no evidence of correlation of any one character
with the others, with the possible exception of the case of the three very
early ^ plants in which very early flowering is associated with a green
bine and many leaf-glands. The cases of apparent correlation (to which
reference has been made on p. 256) between (1) the number of glands on
the leaves and the " condition " of the hop-cone, and (2) between the
colour of the bine and petiole and the colour of the strig, cannot perhaps
be considered as the association of distinct characters but rather as mani-
festations- of the one character (glandulation in the one case — production
of a red colouring matter in the other) in different organs.

A classification to include all forms is at present impossible as the
evidence at hand suggests that given an unlimited number of seedlings
a gradual transition would be found between the extreme types of each
series and even an approximate estimate of the possible number of com-
binations of characters cannot be given.

The accompanying table however may prove of interest as indicating
the combinations of character with respect to time of flowering, colour
of bine and the relative number of glands on the leaves as shown by the

Green

Colour of the Bines

Mottled green and red

Red

/ Early $

(* very early) $

Time I ,,. n

f j Midseason ^

Flowering " +

Late $

^ „ ?

3
1

17
22

4
11

5
4

46
62

3
50

6
1

49
42

1

28

Many Moderate Few Many Moderate Few Many Moderate Few
number number number

Relative number of Glands on the Leaves

1 No "cross "-breeding of the different forms has yet been carried out. Seeds collected
from " open " flowers of a $ seedling (Z 25) immune to mildew produced 33 seedlings,
which when tested for mildew-susceptibility in the greenhouse, proved to consist of
24 immune and 9 susceptible plants (7).

266 Variation in Wild Hop

395 plants which have been examined ; the figures indicate the number
of plants in which any particular combination in the three series was
found.

It is to be observed that the above divisions are altogether arbitrary
and only show the range of variation in the plants which came under
observation ; with a large number of plants probably more extreme cases
would be met with making necessary a modification of the divisions. It
is also to be noted that only the extreme forms and the obviously inter-
mediate ones can be readily assigned to their particular divisions ; the
rest are placed as they tend on the whole, with the records to hand,
towards one or the other extreme or towards the intermediate type\
Thus with the three series of characters selected, each considered as com-
prising three divisions, there are 27 possible combinations of characters
for each sex, of these 17 have been observed among the (^ plants and 16
among the % plants in an examination of 895 plants in all.

It will be seen from the table that the great majority of the plants
have red bines, that most of these again are included under " midseason "
or "late," and have few (or a moderate number of) glands.

Of all the plants examined not one has proved to be of any com-
mercial value, the majority being much too late and all having hops too
small to make cultivation of the plants a profitable undertaking under
modern conditions. The fact that resistance to HojT Mildew is a character
of some of the seedlings suggests that this desirable quality might by
breeding operations be combined with other qualities of a good hop ; in
one plant, OA 34, it is combined with relatively good " condition " and
a pleasant aroma.

We have not been able to find any detailed account of variations
observed in the wild hop. Braungart (3), in his elaborate monograph on
the cultivated varieties of the hop, deals only very cursorily with the
wild species I Systematists have separated one variety, viz. hrachy-
stachyus Zapal. Gonsp. Fl. Galic. ii. 94 (1908), described as follows:

1 No individual seedling has been met with which, as regards the characters noted in
the present paper, has shown fluctuations ranging from one extreme to the other.

^ Braungart (3), p. 146, referring to the wild hop growing in Switzerland, says: "Die
in unmittelbarer Niihe stehenden Individuen sind sich dabei oft nach alien Richtungen,
auch in den Zapfen, sehr ahnlich, oft aber auch nicht, und welter entfernt voneinander
stehende sind meist nach alien Beziehungen sehr abweichend." In a paper by J. Schmidt
(C. rend. trav. Lab. Carlsherg, Vol. xi. 314 — 329 (1917)) entitled " The occurrence of the
wild hop in Denmark," a reference is made to " the numerous forms in which this plant
occurs" but their distribution, and not their characteristics, is mainly discussed. Further,
Schmidt uses the term "wild" merely to indicate " not cultivated," and it seems probable
that he is dealing partly with escapes from cultivation.

JOURNAL OF GENETICS, VOL XI. NO. 3

PLATE XXV

E. S. Salmon and H. Wormald 267

" Blatter kleiner, Fruchtstande etwa um die Halfte kvirzer, nur 1-1,5 cm.
lang. Schuppen zmn grossten Theile zugespitzt, ctwas behaart, an dor
Spitze roth werdend. — Galizien." (Ascherson u. Graebner, Hyn. d. mittel-
eur. Fl. IV. 597 (1908-13)^)

Summary.

1. The wild species H. Lupulus is composed of a number of forms
with distinctive morphological and physiological characters.

2. Forms similar morphologically (for the characters taken) possess
distinctive physiological characters (e.g., immunity from, or degrees of
susceptibility to, mildew).

EXPLANATION OF PLATE XXV.

Fig. 1. Laterals of a seedling (of the wild hop) very susceptible to mildew ; the inflor-
escences have been arrested in development.

Fig. 2. Intertwined laterals of two adjacent hop plants one susceptible to, the other
immune from, mildew (see p. 260).

BIBLIOGRAPHY.

\. Salmon, E. S., and Wormald, H. '■'■ Humuhis americanuSj Nutt." Journ. oj
Botany, 1915, pp. 132—135.

2. Wormald, H. "Variation in the Male Hoj), Ilumulus Lupulus L." Joinm.

Agric. Science, Vol. vil. pp. 175 — 196 (1915).

3. Braungart, R. Der Hopfen (1901).

4. Salmon, E. S. "On Forms of the Hop (77. Lupulus L.) resi.stant to Mildew

{S. Hurauli, (D.C.) Burr.)." Journ. Agric. Science, Vol. viil pp. 455 — 460
(1917).
5. . Idem II. Journ. Genetic.^, Vol. viii. No. 2, pp. 83—91 (1919).

6. . Idem III. Aiinals of Applied Biology, Vol. v. 252—260 (1919).

7. . Idem IV. L.c. Vol. vl pp. 293—310 (1920).

8. . " On Erysiphe Graminis D.C, and its adaptive parasitism within the

genus Bromus." Annal. Mycolog. Vol. n. pp. 255—267, 307—343 (1904).

9. Vavilov, N. I. " Immunity to Fungous Diseases a.s a physiological test in

Genetics and Systematics, exemplified in Cereals," Journ. Genetics, Vol. iv.
pp. 49—65 (1915).

10. BiFFEN, R. H. " Mendel's Laws of Inheritance and Wheat Breeding." Journ.

Agric. Science, Vol i. pp. 40 — 44 (1905),

11. . "Experiments on the Breeding of Wheats for Engli.sh conditions."

Journ. Roy. Hort. Soc. Vol. xx.xil. ; Bej^ort Third Lnternat. Conf. (1906) on
Genetics, pp. 373—377 (1907).

1 We are indebted to Mr E. G. Baker for this information.

i

MALE-STERILITY IN FLAX, SUBJECT TO TWO
TYPES OF SEGKEGATION.

By W. BATESON, M.A., F.R.S.
AND Miss A. E. GAIRDNER,

Of the John Innes Horticultural Institution.

(With Plate XXVI.)

In 1912 a single plant was noticed in a patch of Linmn grandiflonnn
sown in our border of annuals. It was remarkable as having a blue
flower and procumbent habit, whereas L. grandifloruiu is deep crimson
and erect. In its low stature the new plant somewhat resembles the
various kinds of flax cultivated for oil in many countries. Several of
these oil-flaxes have in subsequent years been grown here, but we had
none in 1912, which was indeed the first year in which any form of
L. usitatissimum was sown in this garden. All the flaxes cultivated,
whether for fibre or oil, are included under that specific name. The
new plant, which we shall here call the " procumbent," is also evidently
a true usitatissimum, though in being procumbent it differs from any
variety which we have seen. Obviously it has no connexion with
grandifiorum, which is a very distinct species. We have no surmise as to
what its origin may have been, but somehow a stray seed from which
it arose must have got mixed with the grandifiorum.

Since in 1912 experiments on flax were begun with various objects
the appearance of the procumbent was a matter of interest, and its seeds
were collected. Whether the flowers had been covered is not recorded.
Probably they were, but the point is of little importance, for only a
small percentage of crossing occurs in usitatissimum. The procumbent
bred true from the first and has continued true in each year since,
except that a naturally cross-bred plant appeared once in a row raised
from unprotected seed, as may happen with any flax.

The new form differs from any other flaxes that we know in the
following respects.

