•II

JOURNAL OF GENETICS

CAMBRIDGE UNIVERSITY PRESS

C. F. CLAY, Manager

LONDON : Fetter Lane, E.C. 4

LONDON : H. K. LEWIS AND CO., Ltd., 136 Gower Street, W.C. 1
LONDON : WILLIAM WESLEY AND SON, z8 Essex Street, Strand, W.C. 2
CHICAGO : THE UNIVERSITY OF CHICAGO PRESS
BOMBAY, CALCUTTA, MADRAS : MACMILLAN AND CO., Ltd.
TORONTO : J. M. DENT AND SONS, Ltd.
TOKYO : THE MARUZEN-KABUSHIK.I-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 VII. 1917— 1918

Cambridge :
at the University Press
1918 .

an

AIJ6A-

CONTENTS

No.. 1 (October 1917)

PAGE

Robert K. Nabours. Studies of Inheritance and Evolution in Orthop-

tera. II. (With Plates I and II) 1

Robert K. Nabours. Studies of Inheritance and Evolution in Orthop-

tera. Ill 47

A. W. Bellamy. Studies of Inheritance and Evolution in Orthop-

tera. IV. (With Plate III) 55

No. 2 (February, 1918)

A. B. Stout. Fertility in Cichorium Intybus : Self-Compatibility and
Self-Incompatibility among the OfiFspring of Self-Fertile Lines of
Descent. (With Plates IV— VI) 71

JoHS. Schmidt. Racial Studies in Fishes. I. Statistical Investi-
gations with Zoarces viviparus L, (With Plate VII and seven
text-figures) .......... 105

J. S. W. Nuttall. a Note on the Inheritance of Colour in One
Breed of Pigeons — An Attempt to Demonstrate a Mendelian
Type of Transmission . . . . . . . .119

W. O. Backhouse. The Inheritance of Glume Length in Triiicum

Polonicum. A case of Zygotic Inhibition. (With Chart) . 125

Mrs Oneba A. Merritt Hawkes. Studies in Inheritance in the
Hybrid Philosamia (Attacus)ricini(Boisd) <? x Philosamia cynthia
(Drury)9. ( With Plate VIII and two text-figures) . . .135

No. 3 (May, 1918)

D. W. Cutler. On the Sterility of Hybrids between the Pheasant

and the Gold Campine Fowl. (With Plate IX) . .155

C. W. Richardson. A Further Note on the Genetics of Fragaria . 167

J. E. Hull. Gynandry in Arachnida. (With one text-figure) . 171

M. C. Rayneb. Notes on the Genetics of Tettcrium Scorodonia cris-

pum (Stansfield). (With Plate X) 183

vi Contents

PAGE

Edith E. Hodgkinson. Some Experiments on the Rotifer Hydatina 187

Arthur W. Hill. The History of Primula malacoides, Franchet,
under cultivation, (With Plates XI and XII and one text-
figure) , . . . . . . . .193

Rainard B. RoBBiNS. Partial Self-Fertilization contrasted with

Brother and Sister Mating . . . . . . .199

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

in Poultry. I Inheritance of Leg-feathering .... 203

Edith R. Saunders. On the Occurrence, Behaviour and Origin of
a Smooth-stemmed Form of the Counnon Foxglove (Digitalis
purpurea) .......... 215

No. 4 (August, 1918)

A. St Clair Caporn. The Inheritance of Tight and Loose Paleae

in Avena nuda Crosses. (With six text-figures) . . . 229

A. St Clair Caporn. An Account of an Experiment to Determine
the Heredity of Early and Ijate Ripening in an Oat Cross. (With
one text-tigure) . . . . . . . . . . 247

A. St Clair Caporn. On a Case of Permanent Variation in the Glume-
Lengths of Extracted Parental Types and the Inheritance of
Purple Colour in the Cross Triticum polonicum x T. elohoni.
(With Plates XIII and XIV, three charts and one text-figure) . 259

Prepared by Ida Sutton. Report on Tests of Self-Sterility in Plums,
Cherries, and Apples at the John Innes Horticultural Institution.
(With Plate XV) 281

COEUIGENDA IN VOL. VI.
P. 197, line 10, /or +{\ -s) {{u + ^v + ^io] read ■\- {l - s) {lu + ^v + ^wy

„ 12, for I\=B- I "2i^ M + 2^, ^^ + "2?+^ wj

^ ^ , f 2'-+i - 1 1 2'-+i - 1 V*

read F,=sr ^-^^^ " + gm ^ + "2^+2- «'} •

P. 200, line 2, for ,/OZ,^_^!^ = J^^^
' ■> V 4 4-8 2(2-8) V 2(4-8)

/3 T' n /3(2-s)2

''^^ V 4 43^ «^2l2::8) = V nJ^S) '

" ^' •^'"'' v A "^i \/i '■^"^ s/re "^'i = vl •

Volume VII OCTOBER. 1917 No. 1

STUDIES OF INHERITANCE AND EVOLUTION
IN ORTHOPTERA. 11^

By ROBERT K. NABOURS.
(With Plates I and II.)

Due to the efforts of many able and earnest workere the approximate
end breeding results in numerous groups of plants and animals may
now be predicted. If the ancestry of parents for one generation be
known the characters and ratios of a resulting complex progeny may
be approximately prognosticated. As well also may the characteristics
of parents of unknown ancestry be analyzed by the appearance and
ratios of their offspring. Although there is an approach to agreement
in the matter of end results one finds wide diversity of judgment
regarding the fundamental causes.

Most, if not all Mendelians consider the proof of the segregation of
the germ plasm as an insulated substance in embryonic development
adequate. It is considered a fundamental principle that the units con-
tributed by two parents separate in the germ cells of the offspring
without having had any influence on each other. The ideas of the defi-
nite location in the chromosomes of the factoi-s giving rise to characters
and explaining Mendelian phenomena by means of the manoeuvres of
the chromosomes are thought to be satisfactorily supported by a large
body of evidence. (Morgan and students.) On the other hand the
idea of the insulation and continuity of the germ plasm as an entity
independent of the rest of the organism is seriously questioned. There
does not seem to be justification for the attempts to connect particular
factors with particular chromosomes or parts of chromosomes, and the
factorial hypothesis does not necessarily involve the assumption of
factors as distinct entities in the germ. (Child.) These diametrically
opposed ideas, each the consequence of extensive constructive experi-
mentation and consideration, serve to indicate the difficulties involved
in attempts to solve the problems of the mechanism of heredity, or the
physiology of heredity, or both.

1 Contributions from the Zoological Laboratory of the Kansas State Agricultural College
and Experiment Station. No. 10. The first instalment was published in the Journal of
Genetics, Vol. iii. pp. 141 — 170.

Journ. of Gen. vii I ,

2 Inheritance and Evolution in Ortlioptera II

There is much confusion regarding the use of the terms dominant
and recessive, the interpretation and application of ratios, and the
definition, or determination, of characters. The terms dominant and
recessive remain part of the nomenclature, as if they were realities,'
whereas they can have doubtful application only in crosses between
characters allelomorphic to each other in which one character is more
apparent (epistatic) and the other less apparent (hypostatic) ; or in case
of characters which are allelomorphic only to their absences, a character
being considered dominant and its absence recessive. Part of the con-
fusion in this matter has undoubtedly resulted, as will be shown later,
from the failure to recognize the distinction between the two classes of
characters. The 9:3:3:1 and 3 : 1 ratios are used freely and with
assurance, whereas there are in reality no such ratios. There appears
to be a lack, in usage at least, of appreciation of the distinction between
characters which are allelomorphic to each other, never to an absence,
and those characters which are allelomorphic only to their absences,
never- to each other or any other characters, and which exist only in
relation with, and in addition to, characters allelomorphic to each other. ,
These matters are considered in the following paper which presents
further records of end results and some applications of the breeding
work with the Grouse locust, Paratettix Bol.

The colour patterns of the pronota and femora of the jumping legs
are the characters considered. Observations have been continued on
the characters of long and short wingedness, but nothing further has
been ascertained than that these characters appear to be conditioned
by the environment, a tentative conclusion supported by evidence pre-
sented in my former paper.

The suggestion that these forms be designated as species seems to
have been premature and should be withdrawn ; not that it is thought
they may never be considered as such, but Orthopteran taxonomists are
as yet uncertain regarding the taxonomic position these forms should
occupy, and especially because this matter is not germane to the present
study.

No names are assigned to the true breeding forms, or hybrids. They
are designated by letters, the first eight being the same as before, with
six additional ones, two of them, LL and NN, not being figured. (The
G, BG and CG illustrated (Journ. Gen. Vol. ill. PL VI), but not other-
wise used, in the first instalment have subsequently been proven to have
been A A, AB and AC carrying homozygous doses of (i), respectively.)
Each letter means that the pattern for which it stands is a unit which

R. K. Nabours 3

cannot further be broken up, that the pattern is allelomorphic to each
of the others (all of them having been tried with several, and some with
all) ; in other words each letter represents one of a series of multiple
allelomorphs, fourteen being used. The Gre'ek letter 0 represents a
melanic pattern which can exist either in the single or double dose only
with, and in addition to, the true breeding patterns or their hybrids.
This pattern is allelomorphic only to its absence, if indeed it may be con-
sidered allelomorphic at all. It shows well with any of the multiple
allelomorphs, but perhaps best with BB, BG and CC where it is shown
both in single and double doses (heterozygous and homozygous). At a
given time soon after moulting, it can be determined whether most
of the forms are heterozygous or homozygous for S, but after a few days
those carrying only a single dose of © become melanic to the extent
that they are not readily distinguishable from those carrying a double
dose. The distinctions were not attempted except in a few instances,
because it would have required practically constant watch to make
the records at the proper time subsequent to moulting which occurs
irregularly after the third or fourth instar. However, the groups were
separated in matings (242), (244), (345), (246), (247), (254) and (255)
(see table at end of summaries on p. 40), and this is being continued with
extant cultures.