1. It is procumbent during growth, branching much from the base,
and the stems reach about 2 ft. in length, lying at first fiat on

Journ. of Gen. xi ^"

270 Male-Sterility in Flax

the ground, turning upward as flowering begins, and finally
standing more or less erect.

2. The style is very pale in colour, and when the petals of an opening
flower are pulled off it projects well above the sepals, which are
perhaps slightly shorter than those of ordinary flaxes. But like
other forms of usitatissimum the variety is definitely homostyled,
the tops of the anthers just reaching beyond that of the stigma.

3. It is very late in flowering, about 10 days later than any other
variety that we have.

In other respects there is nothing special to note. The capsule and
seeds are of the ordinary size, not large as those of our oil-flaxes are.

The first crosses were made in 1916 by Miss M. Michell of Cape
Town, who fertilised the procumbent with pollen from a tall white-
flowered fibre-flax. The object of this cross was to study the genetics of
height. Though from the extraordinary uniformity in height character-
istic of pure lines of flaxes these plants seem well adapted for such
work, the distribution of height in F. is complicated, all intergrades
occurring, with indications of segregation so imperfect that useful results
could only be obtained by measurements on a very large scale continued
through several generations. Much work of this kind has been carried
out on which we may publish a report later, but at present it must
suffice to say that though segregation in respect of height occurs, and
though the parental types, both procumbent and tall, reappear in F^,
they are rare, and the height-curves of these families show no obvious
dimorphism.

The purpose of the present paper is to give the facts respecting the
behaviour of a male-sterile form which appeared in F<, in 1918. F^ is
erect and intermediate in height, and in colour also (as T, Tammes
found ^ in crosses between blue and white flaxes, and as we have also
often seen in other cases). In F.^ Miss H. Garlick, who was then in
charge of these experiments, observed certain plants, both blue, white
and intermediate in colour, which had flowers with reduced petals, as a
rule scarcely opening at all. In these flowers the anthers are more or
less completely aborted. A row of such plants growing beside a row of
normally flowered flaxes is shown on PI. XXVI, together with photo-
graphs of drawings of the normal and of the abnormal flowers in
various stages.

1 Tammes, T., Rec. Trav. hot. Neerl., 1911, viii. pp. 264-5.

W. Bateson and a. E. Gaikdner 271

Varieties with aborted anthers are familiar in a great number of
plants. Some, as for example that of the Sweet Pea, are quite sharply
defined, others, as for instance that studied by Miss Pellew in Gamjxi-
nula carpatica may exist in many inturgrading forms. As in the
Caryophyllaceae and Labiatae, such tlowers may coexist in various
grades on the same plant, and may even be associated with normal
hermaphrodite flowers. In these and other Natural Orders flowers with
aborted anthers are both by systematists and writers on genetics often
referred to as females, and the species possessing them are called gyno-
monoecious or gyno-dioecious as the case may be. It is not easy to
apply any definition which will distinguish flowers or plants with
aborted anthers from the normal female flowers of some dioecious
species (e.g. Lychnis), but to avoid raising this theoretical question we
propose to follow the usage of some American authors and call them
male-steriles. In these flaxes the sterility of the anthers is nearly but
not quite complete. No grades of plants were recognized beyond the
ordinary hermaphrodites and the male-steriles, but anthers of the steriles
occasionally reach a degree of development sufficiently complete to pro-
duce a little good pollen. Flowers with this low degree of male-fertility
may be found sporadically, as it seems, on any male-sterile plant. From
their pollen a few seeds have been obtained in self-fertilisation which
gave rise again to male-steriles only. No degree of male -sterility was
ever observed in the flowers of ordinary flaxes. In the breeding work
it was not thought necessary to emasculate the male-sterile flowers, and
in only one mating (Expt. 5) is there reason to suppose that disturbance
was caused by their pollen.

In previous experiments on the genetics of male-sterility the male-
steriles were introduced as an already recognized type. In our work
they arose in F., from a cross between two fully hermaphrodite types.
Subsequent experience proved that the new form was brought in by the
pollen of the common flax, and that the procumbent is genetically
hermaphrodite on both male and female sides. Twenty-four fibre-flax
plants of various kinds (Expts. 6 — 10) tested by using their pollen on
male-steriles, gave in all 640 plants, all male steriles. To these 24
which gave a uniform result may be added the white plant, from the
male side of which the original ^2 family containing male-steriles was
derived, and a tall blue plant on similar grounds (Expts. 1 and 2).

One tall plant (Expt. 11) used as male on a male-sterile gave a
hermaphrodite. If authentic, this constitutes a single exception to the
rule that the pollen of normal flaxes carries the male-sterile only.

18—2

272 Male- Sterility In Flax

Whether it should be discarded as an error or not we cannot yet say.
It came in a family containing only two plants, the other being a male-
sterile.

From the evidence of these 26 plants we must conclude that the
fibre-flaxes in general, perhaps always, are heterozygous in respect of
the male-sterile allelomorph, and that in segregation this allelomorph is
relegated to the male side.

As regards oil-flaxes we have only preliminary indications and can
make no statement as yet.

The procumbent itself is clearly hermaphrodite in genetical com-
position on both male and female sides. The male side was tested as
follows. Tall X procumbent (Expt. 3) gave from 3 F^ plants 559 all ^ in
F2 ; white X procumbent similarly gave 157 5 in F^ (Expt. 4) ; also
male-steriles fertilised by procumbents (3) gave 101 5 , and also 2 male-
steriles (Expt. 5) which may safely be assumed to have arisen by self-
fertilisation, the mother-plant not having been emasculated.

But when the procumbent is used as mother and fertilised with
pollen from a fibre-flax, F^ is a normal hermaphrodite, and in F2 the
male-steriles appear as 1 in 4. It is proved therefore that the female
side of the procumbent must in some way be different in constitution
from the female side of the ordinary flaxes. In heterozygosis with the
female of the procumbent the dominant factor for anther-development
passes with its negative allelomorph to both male and female organs of
the offspring, thus producing an ordinary Mendelian result, but when
the same negative is in heterozygosis with the female side of the
common flax, it passes wholly to the pollen, evidently segregating not
later than the constitution of the sexual organs. The ordinary flax,
though heterozygous, is thus able to breed true just as Begonia Davisii
does for singleness (though heterozygous for doubleness) and as the
various Oenotheras investigated by Renner may do, though heterozygous
for several complexes.

In connexion with this example the somewhat cognate evidence
discovered by Miss Pellew in Campanula carpatica should be recalled.
There also very similar plants were shown to possess distinct types of
segregation in respect of the same negative allelomorph. The case of the
flaxes differs from that of the Gampanida, for in that example the plants
in which both allelomorphs went to both sexual sides showed an
irregular distribution, whereas here they follow a normal Mendelian
system. It will be interesting to see whether a cytological basis for
these phenomena can be detected.

W. Bateson and a. E. Gatrdner 273

Attempts to use the male-steriles as males in crossing have hitherto
failed, the pollen produced by them being very small in amount.

Much interest would attach to the genetics of plants made in such a
way that the male-sterility was brought from the maternal side into
combination with one or other of the hermaphrodite factors introduced
from the paternal side. Various possibilities are thus opened up. On
these questions we have as yet no suflScient evidence, Expt. 19 being
the only one bearing on the point. There the hermaphrodite element
was that of the female side of procumbent, and it was brought in from
the male side of a plant homozygous in F3 or a homozygous F2. Tried
on a male-sterile it gave 6 male-steriles and 1 9 . That it should have
given the 9 is worth noting, but work on a large scale is required to
explore the many alternative distributions which will possibly be
encountered. Material for these investigations is being prepared.

Plants in which the distribution of factors differs on the two sexual
sides of the same plant now form such a prominent subject of genetical
consideration that some simple terms are required to facilitate discussion.
We suggest that in the ordinary Mendelian distribution the segregation
might be called amhilateral, effected on both sexual sides of the plant
as distinguished from the unilateral segregation which carries the
allelomorph wholly to one sexual side'.

In other examples, e.g. the double-throwing single in Matthiola, we

might speak of the segregation as partially unilateral. The exception

mentioned above (Expt. 11) may indicate that here also the unilaterality

may on occasion be partial, but unless the phenomenon recurs its

significance is doubtful.

Details of Experiments.

P = Procumbent. il/S = Male-sterile.

The "tall" strain is that first raised here by selection-; "Dark blue" is another
raised since ; "Dark blue tall" is another, now pure, raised by combination of the above.

Expt.

1. From P X white, 3 F^ plants gave in Fo

2. „ Pxtall, 4

2

53
308

MS

9

100

Totals

3. From tall x P 3 Fj plants gave in F^

4. ,, white xP 3 „ ,,

361

559
157

109

Total ... 716 —

5. MS X P

5 mothers X 3 fathers gave ... ... 101 2

The 2 MS resulting were doubtless due
to self-fertilisation

1 Proc. Ro7j. Soc. 1920. - Journal of Genetics, Vol. v. 1916, p. 199.

274

Male-Sterility in Flax

MS X various fibre flaxes.

MS X tall

Parents

used

Expt.

6.

?
15

16

7.

,, X dark blue tall

2

1

8.

,, X dark blue

5

3

9.

, , X white

5

3

10.

,, X Canadian fibre

1

1

Offspring

MS
402
57
72
77
32

Totals

28

24

640

11. In addition to these was the exceptional case in which ilf^^xa tall gave
1 $ and 1 MS

F.) plants from Expt. 2 were tested as follows :
Expt.

Two 5 s selfed gave F3

Two MS X two sister $ s (1 used in Expt. 12) gave

proving that these 5 s were heterozygous.
One ilfS X sister $ gave

proving that this ^ was homozygous.
5 5 s from Expt. 13 selfed gave
45s from Expt. 5 selfed gave
4 MS X sister ? s gave
2 ?s of complex origin (derived from P 5 ) gave ...

12.
13.

14.

15.
16.
17.
18.

§

MS

98

30

14

12

41

378

298

58

43

140
87
58
25

Adding families derived from P 9 , which are expected on selfing to give 3 : 1. There are

Kxpt.

5

MS

1.

53

9

2.

308

100

12.

98

30

15.

378

140

16.

298

87

18.

43

25

Totals ...

1178

391

Adding those expected to give 1 : 1

Expt.

9

MS

13.

14

12

17.

58

58

Totals ...

« . .

72

70

Expt.

19. Case mentioned, p. 273.

MS X 5 from Expt. 14 [viz. a plant which had received MS from
maternal side and hermaphrodite factor from the paternal
side] gave

MS
6

JOURNAL OF GENETICS, VOL XI. NO. 3

PLATE XXVI

Fi''. 1.

Fig. 2. Fig. 3. Fig. 4. Fig. 5.

Fig. 6.

Fig. 7.

Fig. 8.

W. Bateson and a. E. Gairdner 275

EXPLANATION OF PLATE XXVI.

Fig. 1. A row of male-sterile plants (left) growing beside a row of normal plants (right).
Figs. 2 — 8. Drawings of flowers with petals and sepals removed ; Figs. 2 — 4 and 6 from

blue-flowered plants ; Figs. 5, 7 and 8 from whites.
Fig. 2. Normal just before opening.
Fig. 3. Same later, anthers dropped.
Fig. 4. Male sterile, unopened flower, the usual condition.
Fig. 5. Barer condition of male-sterile, unopened flower. Stamens more developed,

occasionally bearing some pollen.
Figs. 6 — 8. Later stages of male-steriles. Anthers shrivelled but persisting.

THE INHERITANCE OF WING-COLOUR IN
LEPIDOPTERA.

VI. DIAPHORA MEN BIG A CL. AND VAR. RUSTIC A HB.

By H. ONSLOW,

Trinity College, Cambridge.

(With Ten Text-figures and Plate XXVII.)

The inheritance of several melanic varieties has been dealt with in
some recent papers, the melanic character being found as a rule dominant
to the type form. In the case of Odontopera bidentata and of Abraxas
grossulariata (var. varleyata), the black variety behaves as a recessive.
The first insect has no doubt a true melanic form, but in the case of
varleyata the existence of the white collar at the base of the wings
suggests a pattern factor rather than melanism. Melanism as a rule
takes the form of a general suffusion with black or very dark scales,
the pattern being sometimes so completely obscured that the insect
appears uniformly black.

The subject of this paper, D. mendica (The Muslin Moth), cannot
be included strictly among melanics. In England this insect is sexually
dimorphic, the male being dark brown or black, with a number of dis-
tinct black spots of very variable size, the female pure white with similar
spots (see Nos. 1, 2 and 11, 12, PI. XXVII). In Ireland, however, there
is a race, the males of which are nearly as pale as the females, but of
a creamy shade rather than dead white (see Nos. 9 and 10, PI. XXVII).
This race, var. rustica Hb., is also found in the Caucasus, Hungary,
Roumania, Silesia and Armenia. Experiments have already been carried
out in crossing the variety with the type form, by Caradja\ Standfuss^
Tutt^ Adkin^ and recently by Cockayne^ Tutt found, as did Caradja,

1 Societasentomologica, Zurich, Vol. ix. No. 7, p. 49, 1894—5, and Vol. x. No. 7. p. 49,
1895—6.

- Standfuss, Handbuch palnarkt. Gross-SchmetterUnge, 189(5, p. 223.

^ Tutt, J. W., British Lepidoptera, Vol. v, p. 45.

* Adkin, R., Entomologist, Vol. xxx. p. 206, 1897.

5 Cockayne, E. A., Ent. Record, Vol. xxxi. No. 6, p. 101, June 15, 1919.

278 The Inheritance of Wing -Colour in Lepidoptera

that the males of the Fy^ generation were a little darker than rustica.
Adkin described the hybrid males as dusky white, and Standfuss, who
obtained his pupae from Caradja, bred 17 F^ </(/'> the darkest of which
was much lighter than the lightest type. Caradja classified the Fi hybrids
under the names of var. standfussi (intermediate between dark grey
and milk white), var. mus (darker grey but lighter than type), var. clara
(pale dirty yellow), and vai'. mixta (pale as clara, but heavily spotted).
Nevertheless, 20 per cent, of the males could not be classified in any
of these categories. All these authors, however, agree in finding the
-^1 cT </ b^t little darker than those of var. rustica.

The fullest details are recorded by Cockayne, and he concludes that
" the results show no segregation of the dark brown colour of the type

form takes place, either in F^ or ^5 generations ; even with the small

numbers bred I think it safe to say that the pale colour of var. rustica,
and the dark brown of the type do not behave as Mendelian unit
characters." The present writer does not consider that sufficient evi-
dence is advanced to justify these conclusions, and, moreover, the
experiments recorded later suggest that segregation does take place,
though it is partly obscured by modifying genes which tend to make
the variation appear continuous. The 46 F^ ^^ recorded by Cockayne
vary from creamy brown to the same pale colour as rustica. In the
F2 generation, which in the event of segregation should approximately
contain 3 pale or buff insects to every black one, only 4 (/*(/ are recorded,
and one of these was described as a dark intermediate. Moreover, from
the data given, all the F.^ generation came from matings of the form
DD X BD or DD x DR (where D stands for the imperfectly dominant
factor for rustica. and R for the factor of the recessive black type
form). This may be more clearly seen from the pedigree given below.
Only two F. broods are recorded, namely (i) and (ii), the male parents of
which came from the F.^ generation, and are described as " pale rustica,"
thereby almost certainly having the constitution DD. The two female
parents came from the same Fo, family as the male parents, but being
of course white they might have had any constitution, DD, DR or RR.
Since, however, the 1\ ^ ^ recorded in brood (i) were as pale or rather
paler than rustica (one was said to be like var. mixta), the % must have
been DD or at least DR. In brood (ii), 6 of the cfcT recorded were as pale
as rustica, and the others (number not mentioned) were a pale creamy
brown. The male parent of this brood was, as has been pointed out, DD,
and therefore the female parent must have been DR, because the male
offspring were, as would be expected, partly riistica and partly buff {DD

H. Onslow 279

and DR). Thus we see that in the F.j generation no type insects could
be expected to appear in any circumstances. Moreover, of the 4 cfcT of
the Fo generation, at most one should have been a type. Surely this is
not sufficient evidence on which to deny the occurrence of segregation !

I'able showing the Relationship of the Insects bred by K. A. Cockayne.
Fi (Brood i) $ (DR) x pale ^ (DR)

2^9 {Brood iv) $ x ^ " pa.\e rnstica." o x ^ " p&le rustica."

? DD \ DD ?DR \ DD

F^ (Brood i) (Brood ii)

21 o (? " Pfl'le rustica." (\ ^ ^J "pale rustica,^^

the rest pale buff.

Experimental.