In most of the crossings between different pure types the resulting
hybrid is readily discernible from either parent, usually intermediate,
though in some cases the one pattern is more apparent than the other.
In all crosses between any of the forms and A A, the hybrid superficially
resembles A A less than the other parent form, and in some instances
cannot be distinguished by casual examination; e.g., the crosses AA x II,
AA X PP, and AA x SS produce progeny that can scarcely be discerned
from //, PP, and SS, respectively. However, this is only an apparent
dominance of /, P, and S and others over A ; for when careful exami-
nations are made it is found that the pigmental elements of each pattern
are present in the hybrid in about equal proportions (5).

In most instances where A and single and double doses of 0 were
involved I grouped together in the records the members of a progeny
superficially similar in appearance. On this account some groupings
were made that should give the apparent 9:3:3:1 ratio (matings
(175), (176), (177) {ABS x ABS), and (185) {AC^ x ACS)). Also on
this account there were from two to nine- possibilities, including reci-
procals, in the parents, and from one to four in the progenies (in
matings (315), (446), (220), (221) and others), where the four kinds

1—2

4 Inheritance and Evolution in Orthoptera IT

{AB®, A®B®, BBS, BSBS) so closely resemble each other that they
were grouped together, and their progenies not bred out.

On Plate II it is attempted at this time to illustrate only twelve of
the fourteen true forms (multiple allelomorphs) used in this paper and
thirty-six of the hybrids, and those forms and hybrids carrying 0. The
distinctness of patterns is nowhere overdone and in some instances
underdone. Some of the forms illustrated in my former paper appear
again, a few of them, e.g., EE, II, and some of the hybrids, much better
represented. The form QQ resembles CO greatly except for the red
legs which are conspicuous, but it behaves as a true form.- The form LL
(not illustrated) has three light brown stripes running longitudinally
along the pronotuiu, with the rest of the pattern somewhat mottled as
in most of the others. The form NN (not illustrated) has a rufous
brown, somewhat bronzed, pronotum as the striking feature of its
pattern. Along with LL and NN, many hybrids, and forms and hybrids
containing S, used in the matings are not illustrated.

The results are given in the form of tables. The arrangement is
such that the ancestry of the individuals of any mating may be traced
back, and the progeny traced forward, as far as there are any records.
In the matings the gametic composition of the parents and the somatic
composition of the offspring are given, and, unless marked (R), the male
is on the left and the female on the right. The first line of figures
(totals) gives the actual numbers and the second line the Mendel ian
expectation. The letters indicating the progeny are placed alpha-
betically, fi-om left to right.

At this time the arrangements of the tables will suffice for the
presentation of the data. Matings 1-159 show exclusively the behaviour
of multiple allelomorphs, a continuation of the first instalment. Matings
160-468 exhibit the behaviour of the character © (or the factor for it).
This part also indicates further the behaviour of the multiple allelo-
morphs; for there is no possibility of observing such a character, or
factor, as © except along with characters like A, B, G, D, etc., and
their hybrids, from which it is, by its nature, inseparable. (Two other
characters apparently of the nature of 0, one of them appearing as
white specks on the femora and anterior pronotum and inconspicuous,
the other a bright redness all over, head, body and legs, as well as
pronotum, almost as conspicuous as @, have been discovered, but the
forms bearing them are only now being bred.) The cultures of both
groups have been carried on simultaneously, and, as the cross references
indicate, individuals from one group were frequently used for matings
in the others.

R. K. Nabours

Table for Use in Making Pedigrees.

One or both of the parents of the numbers inside the parentheses came from the
number outside. This table, with the source numbers in connection with the matings,
enables one to trace the ancestry of any matings as far back, and the progeny as far for-
ward, as there are any records.

1 (2, 314); 4 (6); 6 (7); 8 (9); 10 (15); 11 (1, 198); 13 (187); 14 (6, 12, 18, 45, 46, 47,
48, 53, 82, 131, 231); 15 (68); 17 (3); 18 tl6, 54, 105; 338); 19 (112); 21 (115, 116, 117);
22 (25, 40, 189) ; 23 (69); 24 (52, 57, 58, 71, 292) ; 26 (74, 80, 106, 135) ; 27 (33, 53, 80,
145,146,147); 28(71); 29(64); 30(4,31,211); 31(152); 34(100,108,406); 36(115,
117); 37 (38, 197, 200, 201, 202, 203); 39 (11, 96); 40 (51); 43 (437); 44 (39, 40, 122,
186, 441, 444) ; 47 (37) ; 48 (41, 83, 144) ; 51 (81, 102, 123) ; 52 (77, 93, 334, 337, 372) ,
53(103); 59(92); 60(67); 68 (136, 137); 69 (91, 119, 206); 70 (73, UO, 194); 71(77);
72 (70, 76, 78) ; 73 (69, 84) ; 74 (75, 139, 142, 143) ; 78 (23) ; 85 (42, 43, 91, 287) ; 85 (38,
223, 224, 225, 349) ; 86 (76, 87, 348) ; 90 (283, 288, 289) ; 96 (42, 43, 95, 99, 155) ; 100 (107,
156, 279, 280, 408, 409, 412) ; 106 (93); 107 (109, 172, 407, 465); 108 (110); 114 (111);
115 (118); 116 (113, 114, 149, 428, 429); 119 (79); 120 (78, 141); 121 (267); 124 (434);
127(67); 129(138); 133(128,134); 135(130); 137(129); 140(29,92); 146(144,148,
464); 147(463); 149(150); 152(109,360,415); 154(110,151); 156 (32) ; 157 (18, 131,
132,133,135); 158(159); 165 (8); 166 (167) ; 169(5,161,162,163,164,168,170,177);
170 (172, 465) ; 173 (176, 178) ; 175 (1, 166, 173, 220) ; 176 (179, 181, 205) ; 178 (315) ;
179(180, 192,277, 405); 180(160); 181(195); 184(192,388); 186(248,249); 191(15,
348, 434); 192 (163, 193); 194 (175); 196 (236); 198 (242, 243); 205 (156, 168, 183, 374);
206 (207, 215); 208 (89, 183, 195, 196, 209, 317); 211 (342, 264); 213 (427); 214 (363);
216 (217, 218,219, 222, 224, 225, 226, 227, 228); 220 (221); 221a (239); 231 (464);
232 (210); 233 (268); 234 (132, 187, 231, 247, 359, 406); 235 (89, 270, 317); 236 (28,234,
265, 425); 237 (257); 242 (199, 208, 211, 235, 258, 278); 243 (199); 244 (223); 245(334);
249 (214, 250, 251, 358, 368); 256 (284); 258(264, 279, 409, 412); 264 (32, 213, 236, 407,
415) ; 265 (85, 244, 245, 256, 286, 344, 357, 365) ; 266 (63, 222, 371) ; 268 (66, 320, 380) ;
275(260,261); 276(280); 278(374); 279(408); 280(360); 282(25,64,188,229,233,

237, 283, 294, 295, 300, 305, 309, 354, 370, 375, 376, 377, 382, 389); 283 (49, 51, 268, 310,
316, 378, 431) ; 289 (299, 311, 381); 292 (88, 338, 369, 372, 463) ; 296 (291, 297, 313, 430,
431, 432); 300 (173); 303 (433) ; 305 (206, 221a, 316, 323, 324, 331, 453); 310 (66, 123,

238, 269, 321); 312 (62, 239, 240, 241, 304, 318, 325, 332, 385, 387, 391, 392, 393, 430,
452); 314 (235, 404); 315 (174, 196, 293, 322, 436, 438, 442, 443, 446); 316 (306, 319,
327, 328, 329,330); 317 (213); 320 (315); 324 (212, 325); 325 (314, 333, 445, 455);
331 (240, 390, 392); 333 (298, 326); 335 (50, 52, 104, 232, 254, 255, 263, 271, 272, 273, "274,
275, 281, 285, 290, 396, 397, 398, 400, 401, 402, 403); 339 (193) ; 341 (37, 158, 216, 281,
290, 340, 343, 344, 345, 346, 357, 396, 467, 468) ; 348 (44, 282, 373) ; 351 (352) ; 360 (395,
413); 363 (252, 253, 258, 262,270, 276, 277); 364 (198); 365 (200, 201, 202, 203, 210, 226,
227, 349, 366, 367, 371) ; 368 (363) ; 373 (94, 389); 380 (278) ; 382 (84, 221a, 364, 383,
884, 448); 383 (332, 385, 390); 384 (126, 391, 440); 386 (387, 393); 389 (457); 393 (184,
333,445,455); 395(394,428); 396(197,336,337); 397(350,399); 407(82,265,335,
341, 347, 355, 356, 411, 416, 417, 425, 426, 427) ; 408 (416) ; 411 (292) ; 413 (351, 353,
361, 362, 414, 423, 424, 429, 466); 416 (340, 410, 418, 419, 420, 421, 422); 417 (286);
425 (230, 246, 266, 343) ; 426 (263, 346, 397) ; 434 (439) ; 435 (312, 331, 379, 386) ;
436 (190); 439 (122, 125, 435, 441, 444, 447); 441 (296, 301, 302, 303, 307, 312); 444(257);
445 (388); 447 (288, 289); 448 (287, 449); 449 (437, 450, 451); 453 (454) ; 457 (456, 458,
459,460,462); 465(356); 468(228).