The strain of rustica used in the following experiments came from
Ireland (Co. Tyrone), and the type insects came either from Bexley, Kent,
or from Cambridge, where they were attracted by caged females. The
work was commenced in 1915, but owing to the disease which has been
recorded by every observer, the stock nearly died out on more than one
occasion. This disease, which only attacks the full-grown larvae, is said
by Cockayne to show no signs of being transmitted through the ova,
though as long as the majority of the families are affected, this must bo
a difficult point to decide. Cockayne moreover suggests that the infection
may be carried by the food, which was in his case nettles, since he
observed that neither the sterilisation of the sleeves nor the cages
prevented it.

At the commencement of these experiments, the larvae were fed on
plaintains grown in large earthenware pots. These proved so cumbersome,
and the incidence of the disease so severe, that in the following season
the young larvae were put in glass cylinders filled with willow. All the
cylinders were sterilised, and in order to prevent the spread of infection,
each brood was placed upon fresh clean paper while renewing the food-
plant, and the hands, brushes and everything used in feeding were also
sterilised. With these precautions, unless the disease was transmitted
through the ova, or carried by the food-plant, little infection should have
taken place. Nevertheless, some families developed the disease, but at
the same time it was possible to keep others completely free from it, so
that nearly all the pupae would emerge. In other cases the disease would
only appear at the last moment when the larvae were beginning to pupate.

280 The Inheritance of Wing-Colour in Lepidoptera

In these circumstances a large number would pupate successfully, but
next year only from 25 to 50 per cent, would emerge. When the disease
appeared some time before pupation, the case was usually hopeless, and
the family was thrown away to prevent further infection. From the limp
and disintegrating look of the sick larvae the disease appeared as if it
might be due to a proteolytic bacterium, but smears of the blood failed
to show any definite organism. It was possible, however, that the ex-
amination was not made at a sufficiently early stage.

The ^1 cTc/* hybrids from the first cross of rustica x type, like those
of previous observers, varied from cream to a dark buff (see Nos. 3 to 7,
PI. XXVII),but were sometimes milk white like var. rustica, or white with
a fleck or two of grey, especially on the hind wings (see No. 8, PI. XXVII),
but the range of colour was considerable. In order to classify these
intermediate forms for analysis, it was necessary to give them definite
colour values, which was done by means of the tintometer in the manner
previously described^ To match the various shades of grey it required
approximately the same number of units of blue, red and yellow. In the
case of the dark insects orange-red predominated, and in the case of the
cream and butf coloured insects, orange-yellow. The following table gives
the colour values of some characteristic insects :

Insect

Black

Red

Orange

Vello\

'18 W^ 3 (type)

6-4

0-4

0-6

—

'18 W^2 (type)

5-2

. —

0-8

—

'19 I ^ 15 (jPi hybrid)

2-5

—

0-5

0-2

'19 K (J 14 (Fi hybrid)

1-5

—

0-4

0-1

'16 Z) ^ 2 (rustica)

0-3

—

0-3

0-3

'16 £» (J 6 {rustica)

0-2

—

0-3

0-5

'16 D (J 7 (rustica)

0-1

—

0-5

0-6

Typical ? ?

0-4

—

—

0-2

Unfortunately there Avere a certain number of insects which had light
veins (cf. No. 6, PI. XXVII) and others which showed a central area paler
than the periphery (cf No. 3, PI. XXVII). It at first appeared very diffi-
cult to give such insects a value which would represent the mean colour,
but after several attempts the following method was adopted. One or
both wings were removed from the body and mounted at the centre of
a large cork, which was then fitted to the shaft of an electric motor. On
revolving the motor at a sufficient speed, the markings on the. wings
blend, as on the colour-wheel, into a uniform field, which can be measured
by the tintometer in the usual way. •

From the values in the table above it is clear that the figures for
orange and yellow are small enough to be insignificant, and since their
^ Onslow, H., Journal of Genetics, Vol. viii. p. 225, September, 1919.

H. Onslow

•281

inclusion would greatly coinplicatc the record, they have been omitted,
the black values alone being given. From these figures a number of
curves have been constructed, showing the distribution and percentage
distribution of the colour values, on the system previously adopted, to
illustrate the variation of the insects in the principal families of each
type of mating. The colours of the parents are shown by arrows, and in
addition to the curves, tables have been given which include every in-
sect bred. For simplicity, the male insects in these tables had to be
classified into three groups. The black group, consisting of insects like
the wild type, and the white or cream coloured group of insects like
var. rustica, obviously represent the DD and RR classes, but the BR
class of insects presents a difficulty. It includes all the grey or buff
coloured males of the F^ generation, and as these have been called var.
standfussi, that name might have served to designate the whole class. The
name is, however, misleading, because whereas buflf or grey males are
always intended, the name should include every insect of the DR class,
some of which, it has been pointed out, are so pale as to be indistinguish-
able from pure rustica (see families '16 J. and '19 5 in Fig. 1). Thus

10 20

Number of Insects

^ Q parents of
'16^, '19 7.'
and '16 Z>

Fig. 1 (cf. Fig. 2). Curve showing the distribution of the black values of the ^ ^
in family '16 D bred from rustica $ x rustica^; also the values of the rj r^ in
the two families "16 A and '19 B bred from (Fj) x rustica.

there may be some males, which, though they appear to be pure rustica,
are genetically heterozygous for the black pigment, the white form being
completely dominant. For this reason, the DD class has been called
"pure rustica" to distinguish it from the DR class, called ''rustica and
standfussi" which includes both standfussi and the heterozygous males
indistinguishable from rustica. It was also necessary that the classes

282 The Inheritance of Wing-Colour in Lejndoj^tera

should correspond to definite portions of the colour scale, and as far as
possible the following values have been taken :

Class of Insect

Black colour values

" Pure rustica" (DD)
inistica and standfussi (DR)
TjTpe{RR)

0 0— 0-5
0-5— 4 0
4-0— 6-0

In the table (p. 286) which shows the F.^ generation, the range of colours
is not quite the same, the whole scale being slightly shifted towards
the black.

The white variety is, as has been said, usually imperfectly dominant
but sometimes the dominance is complete. The heterozygous males are
grey or butf, and in the F^ and other generations, in spite of a considerable
variation, the accompanying curves, constructed from the colour values,
show that the white and black homozygous insects tend to segregate in
accordance with the ordinary Mendelian ratios. The variation of the
colour is most probably caused by the presence of a modifying gene or
genes, in much the same way as the very variable radiation of the black

60

1

1

-

50

1
1

40

-

30

-

20

-

10

^160

■•16 A'19 B

U

1

2

3

0

1 2

(J cJ parents of Colour Units

'16 A, '19 B and '16 D

Fig. 2 (of. Fig. 1). Diagram showing the percentage frequency of
the black values of the pure rustica ^ J in the last figure ; as
well as of the (J cj in the two (I<\) x rustica families '16 A and
'19 B which have been combined.

H. Onslow

283

forms of the allied species tiinlosoma luhricipeda, the Heligoland variety
zatima Cr., and var. intermedia Standf , the genetics of which have been
worked out by Federluy'.

to

J ^

T r

18 W

A

'19 K

TYPE Q O

10

20

30

(J parent of
19/

^ parent of
19 A'

40 50

Number of Insects
Fig. 3 (cf. Fig. 4). Curves showing the distribution of the black values
of the (J (J , from a wild type family, '18 W ; as well as the values for
the ^ (J in the two Fi families, '19 I and K. The value for $ $ of
any breeding is given by the line on the right-hand side.

From matings of rustica x rustica only rustica were obtained. Refer-
ence to Figs. 1 and 2 (Family 10 D) shows that the black values of
these "pure rustica' all fall round 0'2. Similarly, matings between wild
type insects gave nothing but dark males, the colour of which varies
considerably within certain limits. The extent of these may be clearly
seen from family '18 IT, Figs. 3 and 4.

The colour variation of several typical F^ hybrids is shown by families
19 K and 19 / (see Figs. 3 and 4, also Nos. 3 to 7, PI. XXVII). These
hybrids appear to be the same whether the male parent is white or black.
Standfuss and Caradja consider that the tnale has more influence in deter-
mining the appearance of the otfspring than the female, but this is
probably due to the comparatively small numbers they (obtained. In

1 Federley, H., Hereditas, Vol. i. p. 221, 1920.

284 The Inheritance of Wiug-Colowr in Le^ndoptera

7'ustica X rustica.
[BD X DD\

Imagines

Male Female

•' Pure rustica " rustica and Type

Family (PD) standfussi(DR) {KR)

'l&D 28 — — 21

'17 C 2 — — 2

'18£ 5 _ _ 5

Totals 35 — — 28

the following table the F^ hybrid (/•(/ have all been included under the
heading "rustica and standfussi." Many of these (/'(/' were pale, but as a
rule they were not quite so pale as "pure rustica," or the DR insects of
families '16.4 and 'IdB (see Fig. 1). The very dark ^ (Fig. 3) with a
value of 40 is so black that it is probably a type insect included by
mistake. The average value for these Fj insects is seen from Fig. 4 to
be not much above I'O, so that it may be inferred the dominance of
7'ustica is nearly complete.

rustica x Type.
[BB X RE].