6 Inheritance and Evolution in Ortlioptera It

Explanation of the Tables. Matings of the same kind are grouped. The <? is ou the
left and the 9 on the right of the x, except when (R), which indicates a reciprocal,
is used. Beginning at the left, the (R), when used, equals a reciprocal ; the number in
heavy type (e.g. 3) is that of the mating; the next number, or numbers, indicate the
progeny; the last number, or numbers, in square type (e.g. 17) indicate the sources of
the parents, the c? being on the left and the ? on the right when two numbers are given;
(I in.) = from the cultures of the first instalment {Journal of Genetics, Vol. iii. p. 141);
(N) = from nature. The first line of totals gives the actual numbers and the second, in
italics, the expectation.

AA X AA

AA

1 38

175 11

a 5

1

3 17

17

4 7

N 3b

S 7

169 4

6 22

14

7 129

6

8 74

165

9 32

8

Total 331

AA X FH

AF

AH

lO

24

29 1 in.

Expect.

26-5 26-5

AB :

X AB

1

1

AA AB BB

11

1

1 39

IS

t 5

29 1

4

Totals

6

30

Expect.

9
AB

27
X AC

r
1

AA AB

AC

BC

13

9 14

20

17

N

R 14

30 32

23

15

N

Totals

39 46

43

32

Expect.

40 40

40

40

AB X AH

~"j

j

~r 1
1

15

1
AA

0

AB AH
2 1

BH
1

191 10

Expect.

1
16

1 1

AC X AC

1

1

1

1
AA AC

CG

5

LI

18

Expect.

AA
4

4
AC X AL

12

17

AC AL
3 1

I
CL

2

N

Expect.

2-5

2-5 2-5

2-5

AQ X BC

18

AB
12

AC
10

BQ
23

CQ

14 157 14

Expect. 14-7 14-7 14-7 14-7

AD X AD

19

Expect.

AA AD

12
16-7

I
DD

55 N
50-2

R. K. Nabours

1 —
1
AA

AD X

AJ

AI X

AP

Jd

' — r
1
AJ

1

1

DJ

AA

AI

AP

IP

20

5

3

i

3

N

27

8

9

12

U N

21

8

6

10

11

N

Expect.

10'7

10-7

lO-?

10-7

Totals 13 9 14 14

Expect. 12-5 12-5 125 12-5

J£ X AP

1

1

r
1

1

1

1
A A

AE EP

1
AP

22

2

3

3 r

Expect.

2

4
JJ X EI

^

AE AI

Expect.

23

4 1

2'5 2-5

N 78

(?) X AE

A A AB
24 42 52

.4£ BE

47 49 8

J£ X BC

1 N

25 16

Expect. 15

AG
15
25

BE
14
15

15
15

22 282

^7 X BB

Expect.

AB
28 21

24

— I
I
BI

27 N 236

^/ X CI

AC

CI AI

29

Expect. 6-2 6-2

AS X AJ

I
II

10 N 140
1^-5

30

i I i

AA AJ AS
111

AJ X SS

i
, L

JS
0 N

31 91

Expect.

75

AN X BS

I
JS

59 30 N

75

i ^^1

^-B AS BN NS

32 4 9 5 4 156 264

Expect. 5'5 5'5 5'5 5-5

26

Expect.

AI X AB

AA AB AI BI

19 17 21 12 N

i?-;S? 17-2 17-2 17-2

AP

X ^P

1
1

1

.JP P

33

18

51

Ixpect.

17-2

51-7

27

8 Inheritance and Evolution in Orthoptera It

AS X AC

I

^j ^

r I I

AA AG AS

34 23 13 9

Expect. 15 15 15

AS X AD

CS
15
15

i \ r~i

AA AD AS DS
36 8 11 20 8 N

Expect. 11-7 11-7 11-7 11-7

AS x'dS
AD AS SS DS

36 20

Expect. 23'5

46
47

BB X BB

28 N
23'5

BB
37 31 341 47

BB X GG

BG
38 168 37 85

BG X AN

AB AG BN GN

39 58 40 40 39 44 N

Expect. 44-2 44-2 44-2 44-2

BG X AP

AB AG BG GP
40 0 0 0 1 44 22

BG X BB

BB BG

41 7 13 48

Expect. 10 10

B

G ;

< BG

BB

BG

1

1

GG

42

1

11

7

96 84

43

14

27

9

84 96

44

14

26

13

348

45

8

7

6

14

46

15

24

14

14

47

1

0

0

14

48

24

57

32

14

49

13

29

13

283

SO

11

27

11

335

Totals

101

208

105

Expect.

103-5

BG X

1

1

207
GP

103-5

"1

BG BP GG GP
51 19 22 13 28 283 40
Expect. 20-5 20-5 20-5 20-5

BG X EP

I —
I
BE

52 25

Expect. 25

BP
34

25

GE
24
25

"1

GP

17 335 24

25

BG X IP

r-— 1 1

BI BP GI GP
63 12 12 16 14 14 27

Expect. 13-5 13-5 13-5 13-5

BQ X BQ

BB BQ QQ

64 14 20 17 18

Expect. 12-7 25-5 12-7

BB X BG

BB BG BE GE
65 7 7 6 9 I in.

Expect. 7-2 7-2 7-2 7-2

%

R. K. Nabours

BE X BI

i

"1

EE

BH X AS

1

1

BB

BE

AB

1 1

1 1

AH BS

— 1
HS

56

8

19

6

1 in.

68

1

5 2

3

15

N

67
58

4
5

18
13

14
6

24
24

Expect.

2-7

^r X AE

2-7

Totals
Expect.

17
BE

50

46-5

X EE

26

23-2

1

!

AB

1 1

1 !

AI BE

— 1
1
EI

69

8

8 7

10

73

23

BE

EE

Expect.

8-2

8-2 8-2

8-2

59

Expect.

16 7 1 in.
Uo 11-5

BF X BF

♦

BI X BB

72

1
1
BB BI

1 ■

FF

70

5 7

BB

BF

Expect.

6 6

60

0

1

0

1 in.

61

8

23

9

N

BI X BE

62

36
5

73
19

41

7

312
266

63

/«

1
1
BE BI

i
EI

Totals

49

lie

57

Expect.

55-5

111

55-5

71

12

15 10

13

28 24

BF X

CI

Expect.

12 5

12-5 12-5
BIx BI

i^-^

BC

BI

CF

FI

1

— ,

64 13

Expect. 9-7

12

9-7

8
9-7

6
9-7

282 29

72

BB BI
3 13

II
5

1 in.

BF :

K FF

73
74
75

11 31

11 46

2 5

14

27

3

70
26
74

BF

FF

Totals

27 95

49

65

6

5

1 in.

Expect.

^^•7 S-5-5

42-7

66

3

0

268

310

BI X CE

Totals

9

5

Expect.

7

BF X

1

7
FH

BC

BE CI

"1

£1

1

1

60 127

R 76
77

10
2

12 4
0 3

12

1

86
71

72

BF
67 6

BH
6

FF
2

FH

8

52

Totals

12

.12 7

13

Expect. 5-5

5-5

5-5

5-5

Expect.

11

11 11

iJ

10 Inheritance and Evolution in Orthoj^tera II

I —
I

BE

78 0

79 1

BI X EI

BI EI
1 1
4 3

"1

II
0
1

72 120
119

Totals 15 4 1

Expect. 2-7 2-9 2'7 2-7

BI X IP

r

BI

BP
80 4 5

Expect. 5'5 5'5

IP

5-5

— 1
I
II

5 26 27

5-5

86
87
88

Totals
Expect.

CE X CE

I

CC Olid EIl
8 6 3

4 12

I in.
86

12

17

14 292

Expect.

24 24 19

16-7 33-5 16-7

CF X BB

89

BG BE

17 29 235 208

23 23

CHxBI

BP X BP

81

Expect.

I
BB

17

19

BP
39
38

PP

20 51

19

CC X BC

I

BC

82 25

83 33

CC

30 407 14

28 48

Totals
Expect.

58
58

58
58

CC X BI

I

84

BC
4

Expect.

CI

4
4

382 73

CC X CC

85

CC
59

265

r

BC
90 11

Expect. 9

12

9

CI

7
9

CI X B£

2

BC

91 1

92 7 8

93 28 28

CE

3

6
31

"1

HI
6 N 70

EI
4 84 69
7 140 59

35 106 52

Totals 36 38
Expect. 40 40

40
40

46
40

CJ X CJ

I
CC

94 9

Expect. 9

CJ

18

18

"1

9
9

373

CN xB2^

BC

BN

1

1

CN

1

NN

95

23

23

25

16

96

R 96

13

9

6

11

39

Totals 36 32 31 27
Expect. 31-5 31-5 31-5 31-5

1

R. K. Nabours

11

CN X CN

CS X

BI

r T 1

CC CN NN

1

1

-\

BC

BS

CI

IS

97

4 9 3

N

106

5

8

4

3

98

6 12 5

N

Expect.

5

5

5

5

Totals 10 21 8

Expect. 9-7 19-5 9-7

GS X CS

34 26

CN >

NN

1 1
1 1

1

1

36

107
108

1 1
GG CS

9 31

48 77

SS

CN
99 12

Expect. 10

' NN
8
10

12 100
37 34

Totals
Expect.

57 108
53-5 107

49
53-5

CN >

. SS

CS X

JJ

CS

NS

1

lOO 44
R lOl 19

46
32

N
N

109
R IIO

CJ

7
3

JS
7 107 152

Totals

63

70-5

78
70-5

10 154 108

Expect.

Totals

10

17

GP X BP

Expect.

13-5

13-5

DD X DD

1

\

BC BP

GP

PP

1
DD

102
103

6 11
6 7

8
5

11

8

51
53

•

111 32

DD X .
AD X J

114

Totals
Expect.

11 18 13
15-2 15-2 15-2

19
15-2

DD,
DD or

•

GP >

< GP

DD X AD
DD or

AD and DD

CC C

P PP

112 40

19

104

19 1

9 16

335

Expect,

13-5 2
CQ

7 13-5

X CQ

DD X

DJ

R 113

DD
2

DJ

1

0 116

CC G
20

Q QQ

18

114

31

36 116

105

57

Totals

33

36

Expect.

19-2

o7-7

Expect.

34 5

34-5

12 Inheritance arid Evolution in Ortkoptera II

DJ X

AS

Fl
BF

^ X BP

1

1

1

1

AD

1
AJ

DS

JS

FF

115 42

32

43

39 21 36

123 8

10 310 51

Expect. 39

39

59

39

Expect.

9

9

DJ >

< DJ

FI

' X FF

1

1

FF

DD L

J JJ

124

33 1 in.

116

49 6

5 22

21

125

20 439

Expect.

34'li 68-5 34 -a
DJ X DS '

Total

126

3 384

FI

56

r

DD

1

1
1

1

X FH

DJ
10

DS
5

JS

12 21 36

j

117 9

1
1

Expect. 9

9

9

9

FF

FH

127 1

1 N

DS X

JS

S/Sf

Expect.

1
HH

1

DJ

DS

JS

X HH

118 0

0

1

3 115

HH

Expect. 1

1

1

i

128

18 133

EI X

BE

HI
129

I X SS

1

BE
119 7

1

BI
9

EE
11

1

EI

10 69

HS
37 137

Expect. 9-2 9-2

9-2

9-^

Hi

' X HI

EI

< EI

r

HH

]

1 ~

r 1
1

HI II

EE E

1
I 11

130

3

2 2 135

120

Expect.

16 3
25-5 3

0 16 1 in.

1 15-5

Expect.

1-7

3-5 1-7

FF>

< BG

1 in.
439 44

131
R 132

BH

4

12

HQ X

BG

BQ
11

4

GH CQ
7 2 157 14
9 6 234 157

121
122

1

BF

7

45

52

CF

8

35

43

Totals

Totals

Te"

15

16 8

Expect.

47-5

47-5

Expect.

13-7 13-7

13-7 13-7

t

ai

R. K. Nabours

13

HQ xHQ

HH HQ QQ

133 30 56 16 157

134 10 1 133

Totals
Expect.

31

26

56
52

17
26

Total

HS X BI

r

BH
135 21

Expect. 18-5

BS
14

18-5

HI
30

18-5

IS
9 157 26

18-5

HS X BS

BH BS HS SS
136 86 80 72 67 68
Expect. 76-2 76-2 76-2 76-2

HS X HS

I

I 1 1.

i I • I

HH HS SS

137 27 51 38

138 4 9 1

II X II

I
II

141 16

142 18

143 90

120
74
74

I
BI

144 3

Expect. 4-2

68
129

Totals
Expect.

31

32-5

60
65

39
32-5

AL

145 0

Expect, i

I
IJ

146 3
R 147 5

124

IJ X BG

BJ
6

4-2

— I —

I
CI

4

4-2

IP X AP

AP PP

3

2

IP X JS

IS
9
5

— \ —

I

JP

7

6

— 1
I
CJ

4 146 48

4-2

"1
I
IP

1
1

27

— I
I
PS

9 27 N
11 N 27

Totals 8 14 13 20
Expect. 13-7 13-7 137 13-7

139

Expect.

II X BI

BI

29
20-5

II
12

20-5

74

IS X IS

r-t-1

II IS SS
14S 18 5 146

Expect. 3-5 7 3-5

140

Expect.

II X CI

CI
10
11

II

12 I in.

11

149

Expect.

JJ X

1

1

DJ

1
DJ

1

1

JJ

3

4

3-5

3-5

116

14 Inheritance and Evolution, in Orthoptera II

151

Expect.

JJ X J J

I
JJ

ISO 12 149

JJx JS

I

JJ JS

23 20

^i-5 21-5

154

A A

X 449

1
1

1

AAQ

160

9

13

180

161

30

26

169

162

11

20

169

Totals

50

59

Expect.

54-5

54-5

AAQ X

AAQ

152
153
154

r

JJ

3

6

14

Totals
Expect.

JS X JS

I

— +—
I

11'

10
40

1

SS
6
5

25

23

30

61

60

NN X NN

36

30

155

NN
23

96

NS X AB

31

N

N

4iV
156 5

Expect. 4-<^

— I —

I
AS

4

4-2

BN
4
4-^

163
164
165

r'
I

AA

13
2
10

449 4949

54 169 192

1 169
34 N

Totals
Expect.

25

28-5

89

85-5

449 X AB

"1

4 100 205

4-2

166
167
168
169
170
171

r

AA

0
10

1
10
27
13

AB
1
4
4
12
11
15

AAQ

2

2

2
17
23
24

1

4E9
1 175

2

1

9

31

166

169 205
N
169

16 N

Totals 61 47_ 70 60
Expect. 59-5 59-5 59-5 59-5

%

QS X AH

449

X

SS

AQ
18

y-L-r-

157

4.S HS
21 20

QH
22 N

AS
172 6

ASQ
5

170 107

Expect.

20-2

20-2 20-2 20-2

Expect. 5 '5

5-5

SS X SS

1
SS

158 127

159 95

341
158

AQAQ

X

BF

ABQ
173 11

Expect. 9

AFQ

7
9

175 300

Tptal

222

R K. Nabours

15

174

3PCt.

AB

X BFBOFe

.•

r T

AB BB

32

30-7

AF

12
15-3

1

1

BF ABQ BBQ

T 1

AFQ BFQ

Exi

10 35
15-3 30-7

18 16 N 315
15-3 15-3

ABAeBOxABAQBe

1 ■■

i
AA

1

1
AB BB

26

AAQ AQAQ ABQ

1
1
AQBQ BBQ

1

1

BQBQ

175

25

49

194

176

0

6

4

19

173

177

1

11

14

48

169

Totals

2

43

43

116

Expect.

12-7

38-2

38-2

114-6

i 1 i

AA AB AF

178 0

Expect. 1-1

ABAQBQ X AFAQFQ

4
11

I . I I I I I I I I
BF AAQ AQAQ ABQ AQBQ AFQ AQFQ BFQ BQFQ

0 3
11 11

2
3-3

3

3-3

3

5-3

3

3-3

173

179
180

ABAQBQ X AS

I . I I I I I I I

AA AB AS BS AAQ ABQ ASQ BSQ

0 0 2.1 2 4 0 0 176 N

26 11 19 6 9 20 8 17 179 N

Totals 26 11 21 7 11 24 8 17

Expect. 15-6 15-6 15-6 15-6 156 15-6 15-6 15-6

AQBQ X AAQ, or
ABAQBQ X AQAQ

I 11 I

AAQ AQAQ ABQ AQBQ

181

Expect.

176

16 Inheritance and Evolution in Orthoptera II

AQBQ X AFAQBO, or
ABAOBe X AQFQ •

I

I I I I I I I I

ABAO AAG AQBO ABQ AQFQ AFQ BQFQ BFQ

182 2 2 3 3 N

Expect. 3-5 2-5 2-5 2-5

AQBQ X BB, or AC x CQFQ

BQBQ X BB J^

1

r
III.
ACQ AFQ CCQ CFQ

184 10 0 1 393

BBQ ABQ
183 19 205 208 Expect. OS 0-5 0-5 0-5

ACAQGQ X ACAQCQ

I I I I I I I I

A A AC CC AAQ AQAQ ACQ AQCQ CCQ CQCQ

185 0 3 9 14

Expect. 1-6 4-8 4-8 ' 14-6

ACAQCQ X BB

AB BC ABQ BCQ
186 2 1 0 3 N 44

Expect. 1-5 1:5 IS 1-5

ACAQCQ X BC ,

I I I I I I I I

AB AC CC BC ABQ ACQ CCQ BCQ

187 17 29 10 9 24 21 234 13

Expect. 13-7 27-5 13-7 13-7 27S 13-7

ACAQCQ X BF

t

. , I , . , ,

AB AF BC C'F ABQ AFQ BCQ CFQ
188 11 6 9 7 8 8 8 10 N 282

Expect. 8-3 8-3 8-3 8-3 8-3 8-3 8-3 8-3

R K. Nabours

17

AE X cce

AC CE ACQ CEQ
189 6 8 9 7 22 N

Expect. 7-5 7-5 7-5 7'5

AF X AFABFO

AA AF FF AAQ AFQ FFQ

lOO 0

Expect. 2

436

AFAQFQ X BFBBFe

I 1 r — n 1 r -^^ i \ ~i, 1 i

I I I I I I I II I I \-

AB AF FF BF ABB AQBS AFQ AQFO FFQ FOFO BFQ BOFB

191 5 2 0 5 18 16 N

Expect. 2-8 5-7 2-8 8-6 17-2 8-6

AS X ACAOCe

I
AA

192 9

Expect. 7-3

-~l i 1 — 1 1 1 1

I I I I I I I

AC AS CS AAQ ACQ ASQ CSQ

10 12 7 6 5 9 1

7-3 7-3 73 7-3 7-3 7-3 7-3

179 184

ASAQSe X ASAQSQ

193

Expect.

r"

AA

3

0-3

.1 I r I I I I I

AS SS AAQ AQAQ ASQ AQSQ SSQ SQSQ

0
1-1

0
I'l

3
3-3

192 339

BB X AAQ

I
, L_

AB
194 19
196 8

ABQ
16 70 N
10 208 181

BB X BFBQFQ

r

BB

Totals 27 26 196

Expect. 26-5 26-5 Expect.