Imagines

Male

Female

" Pure rustica "

rustica and

im

Family

Female x Male {DD)

standfussi

(l)lt)

'16 G

rtistica x type —

7

—

11

'16 H

>, X „ —

1

—

2

'16 F

type X rustica —

2

—

4

'16 —

Ova from Mr L. Newman —

3

—

1

'11 F

rustica x type —

1

—

3

'111

type X rustica —

1

—

3

'19 F

,. X >, —

2

—

4

'19 a:

,, X ,, —

27

—

42

'191

rustica x type —

9

—

13

Totals ... — 53 — 83

The F2 generation, obtained by crossing two F^ insects, might be
expected to give approximately 7^ustica, standfussi and type, in the ratio
1:2:1. As the table on p. 286 shows, this appears to be the case, except
that the DD class is rather too large, owing of course to the fact that it
must contain a few white insects, indistinguishable from rustica, which
are genetically heterozygous (DR). In these circumstances the DR class
only contains standfussi ; and the DD class can only be called rustica.
Although the totals approach the expectation, the numbers in the indi-
vidual families seem to vary widely, and the fact that the ratio is so

H. Onslow

285

closely realised must be due to there being sufficient numbers tu give
an average. A glance at Fig. 6 shows that the curve of the com])iiu'd
families divides fairly sharply into three classes, the DR class being about
twice as numerous as the other two. The individual families, however,
which are seen in Fig. 5, show wide divergences. This is probably due

^ parent
of '19 K

cJ parent
of 191

Colour Units

Fig. 4 (cf. Fig. 3). Diagram showing the percentage frequency of the black values
of the type ^ ^ in the last figui'e. The two t\ families have been combined.

to the variation in the colour of the parents, i.e. to the modifying genes
which shift the range of colour of single families up or down the scale.
Family '20 E is, for instance, shifted towards the dark end of the scale,
whereas family '20 B is shifted in the other direction, and has, moreover,
much the palest male parent. If the curve is considered as a whole, the
classes are all a shade darker than the values given on p. 282 ; for in-
stance the DD class of " pure rustica " extends as far as the black value
1"0. The extracted type males are many of them very dark, some being
like No. 2, PI. XXVII, black rather than brown, and darker than the
ordinary wild male. The extracted rustica are like the Irish race, but
some have a small patch of light grey scales on the milky ground some-
what lighter than No. 8, PI. XXVII, and others are more heavily spotted
than usual, with a black streak down the costa (No. 10, PI. XXVII).
Journ. of Gen. xi 19

286 The Inheritance of Wing-Colour in Lepidoptera

^.

X i^\. -

[Z)A'

xDJi].

f-2

Imagines

Male

A .

Female

Family

rustica (DD)

standfussi (DR)

Im

'17 D

3

—

8

'17 £

4

1

—

3

'17 G

—

2

'18 D

5

8

4

27

'20 A

7

15

17

29

'20 B

10

8

—

26

'20 C

2

8

—

10

'20 D

4

4

2

12

'20 E

7

13

14

37

'20 F

9

9

—

22

'20 G

2

11

3

13

'20 1

1

1

1

5

Totals

54 (31

7o)

78(45%)

41 (24 °U)

194

Expectation..

45

86

43

—

By pairing the F^ hybrids back to type insects, two fairly distinct
classes were obtained; insects as dark as type or darker with black
values above 40, and intermediate insects with black values consider-
ably below 4*0 sometimes even falling below 0'5. Reference to Figs. 7 to
10 will show how clearly these two forms segregate. In family '19 C
(Fig. 7) there is a gap of over two colour units between the RR and the
DR insects, and in family '18 A (Fig. 9) the gap is as large. In the other
families segregation is scarcely less complete. There appears to be no
difference in the colour of the offspring, whether the type insect is the
male or the female parent.

i^i X Type.

[DE X ER].

Imagines

Male

Female

' Pure rustica "

rustica and

im

Family

Female X Male

(DD)

standfussi (DR)

'17 B

Fi X type

7

2

8

'ISB

type X Fi

—

37

30

44

'18 A

Fi X type

—

21

23

42

'19 A

type X Fi

—

11

15

40

'19 C

', X „

17

18

30

'19 D

,, X ,,

—

32

41

89

'19 E

,, X ,,

—

5

5

16

'19 G

Fj X type

—

1

1

3

'19H

type X Fi

—

11

2

11

Totals

—

142(51°/o)

137(49%)

283

Expectation

—

139-5

139-5

H. Onslow

287

M
O

^O

»

1)

c

jS

^

o

^r>

CO

"^

■*^

o

a

n

O

^

as

a

J3

'^

a;

rn

^

(P

•43

<t>

s

a

o

•a

'^

■*

^

M

T3

J2

n

H

0

>

s

^^

c

CD

^s

tic

Cfi

fe

«

<W

0)

0)

u

t*-i

,c

>^

c

US

J3

0

6C

[x<

stiujj dnoioj

19—2

288 The Inheritance of Wing -Colour in Lepidoptera

s; 50

40

30

60

10

i c?c? 3 3

parent parents parents

of of of

'20 B '20 D& '20^,20C,

'206 '20 £

7 8

Colour Units

Fig. 6 (cf. Fig. 5). Diagram showing the percentage frequency of the black values
of the (J (J shown in the last figure, illustrating the segregation of the three
classes DD, DR and ER.

Av. value of rustica
parents oi'19A,E,H

(J parent of '19 0
(J parent of '19 D

10

60 70 80

Number of Insects
Fig. 7 (cf. Fig. 8). Curve showing the distribution of the black values of the
(J (J in five families of the cross (Fi) x type.

H. Onslow

289

When the F^ hybrids were mated back to pure rustica, most of the
offspring were as pale as the rustica parent, and a few were slightly
tinged or flecked with buff or grey, lis may be seen from the colour
values in families 16 A and 19 B, Figs. 1 and 2. There should of course
be equal numbers of " pure rustica " DD and DR, but owing to the im-
possibility of separating the two forms of rustica except by breeding

7 8

Colour Units

a <J ,^ parents of '19 D. b ^ parent of '19 C.
c Av. value of rustica ^ ^ parents of '19 A, E, H.

Fig. 8 (cf. Fig. 7). Diagram showing the percentage frequency of the black values of
the jj (J, in the families of the cross (Fj) x type shown in the last figure.

F^ X rustica.

[DR X DD].

Imagines

Male

Female

Family

rustica and stand,fussi
Female x Male (DD and DR)

Jm

'16^

Fi X rustica 36

—

22

'17 A

„ X „ 3

—

5

'17 if

,, X ,, 8

—

5

'17^

rustica X J^i 10

—

4

.'18 C

Fi X rustica 1

—

2

'19 B

rustica x Fj 28

—

22

Totals

60

290 The Inheritance of Wing -Colour in Lepidoptera

20

30

50

(J parent
of '18^

60 70.

Number of Insects
Fig. 9 (cf. Fig. 10). Curves showing the distribution of the black values of the ^ ^
in two families of the cross (Fi) x type. Note the two distinct plateaux in '18 B.

(J parent Colour Units
of '18 A

Fig. 10 (cf. Fig. 9). Diagrams showing the percentage frequency of the black values
of the (J (J in the two families of the cross (Fj) x type shown in the last figure.

H. Onslow 291

experiments, they have all been placed together in the column " rustica
and standfussi " (DD and DR) and the variation of colour must be judged
by reference to the curves.

There is one more point of some interest. Both Tutt and Caradja
consider that rustica is phylogenetically the older form, and in this
Cockayne agrees, believing that a black race arising in the centre of
Europe has spread in all directions replacing the old white race every-
where except in the extreme periphery of its habitat, as in Ireland and
the Caucasus. In the light of present knowledge this seems to be un-
doubtedly correct, for if evolution has taken place largely, if not entirely,
by the loss of factors, it is more probable that the white race rustica
lost a factor which inhibited the formation of black pigment in the male,
than that the black race lost the power to produce black pigment. If
the latter alternative were true, the white insect produced would be in
the nature of an albino, and as is invariably the case with albinism in
other animals, it would be recessive to the coloured form. It has, how-
ever, been shown that the white race is almost completely dominant to
the black race, that is to say, it behaves as a "dominant white," like
the English rabbit. Now dominant whites are supposed to contain a
factor inhibiting pigment production ; and extracts from the skins of
English rabbits have been shown to contain an inhibitor capable of
preventing the production of pigment by the enzymes from the skins
of black and other coloured rabbits \ whereas extracts from the skins
of albinos fail to do so. Thus the suggestion that the dominant form
is phylogenetically the older is probably correct.