Journ. of Gen. vji

BF BBQ BFQ
19 24 20 16 208 315
19-7 19-7 19-7 19-7

2

18 Inheritance and Evolution in Orthoptera II

BB X BPBQPQ

I

BE HP BBQ BPQ
197 12 11 17 9 37 396

Expect. 13-2 12-2 12-2 12-'2

BOBe X AS

211

Expect.

ABe
15
16

1

Bse

17 242 30
16

BB X cece

1

Bce

198

34

11 364

R 199

19

242 243

200

310

37 365

201

131 '

37 365

R 202

217

365 37

203

38

37 365

Total 749

BBe X A RAQBQ

! ! r-^ \ i 1

III I II

AB BB ABQ AQBQ BBQ BOBO

205 17 44 176

Expect. 15-2 45-7

Total

-

BRe

X BBQ

1

r 1 1

BB BBe BQBG

206

1

11 305

207

7

13 206

208

8

23 242

209

0

2 208

Totals

16

49

Expect.

16-2

48-7

BBQ

X CQCQ

1

\
BCe

BQCQ

210 42

32 232 365

Expect.

37

37

Total

BQBQ X BQBQ,

BBQ X BQBQ, or
BQBQ X BBQ

r \

BQBQ BBQ

212 89 324

213 1 317 264

214 4 249

215 40 206

134

BQBQ X BQBQ

I
BQBQ

216 171 341

217 106 216

218 172 216

219 150 216

599

ABAQBQ X

BQBQ,

BQBQ X

ABAQBQ,

BBQ X

AQBQ,

AQBQ X

BBQ,

AQBQ X

BQBQ,

BQBQ X

AQBQ,

BQBQ X

BBQ,

BBQ X

BQBQ, or

BQBQ X

BQBQ

1
1
lJ5e AQBQ

1
1
BBQ BQB^

220

45 175

221

.30 220

Total

75

R. K. Nabours

19

BBQ X Cece, or

BOBe X GCQ

I

Bce

321 a 1 305 382

Total

BBBQ X CG

• 223

224

R 225

I

Bce

104

91

114

244 85
216 85
216 85

309

Bese X BFBQFe

~i

BBS BOBQ BFQ BOFO

222 7
Expect. 7

7 216 266

BQBQ X cece

I

Bece

226 52 216 365

R 227 9 365 216

Total 61

BeBe X SS

. I

Bse

228 48 216 468

BCBece

X BC

1

1
BB

BC

i

1

CC

BBe

1

cce

229

15

13

5

3

21

17

282

230

27

46

12

10

30

15

425

231

2

10

10

8

15

1

234 14

R

232

1

4

4

4

5

2

335

Totals

45

73

31

25

71

35

Expect.

35

70

35

55

70

35

BCBece

x^C

1,

1

1

AB

11

AC

1

1

cc

BC
20

1

A Be

4

1 ■ 1

yice cce

Bce

233

3

29

8 229 N

234

8

15

10

3

16

7 236 N

Totals

19

18

30

7

45

15

Expect.

16-7

35-

5

16-?

J6-7

33-5

16-7

BCBece X AFAeFe

235

Expect.

I 1 1 1 1 1 ^-1 1 1 1 1 j

I I I I I I I I I I I I

AB AC BF CF ABe AeBe Ace Aece BFe BeFe CFe ceFe

6
3-1

3
3-1

1

3-1

1

3-1

7
9-3

14
9-3

9
9-5

9 242 314
9-3

2—2

20 Inheritance and Evolution in Orthoptera II

Bcsece X BBe

BE

1

1 ~l
BBO BBBO BCe BOCe

236 7

3

7

5

9

7 264

196

237 1

0

7

6

282

N

Totals

8

3

19

22

Expec

t.

6-

5 6-

5 i9-5

Bc/yece

X

1

BCBece

19-5

BB

1

1

BC

OC

i

BOBQ BBe

\
Boce Bce

1

cece cce

238

0

0

1

2

3

2

310

239

1

0

0

1

3

1

312 221 a

240

9

7

3

9

15

5

312 331

241

2

5

2

17

27

10

312

242

0

2

1

5

6

3

198

243

4

6

1

4

10

3

198

244

1

7

6

8

11

8

265

245

1

5

3

8

20

5

265

246

6

3

0

6

13

7

425

247

2

5

1

3

11

6

234

248

2

4

0

1

11

4

186

249

9

18

5

20

65

30

186

250

2

1

0

2

6

1

249

251

1

0

1

1

5

3

249

252

2

5

4

8

16

8

363

263

1

0

0

2

3

2

363

254

3

4

1

11

36

18

335

255

13

13

7

27

49

20

335

Totals

59

85

36

135

310

136

Expect.

4?-5

95

47:5

142-5
BCBQCQ

X

285
BF

142-5

1

\
1

1

1 1
1 1

1

BB

BC

BF CF

BBe Bce

BFe CFe

256

3

3

2 3

5 0

4

1 265 N

Expect

2-6

2-6

2-6 2-6

2-6 2-6

2-6 2-6

BGBece

X BFBQFe

1

T
j

T T

~~\

1
1

— 1 —
1

1

1

BB

B'G

BF

OF BJie BQBB

Bce Bece

BFB

i>e/''e CFe cofb

257 4

0

1

0

2

2

5

5 237 444

Expect. 1-1

J-1

1-1

1-1

3-5

3-5

3-5

3-5

R. K. Nabours 21

BCBOCe X BQCe

r T T"^^T T — 1

BBO BOBQ BCe BQCQ CCQ CQGQ

258 4 10 5 363 242

Expect. 4'7 9-5 4-7

BCBQCe X BP

T~T~T T'

~"T~i

BB BC BP CP BBO BCQ BPO CPQ
260 390 12 7272 275

R 261 35143323 275

Totals 6 14 1 16 10 5 9 5

Expect. 8-2 8-2 8-2 8'2 8-2 8-2 8-2 8-2

BCBOCe X BS

r~T~T~T"^T~T~T~l

BB BS BC CS BBO BCO BSO CSQ

262 5 3 3 6 13 2 363 N

. 263 89 233857 335 426

Totals 22 5 6 9 9 8 9

Expect. 17 8-5 8-5 8-5 8-5 8-5 8-5

BCBeCQ X BSBOSQ

BB BS

BC CS
4 0

1

BBe BeBQ Bse bqsq
5 3

BCQ

BeCQ

1

CSG

• CQSe

264 4

4

5 258 211

pect. 3-1

1-5 1-5

4-6 4-6

Bcsece X cce

4-6

36

4-6

BC
iS 2

1 1 1

CC Bce Bece

4 15

1

cce

cece

26

15 2

407

Expect

4-5

4-5 13-5

13-5

22 Inheritance and Evolution in Ortlioptera 11

BCBeCQ X CF

1

1
BC

1

BF

CG

CF

i^C'0

BFO

1

GGQ

CFG

266

22

9

0

10

13

2

25

13

425 N

267

0

1

2

1

3

2

1

0

N 121

TotalB

22

10

2

11

16

4

26

13

Expect.

13

13

i3

13

13

i5

13

25

BCBeCQ X CFCeFG

1

BG BF
14 4

CC '
6

■CF .

8

T T

Bce Bece

r T

BFQ BQFe (

1

j'ce c<

1

BGQ C

T 1

7^9 C'GFe

R 268

9

19

15

30 283 233

269

2 0

1

2

0

3

2

2 310

R 270

0 1

0

1

1

0

5

1 235 363

Totals

IG 5

7

11

10

22

22

33

Expect.

7-8 7-8

7-8

7-8

23-6

^5-6

^3-6

23-6

BCBOce y CP

BC

BP

T

CC

1
CP

iice

BPe

1

CCQ

cpe

271

18

8

5

1

6

5

7

12

335

272

10

3

1

4

3

1

3

9

335

273

25

22

"4

1

5

2

20

19

335

R 274

1

3

6

4

5

2

6

6

335

R 275

14

8

6

2

6

8

2

10

335

Totals 68 44 22 12 25 18 38 56

Expect. 35-3 35-3 35-3 35-3 35-3 35-3 35-3 35-3

Bece X Bece,
BCBece X Bece, or
Bece X BCBece

BQBQ BBS

Bece Bce

21
17

cece cce

276 8

Expect. 8-3

5 ;

8-5

363

Bece X Bs
BBe Bce Bse cse

277 37 28 27 25 363 179
Expect. 39-2 29-2 29-2 29'2

II. K. Nabours

23

BCQ Bece

278

Expect.

BQCe X CQFQ,

BCBQce X cere, or

BQCQ X CFCQFQ

242 380

BQCe X CS

we BSQ CCQ CS*

R 279 3
280 2

Totals
Expect.

.ve

1 100 258
6 276 100

BC X BSBQSe

r~T^T~T"^~T"~T~l

BB BC BS CS BBQ BCe BSO CSG
281 31525 2 54 335 341

Expect. 3-3 3-3 3-3 3-3 3-3 3-3 3-3 3-3

BC X CFCQFQ

r~"T~"T"-T T~T~T^1

BC BF CC CP BCQ BFQ CCQ CFQ
282 11 3562436 348

R 283 3 0 0 10 2 0 2. 282 90

284 2 13 5 4 4 4 5 3 256

Totals
Expect.

16
11

16
11

10
U

11
U

6
11

10
11

11

11
11

BC X CPGQPQ

285

Expect.

BC
23

BP
27

T

CC
21

T

CP
30

~V

BCQ
30

BPQ
24

CCQ

28

"1

CPQ
20

335

25-3 25-3 25-3 25-3 25-3 25'3 25'3 25-3

BC X CSCQSQ

286

Expect.

I —
I
BC

7

3-7

"T"

BS
4
3-7

CC
3
3-7

CS
4
3-7

BCQ
2

3-7

"T"

BSQ
4
3-7

CCQ
2

3-7

"1

CSQ
4 265 417
3-7

24 Inheritance and Uvolution in Orthoptera 11

BC X FFQ BC X FQFe

BF GF BFQ CFQ BFQ CFO

287 37 49 25 28 84 448 288 0 2 90 447

289 21 15 90 447

Expect. 34-7 34-7 '34-7 34-7

Totals 21 17

Expect. 19 19

BC X SSQ

Bs cs Bse cse

290 25 25 31 26 335 341
Expect. 26-7 26-7 26-7 26-7

BE X BEBOSe

BB BE BBQ BEQ EEQ EE

291 4 12 6 9 17 296

Expect. 6 12 6 12 12

BE X cce

r-rV

BC CE BCe CEO
292 6 11 14 8 24 411

Expect. 9-7 9-7 9-7 9-7

BFBQFQ X AC

r~T~T~r^T~T-T~i

AB AF BC CF ABO AFO BCO CFQ
293 3 71 4 3430 315 N

Expect. 3-1 3-1 3-1 3-1 3-1 3-1 3-1 3-1

BFBQFQ X BBQ

r~T~T~ T~T 1

BB BF BBe BeBQ BFQ BQFQ

294 7 2 16 10 282 N

Expect. 4-3 4-3 IS'l 131

R. K. Nabours

25

295

Expect.

BFBQFe X BC

BB BC BF OF BBQ BFQ BCO CFQ
79245346 282
55555555

BFBQFQ X BE

1

BB

BE

BF

r T

EF BBS

BEQ

1

BFQ

1

296

11

8

1

0 2

1

4

8

441 69

297

6

6

2

0 2

0

5

7

296

Totals

17

14

3

0 4

1

9

15

Expect.

7-9

7-9

7-9

7-9 7-9

7-9

7-9

7-9

BFBQFQ X BS

1

BB

BF

BS

T T

FS BBO

BFQ

B5:e

FSQ

298

2

0

1

1 0

1

1

2

333 N

Expect.

1

J

1

1 1

1

1

i

BFBQFQ X BFBQFQ

1

BB
3

BF

7

I
FF

T T

B2}e £050

T T T ]

BFQ BQFQ FFQ FQI
16 9

^0

299

7

289

800

11

13

1

29

58

32

282

301

0

1

0

0

3

2

441

. 302

1

1

0

3

5

6

441

303

1

0

1

1

0

0

441

304

7

5

4

19

11

13

312

305

14

24

4

35

69

40

282

306

6

0

4

5

0

3

316

307

0

4

2

7

4

4

441

Totals

43

55

17

106

166

109

Expect.

31

62

5i

Bj

BC

95
F50F0

X GC

1

186

95

CF BCQ

CFQ

309

31

7

3

19 282

Expect.

i5

i5

25

i5

26 Imheritance and Evolution in Orthoptera It

BFBQFe X GFCQFe

i

1

BG

BF
9

GF
9

FF

8

BGQ Bece

BFB

1

• BQFQ

1

CFe

1

) GQFO

1

FFe

1

1

1 ii^eFe

R 310 18

16

25

12

26 283

R 311 3

0

0

0

4

1

2

4 289

312 14

12

4

0

11

12

11

23 441 435

Totals 35

21

13

8

31

38

25

53

Expect. 14

14

14

14

4ii

4ii

4^

41i

BFBBFQ X EFEOFe

r~T^T~T~T~T" T~T- T~T~T~1

BE BF EF FF BEQ BOEQ BFQ B^FQ EFB EQFO FFQ FQFO

313 3 5

0 1

3 5

11

Expect. 1-7 1-7

1-7 1-7

5-^ 5-2

BQFe X A A

J

10-5

•

1 1
ABO AFQ

314 2 1 325 1

Expect.

15 15

296

BOFe X BFBQFQ, or
BFBOFQx BOFO

r~T~T T~t~l

BQBe BBQ BOFQ BFQ FQFQ FFQ

315 3 10 4 178 320

316 5 18 6 305 283

Totals 8 28 10

Expect. 11-5 23 115

BQFQ X BQBO,
BFBQFQ X BQBQ, or
BQFQ X BFBQFQ

BBQ BQBQ BFQ BQFQ

317 1 0 235 208

Expect. 0-5 0-5

R. K. Nabours ^7

BQFe X Bece,

BFBQFQ X BQCQ, or
BQFQ X BCBeCe

r~T~T~T-^T~T~T~l

BBe BOBO BCe BQCe BFQ BOFQ CFQ COFG
318 11 14 21 11 312
Expect. U-2 U-2 14-2 U-2

BFBOFe X BQFQ

BBe BOBQ BFQ BQFQ FFQ FQFQ
319 12 1 316

Expect. 12 1

BOFQ X i^eJFe,
BFBeFe X BQFQ, or
BQFe X BFBQFQ

BQBQ BBe

BeFe BFe

2

FeFe

FFe

320

0

1

268

321

i

1

1

310

322

18

35

12

315

323

3

4

1

305

324

2

4

1

305

325

31

46

33

324 312

326

3

4

1

333

327

26

88

31

316

328

5

21

2

316

329

10

19

8

316

330

6

13

6

316

Totals

105

237

97

Expect.

109-7

219-5

109-7

BeFe X cFceFe

r~T"~T~T~^T~T~T~l

Bce Bece BFe seFe CFe ceFe FFe FeFe

331 10 2 0 305 435

Expect. 0-7 0-7 0-7 0-7

28 Inheritance and Involution I71 Orthoptera II

BOFe X cere,

BFBOFe X GOFe, or
BQFe X CFOOFQ

i r r~T^ r r~i

Bece Bce bqfq bfq cqfo cfq fqfq ffq

332 9 17 16 16 312 383

333 12 0 0 325 393

Totals 10 19 16 16

Expect. 15-2 15-2 15-2 15-2

BP X BBCQ BP X CCQ

r~T^"T~i r^T^"T~"i

BBO Bce BPe cpe bc gp bcq cpq

334 14 13 9 6 52 245 335 27 35 29 25 N 407

Expect. 10-5 ' 10-5 10-5 10-5 Expect. 29 29 29 29

BPBQPQ X BPBQPQ

r~T~T~T~T^"T~T~"T~'l

BB BP PP BBQ BQBe BPe BOPO PPO PGPG

336 3 5 4 5 9 3 396

Expect. 1-8 3-6 1-8 5-4 10-8 5'4

BPBQPQ X CP

r^T^T~T"^T~r~T"l

BC BP GP PP BGQ BPe CPe PPQ
337 0 2 8 5 10 8 1 4 396 52
Expect. 4-7 4-7 4-7 4-7 4-7 4-7 4-7 4-7

BQ X BG BBGe

r~T~T~T^T~T~T~l

BB BG BQ GQ BBe BGe BQQ GQB
338 21 22 12 14 15 17 16 22 18 292
Expect. 17-3 17-3 17-3 17-3 17-3 17-3 17-3 17-3

Bs X AAe Bs X cece

AB AS ABO ASe BCe GSB

330 6 5 2 7 N 340 18 16 341 416

Expect. 5 5 5 5 Expect. 17 17

341

Expect.

BB
5

4:5

R. K. Nabours

BSBOSe X BSBOSe

T~T~T-^T~T^T— T— 1

BS SS BBQ BQBe BSQ BQSe SSe SQSQ
7 4 4 13 16 14 6 4 407

9-1 4-5 9-1 4-5 18-2 91 9-1 4-5

29

BSBQSQ X BSBOSe

1

BB'

BS

0
4

3
0

1

BQBe

1

4

BSQ Bese

3
10

1

sse

1

1

sese

842 1
343 3

1 211

7 425 341

Totals ' 4

4

3

5

13

8

Expect. 2-3

4-6

2-3

6-9

13-8

6-9

BSBeSQ X

00

00

X BFBeFe

r-

T" T^

Bc cs Bce cse

344 1 6 11 1 341 265

Expect. 4-7 4-7 4-7 4-7

r~T"^T~i

BC CF Bce CFe
348 5 4 2 6 86 191

Expect. 4-2 4-2 42 4-2

Bese X BB

CC X oeoe

1 1

1
cce

1 1
BBe Bse

349 209 85 365

345 32 21 341

Expect. 26-5 26-5

CC X OP0

Bese X BS

T 1

z?/>'e iise sse

CC

T i

cp cce CFe

350

14

22 18 18 397

Expect.