Conclusions.

1. It is contended that the evidence advanced by Cockayne is not
sufficient to deny segregation in the Mendelian sense, in crosses between
D. mendica and its white variety rustica.

2. The experiments reported show that in the F^ generation the males
are not much deeper in colour than "pure rustica," so that this form
may be considered dominant to a great extent over the type; in fact,
some heterozygous insects are indistinguishable from mstica.

3. The great variation in the buff coloured hybrids is probably due
to the effects of modifying genes. The colour of these buff insects as well
as that of the others was measured in the tintometer. The construction

1 Onslow, H., Roy. Soc. Proc. B, Vol. lxxxix. p. 36, 1915.

292 The Inheritance of Wing-Colour in Lepidoptera

of distribution curves in the manner previously described enables the
segregation of these variable insects to be followed without difficulty.

4. Segregation is found to take place in the F^ generation, the pure
rustica and the black type males segregating from the heterozygous
buff coloured insects in an approximately 1:2:1 ratio. Owing to the
fact that some heterozygous males are indistinguishable from pure rustica,
the latter class is found to be too numerous.

5. The crosses between the F^ generation and type individuals also
show very clear segregation in the males, there being equal numbers of
type, and buff coloured or white insects.

6. The statement that var. rustica is phylogenetically the older
race is supported by a consideration of what is known concerning the
chemistry of black pigments.

I have to acknowledge my thanks to the British Association for
defraying part of the expenses of this research by a grant of £25 during
the year 1920, and I am also indebted to Professor Punnett for reading
the manuscript. I am particularly gi'ateful to Miss Helen Moodie for
her care of the larvae and for her efforts to prevent the spread of infection
while changing the food.

DESCRIPTION OF PLATE XXVII.

Diaphora mendica, var. rtistica and hybrids. Natural size.

1. D. mendica. Typical wild English ^ from Kent.

2. D. mendica. Black type F^,^, extracted from var. standfussi {Fi).

3. Var. standfussi. Dark F^ ^ from cross mendica x rustica, showing pale central areas

on fore wings.
Note the pale colour of the thorax and abdomen, a sure indication that the insect is
heterozygous.
4, 5. Var. standfussi. Dark Fi ^ ^ from cross mendica x rustica.

6. Var. standfussi. ^ from cross Fi x type, showing pale veining.

7. Var. standfussi. Fj ^ from cross mendica x rustica.

The colour of this insect, or a little darker, is about the usual shade oi Fi^ ^.

8. Var. standfusn. J from cross rustica x Fi , showing patches of grey scales, which are

not as a rule so dark.

9. Var. rustica. Typical white Irish ^J.

10. Var. rustica. White Irish ^ , showing black streak on costa.

11. Var. rustica. Irish $ , showing black costa which appears in certain families, in

which the black spots are often rather heavier than usual.

12. D. mendica. Typically English 9 from Kent.

JOURNAL OF GENETICS, VOL. XL NO. 3

PLATE XXVII

'lab., i"* **

10

.4.-

11

ir» i\*

12

THE INHERITANCE OF WING-COLOUR IN
LEPIDOPTERA.

VII. MELANISM IN HEMEROPHILA ABRUPTARIA
(VAR. FUSGATA, TUTT).

By H. ONSLOW,
Trinity College, Cambridge.

(With Plate XXV^Iir.)

The inheritance of the melanic variety of Hemerophila ahruptaria
(The Waved Umber) has ah-eady been studied by T. H. Hamling' and
E. Harris"^. Their experiments seem to show that the black form is
dominant, but the ratio of type to melanic insects, obtained by them in
the cross melanic (heterozygous for type) x type, appears to suggest
that the inheritance is not that of a simple Mendelian character. For
this reason the breeding experiments recorded in the following pages
were undertaken, and although they only serve to show that the surmise
was mistaken, they are nevertheless perhaps worth recording.

One point of interest is that the variety fuscata appears to be con-
fined to the environs of London, there being but one reference to its capture
in the New Forest where it is exceedingly rare. In its distribution this
variety is similar to some of the melanic forms already studied, in that it
does not originate in any of the manufacturing districts of the North of
England. The type insects used in these experiments came from Bexley,
Kent (Nos. 1 — 6, PI. XXVIII), and the melanic strain from Holloway,
London (Nos. 7 — 12, PL XXVIII). In colour the latter was a deep
chocolate brown, though the central band on the fore wings was not always
completely obscured. The depth of colour of this band varies considerably,
and appears to depend to some extent on the size and vigour of the larvae
(see Nos. 11 and 12, PL XXVIII), but there is no difference in colour
between a heterozygous melanic carrying type, and a pure homozygous
black insect. Of the two sexes the male is usually the darker.

1 Hamling, T. H., Trans. City of London Ent. Sac. 1905, p. 5.

■^ Harris, E,, Proc. Ent. Soe. London, 1904, p. Ixxii ; and ibid. 1905, p. Ixiii.

294 The Inheritance of Wing-Colour in Lepidoptera

The larvae were kept in the usual way in glass cylinders, and were
fed on lilac throughout the summer. They feed rather slowly, but when
ready to pupate, pieces of bark should be supplied, on which they can
make their cocoons. If more than about twenty insects were placed to
pupate in a single cylinder, a tendency was observed on the part of the
larvae to congregate in a small area, so that when finished, their cocoons
formed a continuous layer. As a consequence of this it was found that
where insects were surrounded by other cocoons they were often unable
to force their way out. It has been observed before that this species
will produce a partial second brood when in captivity, though it must
be rather unusual for an insect which as a rule hibernates in the pupal
state, to do so in the larval state as well. Insects that remain as pupae
during the winter begin to emerge as a rule in April or May, pairing
and laying freely. Some of the larvae resulting from these eggs will if
well fed pupate and emerge towards the end of July. These again
pair and lay, giving rise to a second generation, which, if properly fed,
and not kept too cold, often pupate before the end of the year.
Such insects emerge in May of the following year, rather later than
those insects which failed to emerge in July of the year before, and
remained as pupae throughout the winter. Privet was used to feed the
larvae after the lilac leaves had all fallen.

Very few records are to be found giving any useful information con-
cerning the genetics of this melanic variety. Pearce^ and Southey^ both
bred the melanic form, but their records are too deficient to be of much
value. The most detailed experiments are those of Hamling and Harris
already mentioned. Their data have been summarised by Tutt in British
Lepidoptera, Vol. v. but are very misleading, owing to no less than seven
errors, four of which concern the colour of the parents of the families
recorded. The following tables have been compiled from the original
records of Harris, and of Hamling.

Type X Type.

Imagines

Melanic Type

Bred by Male Female Totals Male Female Totali

T. H. Hamling (1904) — 1 1 12 6 18

E. Harris (1905) ... — — — 9 9 18

H. Onslow (1919) ... — - — 3 — 3

Totals 1 39

1 Pearce, W. G., Ent. Rec. Vol. x. p. 121, 1898.

2 Southey, W. A., Ent. Rec. Vol. x. p. 122, 1898.

H. Onslow

295

Melanic x Melanic.

Imagines
>■

Melanic

Type

Bred by

E. Harris (1905)
G. Porritt (1905)'
H. Onslow (1917)

Male

34

?

3

Female

34
?

1

Totals

68

20

4

Male

Female Totals

Total

92

Melanic (heterozygous for type) x Melanic (heterozygous for type).

Imagines

Melanic

Type

Bred by
T. H. Hamling (1904)
E. Harris (1905) ...

Male
31

Female

17
?

Totals Male
48 10
39 ?

Female
7
?

Totals
17
18

Totals ...
Expectation

87 (71°/ J
91-5

35(29%)
30-5

Melanic (heterozygous for type) x Type.

Ima^nps

Melanic

Type

Bred by

Female x Male

Male

Female

Totals

Male

Female

Totals

W. G. Peai-ce (1895) ^

Melanic x Wild type

? ?

?

3

?