18

18 18 18

346 22 30 13 341 426

Expect. 16-2 32:5 16-2

CC X Gscese

1
cs cce cse

cc X BCBece

1 i

BC CC nee cce

1 —
CC

R 351

43

5 12 34 413

352

9

12 14 5 351

847 81 36 34 29 407
Expect. 32-5 32-5 32-5 32-5

Totals 52 17 26 39
Expect. 33-5 33-5 33-5 33-5

30 Inheritance mid Evolution in Orthoptera II

CO X sse
cs cse

353

E xpect.

413

354
355

cce X Bc

r"~T^~T~i

BC CO BCe GC9
20 20 16 '19 N 282
5 10 10 12 407

cce X JS

r~T- T~i

cj cs cje cse

R359 30 30 28 27 N 234

360 0 8 6 3 280 152

Totals 30 33 34 30

Expect. 31-7 81-7 31-7 31-7

cce X SS
I

R 361
362

CS
18
22

cse

10 413
26* 413

Totals 25
Expect. 28

30

28

26

28

31

28

Totals
Expect.

40

38

36

38

cce X

, 1

BS

BC

1

CS

BCQ --CSQ

356 3

11

7 5 407 465

357 37

38

30 30 265 341

Cece X BQBQ,
cce X BOBO, or

cece X BBO
Bce Bece

363

12

368 214

Totals 40 49 37 35
Expect. 40-2 40'2 40-2 40-2

CCQ X CCQ

cc ccQ cece

cece X cece,
cce X cece, or
cece X cce

cce cece

358

1

0-5

1
1-5

249

364
365
366
367
368

1

89

135

114

26

382
265
365
365
249

Total

365

CE X BCBece
BC BE CC CE Bce BEe cce CEe

360 12 6 6 6 6

Expect. 6-2 6-2 6-2 6-2 6-2

4
6-2

6
6-2

4

6-2

292

R. K. Nabours 31

CFCeFO X AC

I r~T~T" T"~T~T~1

AC CC AF CF ACS CCO AFQ CFQ

370 45 1 4 17 23 14 282 N

Expect. 26 13 13 26 13 13

CQCe X BF

BCe CFQ
371 42 48 365 266

Expect. 45 45

CECeEO X CE

CC CE EE CCe CEO EEQ
372 15 6 3 4 0 292 52

Expect. 2-3 4-7 2-3 2-3 4-7 2-3

CFCQFQ X A J

r~T~T~T"^T~T-

AC AF CJ FJ ACQ AFQ CJQ FJQ
373 41401301 348 N
Expect. 17 1-7 1-7 1-7 1-7 1-7 1-7 1-7

CFGQFQ X BE

r~T""^T~i

BC BF BCQ BFQ
374 5 5 5 3 278 205

Expect. 4-5 4-5 4-5 4-5

CFCQFQ X CCQ

r~T~T^-T— T~l

CC CF CCQ CQCQ CFQ CQFQ

375 15 4 22 22 282 N

Expect. 7-8 7-8 23-6 23'6

32 Inheritance and Evolution in Orthoptera II

376
377
378

GFGQFQ X GF

r~T~T^~T~T~"1

GC GF FF GCe GFG FFQ

47 35 12 6 33 25 282

9 18 7 4 10 7 282

18 25 7 2 25 16 283

Totals 74 78 26 12 68 48

Expect. 38-2 76-5 38-2 38-2 76-5 38-2

GFGQFQ X GFGQFQ

1.

r

1

GC

0

1
1

1
FF

0

GGQ

1

1 cece

1 1

CP^e CQFQ
0

1 i

Fi^e i^eFe

0

379

1

435

380

4

5

0

4

10

4

268

381

2

1

0

4

15

5

289

382
383

12

0

7
4

4
0

16
3

37
9

20
4

282
382

384
385

0

1

0
0

2
1

0
3

1
8

1
2

382
312 383

386
387

0
0

4
6

2
0

0

7

2
11

1
6

435
386 312

Totals
Expect.

19
14

28
28-1

9
14

38
42-1

88
84-3

43

42-1

Ci^CeJ^e X GQFQ

r~T~T T~T~1

GGQ CQGQ CFQ CQFQ FFQ b'QFQ

388

15 33

19 184 445

Expect.

16-7 33-5

GFGQFQ X FJFQJQ

i

16-7

' GF
389 2

"^T T r

GJ FF FJ
3 12

T T T T

GFQ CQFQ GJQ GQJQ
6 1

T T T [

FFQ FQFQ FJQ FQJQ

8 7 282 373

Expect. 1 S

1-8 1-8 1-8

5-6 B-e

5-6 5-6

GQFQ X GQFQ,
GFGQFQ X GQFQ, or

GQFQ X ci^cei^e

r~T~T^"T~T"1

cece GGQ GQFQ CFQ FQFQ FFQ

390

7

7

10

383 331

391

3

4

1

312 384

392

15

40

15

331 312

393

12

36

18

312 386

Totals

37

87

44

Expect.

42

84 -

42

R K. Nabours

33

cj X cQce

394

Expect.

cce

7
5-5

I
CJQ

4 395

5-5

395

Expect.

I
CC

CJCQJQ X cjceje

"T~T— r^T~T~T~T~1

CJ JJ CCQ CeCQ CJQ CQJQ JJQ JQJQ

3 la 13

10-1 20-3 10-1

30-5

79
61-1

19

30-5

360

CP X BBe

BG BP BCe BPQ
396 33 31 44 36 335 341
Expect. 36 36 36 36

CC
397 16

Expect. 14

CP X CCQ

"T-"^T~1

CP CCQ CPQ

13 14 13 335 426

14 14 14

CPCQPQ X BCBQCQ

nT~T"T-T~T T~T ~T~T~T~1

BC BP CC CP BCQ BQCQ BPQ BQPQ CCQ CQCQ CPQ CQPQ

398 12 0 1 3 3

Expect. 'i-i 11 11 11 3-3 3-3

2
3-3

6

3-3

335

399
400
401

4oa

403

I —
I
CC

1

0
0
0
0

CPCQPQ X CPCQPQ

"T""T"T"^T~T~T~T— 1

CP PP CCQ CQCQ CPQ CQPQ PPQ PQPQ

1
4
3
10
2

1»

5

0

6

2

0
10

2
23

7

2

10

2

34

15

Totals 1 20 14 42 63

Expect. 10-3 20-6 lOS 30-9 61-8

Joum. of Gen. vii

0

397

5

335

1

335

11

335

8

335

25

30-9

34 Inheritance and Evolution in Orthoptera II

CS X ABAQBB

AG

AS

BC

T T T

Bs Ace Ase

BCQ

1

404

0

3

0

3 0 2

3

6

N 314

405

2

3

6

8 7 15

8

6

N 179

Totals

2

6

6

11 7 17

11

12

Expect,

9

9

9

9 9 9
CS X ACAQGe

9

9

AC

,

1

CC

AS

CS ACQ CCQ

Ase

cse

4oe

17

7

9 27

4

6

34 234

Expect.

i:

■5

S-?

8-7 17-5

8-7

«•?■

CS' X BCBece

r""T~T~r T~T~T~1 ■

BC BS CC CS BCe BSe CCQ CSQ
407 1 2 32 2 2 63 107 264

Expect. 2-6 2-6 2-6 2-6 2-6 2-6 2-6 2-6

CS X BSBese

r~T~T~T T-T~T"n

BC BS CS SS BCQ BSQ CSS SSO
408 00000010 100 279

CS X CCQ

CC CS CCQ cse

409 5 4 4 3 100 258

410 6 5 13 11 N 416

411 21 28 26 25 407 412 5 5 100 258

CS X CQCQ

I I

c'ce CSQ

Totals 32 37 43 39 ~ Expect. 5 5

Expect. 37-7 37-7 37-7 37-7

CS X CSCQSQ '

r~T~T^-T^T^l

CC CS SS CCQ CSQ SSQ
413 20 30 19 16 30 15 360
R 414 2 11 4 4 11 1 413

Totals 22 41 23 20 41 16

Expect. 20-3 40-7 20-3 20-3 40-7 20-3

R. K. Nabours

35

cesQ X js

' T T ^

cjo cse jse sse

\

415

2 6 3 2 264

152

Expect.

3-2 3-2 3-2 3-2

cscese x cce

1 —
1
CC

T T T T 1

cs cce cece cse cese

' , ' *■ , '

tf 416

18

1 26 41

408 407

417

1

3 17 8

407

Totals

19

4 43 49

Expect.

U-3 14-3 43-1 43-1

cese X cese,

cscese x cese, or

cese X cscese

T "V J ^

cece cse cese sse ses

1 —

CCQ

re

418

11 38 14

416

419

3 9 1

416

Totals

14 47 15

Expect.

19 38 19

cscese x cscese

r~T~T~T~^T~T~"T~T-n

CC cs ss cce cece cse cese sse sese

420 0

4ai 2

422 2

1 0

1 1

4 0

Totals
Expect.