9

4

T. H. Hamling (1904)

Melanic x Type

7

4

11

7

2

9

T. H. Hamling (1904)

Type X Melanic

4

4

8

7

1

8

E. Harris (1904) ...

Melanic x Type

5

4

9

6

5

11

E. Harris (1905) ...

Melanic x Type

19

15

34

7

8

15

E. Harris (1905) ...

Type X Melanic

8

16

24

3

3

6

Totals

89(63^

Vo^

53(37%)

Expectation...

71

71

It will be observed from these tables that in matings of type x type,
with a single exception which is probably an error, nothing but type
is given, and similarly, in matings of melanic x melanic, when one
parent is homozygous, nothing but melanic is given.

In the families bred from melanic parents, both of which are hetero-
zygous for type, the ratio of melanic to type insects is very nearly
3 : 1, as would be expected if melanism were dominant.

When a heterozygous melanic is crossed with a type insect, the ratio of
type to melanic should be one of equality, but as will be seen from the
figures in the last table, which includes crosses made both ways, i.e. the

1 Porritt, G. See J. W. Tutt, British Lepidoptera, Vol. v. p. 59.
Pearce, W. G., loc. cit. (see p. 294).

296 The Inheritance of Wing -Colour in Lepidoptera

female being sometimes melanic and sometimes type, the ratio is very
far from the expectation. In fact, there are 63 "/^ of melanic instead of
50 °/^, which is rather closer to a ratio of 3 : 1 than to equality. In order
to find out what the significance of this unusual ratio might be, a number
of matings were made, the offspring from which are recorded in the
following table :

Melanic (heterozygous for type) x Type.
First Broods.

Imagines

Melanic

Type

Family

Female X Male

Male

Female

Totals

Male

Female

Totals

H.Onslow (1917)

Melanic x Type

2

1

3

3

1

4

H.Onslow (1919)

Type X Melanic

5

1

6

4

1

5

'19^

,, X ,,

7

1

8

3

3

6

'19B

).. X ,,

2

3

5

3

1

4

'20a^

Melanic x Type

9

5

14

4

6

10

'20aB

,, X ,,

6

6

12

10

10

20

'20a(7

,, X ,,

2

5

7

2

2

4

'20 aZ) ...

Type X Melanic

4

3

7

4

4

8

'20aE

,, X ,,

4

5

9

6

4

10

'20bF

, , X , ,

4

4

8

4

9

13

Totals
Expeeta

tion

79(48
81-5

5°/o)

84 (51-5%)
81-5

In these experiments it seems clear that the factor for melanism
behaves normally, producing as many black insects as type. As however
the numbers were not very large, those insects which emerged in July
1920 were paired, so as to obtain eggs from which a second brood could
be raised. This was accomplished and the numerous larvae were fed
first on lilac and then on privet, but it happened that there was a very
severe frost in December, which caused the leaves of all the privet
shrubs to fall. For various reasons no attempt was made to give the
larvae, which were nearly full grown, any other food, and less than half of
them succeeded in spinning cocoons. Of these only about half pupated
completely and they emerged in May 1921, about a month after those of
the first broods, which had hibernated in the pupal state. The result of
these matings is given in the table on p. 297.

A glance shows that although the numbers are very small, the ratio
ot melanic to type is different from that given by the first broods in the
previous table, and in this the experiments resemble those of Hamling
and of Harris. When it is remembered that the mortality in the second
broods was over 75°/^, the explanation of this difference between the
first and second broods does not seem very far to seek. The inference
is that the mortality was selective, more melanic insects being able to

H. Onslow 297

Melanic

(heterozygous foi

V type) X

Type.

Second Broods.

Female

X Male

Imagine!

i

Melanic

Type

Family

Male

Female

Totals Male

Female

ToUls

'20 cM

Melanic

X Type

6

2

8

1

2

3

'20 cN

Type

X Melanic

2

2

4

—

3

3

'20 cO

Melanic x Type

3

6

9

3

2

5

'20 cP

Type

X Melanic

3

1

4

1

1

2

'20 cQ

X ,,

1

2

3

—

1

1

'20 cii

X ,,

—

1

1

_

'20 cS

X ,,

—

—

—

2

2

'20 cT

X ,,

1

—

1

'20 c 17

X ,,

3

2

5

1

1

'20 cF

X ,,

1

1

'20 cZ

X ,,

—

—

1

1

'20 cF

X ,,

1

—

1

1

—

1

Totals

37(66°/,)

19(34 7o)

Expectation

28

28

pupate successfully than type. That this was so, seemed to be indicated
by the fact that several of the type insects were too weak to emerge
and had to be assisted from their cocoons, whereas most of the melanic
insects emerged normally. Whether this will also explain the excess of
melanics obtained by Hamling and Harris is doubtful, especially as there
is no record as to whether these insects belonged to the first or second
broods. It does not seem improbable, however, that in their experiments
there was a mortality which favoured the stronger melanic insects at
the expense of the type. It should be mentioned that Southey^ is not
of this opinion, as in his experience the melanic larvae were the weaker.
If this explanation is correct it emphasises an important point which
has been mentioned before", namely the fact that many melanic forms
appear more robust and hardy than the corresponding type forms.
It is obvious that if a constitutional hardiness accompanies melanism,
selection would very soon cause these forms to replace type insects in
localities where the mutation has occurred. Other factors might assist
in bringing this about, but they would probably be much less important.
From this view less stress need be laid on the occurrence of melanic
varieties in manufacturing districts, and the numerous black forms that
have arisen in the rural districts of the south and elsewhere do not
appear anomalous.

1 Southey, loc. cit. (see p. 294).

- Onslow, H., Journal of Genetics, Vol. ix. No. 4, 1920.

298 The Inheritance of Wing -Colour in Lepidoptera

Conclusions.

1. The observations of Hamling and Harris seemed to show that
although the factor for melanism in Hemerophila abruptaria is dominant
to that for the type form, yet the ratio of type to melanic is not always
a simple Mendelian phenomenon.

2. Nevertheless, a repetition of their work shows a ratio of type to
melanic in the first broods, very closely approaching the equality expected.

3. In another set of matings, which produced the second broods,
the mortality for various reasons was over 75°/^. In this case, as in
the experiments of Hamling and Harris, the ratio was nearer 3 melanic
to 1 type than equality. This suggests that there may have been a
selective mortality in favour of the melanic insects, which disturbs the
normal ratio.

4. That a certain constitutional strength seems to accompany many
melanic forms has been pointed out by other authors, and attention is
here drawn to the fact that this would account more easily than other
hypothesis for the rapid replacement of type by melanic varieties.

In conclusion I have to thank the British Association for a grant of
£25 which helped to defray the expenses of 1920-1, and Miss Helen
Moodie for her help in the routine work. I am also indebted to Professor
Punnett for reading the MS. of the paper.

DESCRIPTION OF PLATE XXVIII.

Hemerophila abruptaria and ya,r. /uscata, Tutt. Natural size.

1, 2. H, abruptaria. Typical $ $ .

3, 4. H. abruptaria. Dark $ $, showing the same depth of colour as the typical ^.

5, 6. H. abruptaria. Typical ^ ^.

7, 8. Var. fuscata. Melanic $ $ , in which the ground colour and that of the band on
the fore wings are the same.

9, 10. Y&r. fuscata, Melanic <J ^, the band in which is slightly lighter than the ground

colour.
11, 12. Var. fuscata. Melanic ^^ from the second broods. These insects are smaller
than those of the earlier broods, and the central area of the fore wings is some-
what paler.

JOURNAL OF GENETICS, VOL XI. NO. 3

PLATE XXVIII

■>^ V

Vi?«k^-

'V^

^^^^^B

/^5

'' -spT*

\»k, '-;;j>»^' ^

li

^■"-^<X^

n

WW

mm

10

11

12

KEMAEKS ON THE METHOD OF CALCULATION ^
PEOPOSED BY ME H. L. TEACHTENBEEG FOE
DIALLEL CEOSSINGS.

By KIRSTINE SMITH, D.Sc.

From a mathematical or statistical point of view it may be argued
that an arbitrary assumption about the generative value of one of the
parents is formally unsatisfactory, although one must admit, consider-
ing the nature of generative values, which have no natural zero point,
that in reality it is of absolutely no consequence.

The remedy proposed by Mr Trachtenberg, however, seems to hie
worse than this slight lack of form. While professing to avoid an
arbitrary assumption it appears that the author only replaces it by an
unmaintainable assumption. He does not seem to notice that in
supposing that the personal values of parents represent their generative
values with the same degree of accuracy as that with which the generative
value of a group of offspring is observed from a single of its individuals
he really makes a new assumption and, as we shall see, a wrong assump-
tion. To realise this we only need to look at Mr Trachtenberg's results
comparing personal and generative values for parents and offspring,
from which I have calculated the following diflferences and their
squares :

Parents

Offspring

papers, -gener.

d,^

d^ = oha.-c&\c.

da^'

- -22

•048

+ •03

•0009

- -25

•063

+ •05

•0025

+ -17

•029

-05

•0025

-1-44

2-074

•00

•0000

+ -76

•578

+ •01

•0001

+ -10

•010

-•03

•0009

- -28

•078

+ •01

•0001

+ •01
+ •02

•0001
•0004

Total...

2-88

Mean value of d^

0^411

-•06
+ 03

•0036
•0009

■00

•0000

Total ^012

Mean value of dj^ ■ 0^0010

1 " The Analysis of the results of Professor Johannes Schmidt's diallel crossings with
trout." By H. L. Trachtenberg, B.A., in Journal of Genetics, Vol. xi. No. 1, 1921, pp 75—78.

300 Method of Calculation for Diallel Crossings

According to Mr Trachtenberg's supposition the squared standard
deviation of a personal value of a parent is 50 times that of the mean
for a group of offspring. Now we find here that deviations between
observed and theoretical values are on an average more than 400 times
as great for parents as for offspring. We must therefore conclude that
the supposition is wrong and reject it.

This result ought to be emphasised because it is a fundamental part
of the plan of the experiments with diallel crossings that the environ-
mental conditions are kept uniform for all groups of offspring, while
on the other hand, for all we know, the parents have developed in
quite different environments. Hence there is no reason to expect that
their personal values, apart from an unknown constant common to all
of them, should coincide with their generative values to the same order
of accuracy as we find they do for the offspring.

• While thus the fundamental idea of Mr Trachtenberg's method of
calculation is wrong it is only fair to admit that practically it does not
give so bad results, because the 7 parental observations weigh very
little as compared with the 12x50 observations of offspring.

The correct proceeding when introducing the personal values in the
calculation is of course to give to each of them a weight of only ^^ of
that of the means of a group of offspring. But then the ratio ^^ is
somewhat uncertain and has to be determined empirically.

It has therefore been found preferable to carry out the calculation
in two tempi. At first the generative values are calculated according to
the method of least squares, fixing an arbitrary value for one of the
parents, and secondly the arbitrary zero point of the generative values is
determined so as to make the squared differences between generative
and personal values a minimum. This method has been followed in
some of Dr Johannes Schmidt's publications concerning diallel crossings
with trout while in others the second part of the calculation has been
omitted as really unessential.

PRINTED IN ENGLAND BY J. B. PEACE, M.A.
AT THE CAMBRIDGE UNIVERSITY PRESS

CAMBRIDGE UNIVERSITY PRESS

Devonian Floras. A Study of the Origin of Cormophyta.
By E. A. Newell Arber, MA., Sc.D. With a Preface by
D. H. Scott, F.R.S. Demy 8vo. With a frontispiece and 47
text-figures. 17s 6d net.

Water Plants. A Study of Aquatic Angiosperms. By
Agnes Arber, D.Sc, F.L.S., Fellow of Newnham College, Cam-
bridge. Royal 8vo. With a frontispiece and 171 text-figures.
31s 6d net.

Organic Evolution. Outstanding diiBculties and possible
Explanations. By Major Leonard Darwin, Hon. Sc.D. Cantab.
Demy 8vo. 4s net.

The Influence of Man on Animal Life in Scotland.

By J. Ritchie, M.A., D.Sc, F.R.S.E. Royal Svo. With 91 illustra-
tions, including many original photographs and drawings, and
8 maps. 28s net.

The Concept of Nature. Tamer Lectures delivered in
Trinity College, Cambridge, November, 1919. By A. N. White-
head, Sc.D., F.R.S. Demy 8vo. 16s net.

An Enquiry concerning the Principles of Natural
Knowledge. By A. N. Whitehead, Sc.D., F.R.S.
Demy 8vo. 16s net.

An Introduction to the Study of Cytology. By L.

Doncaster, Sc.D., F.R.S., late Derby Professor of Zoology in the
University of Liverpool. Demy 8vo. With 24 plates and 31
text-figures. 24s net.

The Determination of Sex. By L. Doncaster, Sc.D.

Demy 8vo. With 23 plates. 12s net.

Menders Principles of Heredity. By W. Bateson,

M.A., F.R.S., V.M.H., Director of the John Innes Horticultural
Institution. Third impression with additions. Royal 8vo. With
3 portraits, 6 coloured plates and 38 figures. 15s net.

The Methods and Scope of Genetics. An Inaugural

Lecture delivered 23 October 1908 by W. Bateson, M.A., F.R.S.
Crown 8vo. 2s net.

Growth in Length: Embryological Essays. By

Richard Assheton, M.A., Sc.D., F.R.S., Trinity College, Cam-
bridge. Demy 8vo. With 42 illustrations. 2s 6d net.

Cambridge University Press

Fetter Lane, London, E.G. 4 : C. F. Clay, Manager

CONTENTS

All Rights reserved

PAGE

BuNGO MiYAZAWA. Dwarf Forms in Barley. (With Plate XX) . . 205

F. R, Parnell. Note on the Detection of Segregation by the examina-
tion of the Pollen of Rice. (With Plate XXI) . . . .209

Rudolph Beer, Notes on the Cytology and Genetics of the Genus

Fuchsia. (With Plates XXII— XXIV) 213

J. S. Huxley. Linkage in Oammarus chevreuxi 229

R. C. Punnett and M. S. Pease. Genetic Studies in Poultry. IV. On

the barred plumage of certain breeds. (With two Text-figures) . 235

E. S. Salmon and H. Wormald. A study of the Variation in Seedlings

of the Wild Hop (Humulus Lupulus L.). (With Plate XXV) . 241

W. Bateson and Miss A. E. Gairdner. Male-Sterility in Flax, subject

to two types of Segregation. (With Plate XXVI) . . . . 269

H. Onslow. The Inheritance of Wing-Colour in Lepidoptera. VI. Dia-
phora mendica CI. and var. rustica Wo. (With ten Text-figures
and Plate XXVII) 277

H. Onslow. The Inheritance of Wing-Colour in Lepidoptera. VII. Me-
lanism in Hemerophila abruptaria (var. fuscata, Tutt). (With Plate
XXVIII) 293

KiRSTiNE Smith. Remarks on the method of calculation proposed by

Mr H. L. Trachtenberg for Diallel Crossings 299

The Journal of Genetics is a periodical for the publication of records of original
research in Heredity, Variation and allied subjects. The Journal will also, from time to
time, contain articles summarising the existing state of knowledge in the various branches
of Genetics, but reviews and abstracts of work published elsewhere will not, as a rule, be
included. It will be issued in parts as material accumulates, and a volume will consist oi
three such parts.

Papers for publication may be sent either to Mr Bateson, Manor House, Merton,
S.W. 19, or to Professor Punnett, Whittingehame Lodge, Cambridge. Other com-
munications should be addressed to the University Press, Cambridge.

Papers forwarded to the Editors for publication are understood to be offered to the
Journal of Genetics alone, unless the contrary is stated.

Contributors will receive 50 separate copies of their papers free. A further 50 may be
had at cosu price : these should be ordered when the final proof is returned.

The Subscription price for a volume is 42s. net (post free) payable in advance ; separate
parts 15s. net each. Subscriptions may be sent to any bookseller, or to Mr C. F. Clay,
Manager, Cambridge University Press, Fetter Lane, London, E.C. 4.

Volumes I — X (1910 — 1920) are now ready. The price of each volume, in four parts,
stiff paper covers is 42s. net. Single parts can be supplied at 12s. 6d. net each. Quotations
can be given for bound copies of these volumes, as well as for supplying binding cases and
for binding subscribers' sets.

The Cambridge University Press has appointed the University o-f Chicago Press agents
for the sale of the Journal of Genetics in the United States of America, and has authorised
them to charge the following prices : — Annual subscription, ^ 8.50 net ; single copies,
$3.00 net each.

SPECIAL NOTICE. To effect economies in binding and postage a volume will, iu
future, be issued in 3 parts instead of in 4. The size of the volume will however remain
as before.

PKINTED IN ENGLAND BY J. B. PEACE, M.A.
AT THE CAMBRIDGE UNIVERSITY PRESS

5* r'A^ij

lrSw"S'

New Yorl. Botanical Garden Librar

■■■■■ |l'llllllllllll|H1!r

3 5185 00266 1880