4

3-8

6

7-7

14

11-6

4

4

21

29
23-2

0 416
0 416
8 416

8
11-6

CS X sse

423

R 424

CS'

9

14

.1

SS
6
18

cse sse

9 9

17 25

413
413

cscese x be

r-
I
BC

425 7

Expect. 8

Totals 23 24 26 34
Expect. 26-7 26'7 26-7 267

Bs Bce Bse

8 13 4 407 236

8 8 8

3—2

I

36 Inheritance and Evolutio7i in Orthoptera II

cscesQ X BCBoce

r~T~T~T-T~T T~T^T~T^T— 1

Bc Bs cc cs Bce Bece bsq boso cce cece cse cese

426 0 4 1

Expect. 1-2 1-2 1-2 1-2

0 4 3

6

1-2 3-7 3-7

3-7

cscese x bqbo

" T T 1

Bce Bece Bse Bose

427 2 4 40*7

■ 213

Expect. 3 3

2

3-7

407

DJ X cjceje

r~T^T~T-^T~T~T~l

CD CJ DJ JJ CDS CJe DJQ JJQ

428

0

2

0

12 0

3

116 395

Expect.

1

1

i

1 1 1

DJ X cscese

2

^T~T~T T~T~T~1

CD CJ Ds js cue cje use jse

429 7 6 4 7 8 11 11 6 116 413

Expect, 7-5 7-5 7-5 7-5 7-5 7-5 7-5 7-5

EFEOFO X BC

±

BE BF CE CF BEQ BFQ CEO CFQ

430 11210 5 04 296 312

431 2 0 1 2 0 4 0 2 296 283

Totals 31330906

Expect. 3-1 3-1 3-1 3-1 3-1 3'1 3-1 3-1

I —
I
EE

432 1

Expect. 2'5

EFEOFQ X EFEQFQ
EF FF EEQ EQEQ EFQ EQFQ FFG FBEQ

4 0

5-1 2:5

11

7-6

14
15-3

11

7-6

296

R. K. Nabours

37

FF X BBQ

BF BFQ

433 2 1

Expect. 1-5 1-5

303

FF X BQFQ

\

1
BFG FFQ

434

1 3

124 191

Expect.

2 2

FF X CCQ

r-i-n

CF CFe

435 56 47 439 N

Expect. 51-0 51-5

FFQ X ABAQBQ
AF BF AFQ AQFQ BFQ BQFQ

436 6 4

Expect. 6-5 6'5

FFQ X BB

Expect.

437 8

7

BFQ
6 449 43

7

FFQ X BS

-T^r

L'F i^Sf 5Fe FSQ
438 9 9 4 14 315 N

Expect. 9 9 9 9.

FFQ

X FFQ

439
440

FF

0

1

FFQ FQFQ

10 434
. 2 384

Totals
Expect.

6
4-5

12

13-5

20

22 315 N

19-5

i9-5

2<^0Fe X 55

BFQ

■

441 80 439 44

FQFQ X 5050,

^'^0 X BQBQ, or
JTOF0 X 550

1
BQFQ BFQ

442 14 315

R 443 112 315

Total

126

FQFQ X BG

r "1

BFQ CFQ

444 20 17 439 44

Expect. 18-5 18-5

38 Inhei'itmice and Evolutio7i in Ortho2)te7'a II

FQFQ X

ceFQ,

FFQ X

CeFe, or

FQFQ X

r T

CQFQ CFQ

CFCOFQ

T 1

FOFO FFO

445 1

0 325

393

Expect.

0-5

0-5

FQFQ X

BQFO,

FFQ X

BOFO, or

FOFQ X

BFBOFe

r T

BQFQ BFO

446 22

21 . 315

Expect.

21-5

^i-5

FOFO

X FOFQ

FFe

X FeFO

FOFe

X FFQ

1 —
1

-—

— 1

FOFQ

FFQ

447

22

439

448

16

382

449

.11

448

450

45

449

451

56

449

452

13

312

453

8

305

454

50

453

455

3

393 325

Total

224

FJFQJQ X FJFQJQ

r~T-~T"T~T T~T~T^

FF FJ J J FFQ FQFQ FJQ FQJQ JJQ JQJQ

456 2 4 6

Expect. 1-5 3-1 1-5

4-6

5
9-3

0 457

4-6

FJFQJQ X FQJQ, or
FQJQ X FJFQJQ

r~T~T T~T~]

FQFQ FFQ FQJQ FJQ JQJQ JJQ

457 12 14 7 389
Expect. 8-2 16-5 8'2

FQJQ X FQJQ
FJFQJQ X FQJQ
FQJQ X FJFQJQ

FQFQ FFQ FQJQ FJQ JQJQ JJQ

15

5.

10

5 457
4 457
3 457

Totals
Expect.

22

16

30

32

12
16

U. K. Nabours

3d

nsHese x ss

1

r

HS

-rV

ss use

— 1
1
SSQ

461

2

4 5

4

N

»ect.

3-7

3-7 3-7

3-7

463

Expect.

I

I
BJ

42

31-7

48
5i7

JQJO X JGJG
jje X JQJe

joje X jje

CJ
29

I 1

I 1

jQje jjQ

462 40 457

jp X BCBece

'T^"T^T~T~1

CP Bje BPQ eve CFQ
27 28 28 25 27 147 292
31-7 31-7 31-7 31-7 31-7

PS X iiiJe

SS X cscesQ

BP

iJS iiPe BSQ

1

1

CS

"^^ T 1

SS CSQ SSQ

464 7

9 5 6 146 231 466 17

7 11 13

Expect. 6-?

6-7 6-7 6-7

Expect. 12

12 12 12

SS X ABAQBe

SS X sse

T 1
JSS ASQ Bse

1 "I

AS

-

ss sse

465 3

1 1 1 107 170 467

6 8 34

Expect, i-5

15 1-5 15

Expect.
SSe X SSQ

r - 1

ss sese ssq

7 7

468

25 78 341

Expect.

25-7 77-2

413

In a few matings separations of BBS from BSB%, BGS from B®GS,
and C(7H from G@G@ were made in the records. As has ah-eady been
indicated, this is a very tedious and difficult undertaking, as the dif-
ference between an individual homozygous for 0 and one heterozygous
for it is only slight at best. If the determinations are not made within
a few days after moulting it is necessary to group them, as was done in
most of the matings. Mating (341) {BS& x BS®) was separated with
special care, the result being shown in the table. The following
matings, grouped in the table, were thus separated, but the mortality

GC

BG

BBQ

Bej5e

Gce

CQce

BGO

Bee

1

2

4

1

2

1

3

3

G

7

6

2

4

4

8

3

3

5

6

2

5

0

15

5

0

3

5

1

3

4

9

4

1

5

2

1

6

0

8

3

1

4

5

6

10

8

23

13

7

13

19

8

11

9

34

15

40 Inheritance and Evolution in Orthoptera II

of the individuals mated from them was so great that the correctness
of the separations was not adequately proven by further breeding.

BGQ X BCQ

BB

242 0

244 1

245 1

246 6

247 2
254 3

256 13

Totals 2fi 19 39 47 21 41 26 100 46

Expect. 23 23 ' 46 46 23 46 23 92 46

Note. The mortality on the whole due to fungus and other disease, lack of proper

care, etc., was very great. However, recently, considerable improvement has been made.

Of the 974 matings made in this experiment only 470 produced young that could be

recorded. Of the 43,914 young transferred from the mating jars only 21,686 (those used

in this paper) were recorded.

The nature of this material permits us to observe the difference in
behaviour between the multiply allelomorphic characters, allelomorphic
to each other, never to absences, and the character ©, allelomorphic only
to its absence, never to anything. Application of this conception may be
made outside the Paratettix material ; the case of inheritance of combs
in fowls, for instance. Allowing P to represent the factor for pea, 8 the
factor for single, each allelomorphic to the other, and ^ the factor,
allelomorphic only to its absence, which, when present, modifies 8S
to make the two kinds of rose, and PP and PS to make the four
kinds of walnut, w^e have the same situation that is produced by B, G,
or any other two multiple allelomorphs and @ in Paratettix. We then
secure the same ratios, and exact parallelism in all respects. When
Mr Bateson (2), p. 63, describes a cross, rose x single, with the resulting
dominance of rose in F^ and the usual 3 rose : 1 single in F^, he really
shows a cross of single homozygous for the modifying factor <I> and pure
single {8^8^ x 88) which produces 88^. Then when these are inbred
there results in F^,

SS SS^ S^S^ S^S^ X SS Parents

12 1 I ^1

13 I

^ "* SS<P X SS<^

R. K. Nabours

41

Rose X Pea : (2), pp. 63, 64. This cross was undoubtedly PP x S^S^
which produced PS<^ (one of the four kinds of walnut) in F^. These
inbred would give F^,

SS PP PS

1 12

Single
1

Pea
3

SSi> S^S^
2 1

Bose
3

PP^ P<I>P* P^S^ PS^
2 12 4

Walnut
9

This situation is exactly paralleled by matings (198), (242), and (243)
in Paratettix:

Paratettix.

BB X cece

I (198)

Bce

34

I

BC Bece X BCBQCQ

BB

BC
2

CG

1

BBQ

BOBO

Bce Bece

QC<d

cece

0

5

6

3 242

4

6

1

4

10

3 243

Totals 4

8

2

9

16

6

2-8

5-6

2-8
2

8-4
9

16 -S

8-4

12

22

8-4

2-8

8-4

25-2

Combs of Fowls.

PP X S^S^

PS4> Fi

1

1
PSP^S^ X PSP^S'P

r 1

1 1

»P PS

|.. ...

1
SS

PP*

1

1 1
P^P^ PS*

1

P<I>S*

1
1
SS^ S-PS^

1 2

1

2

1 4

2

2 1

3

1

9

3

Pea

Single

Walnut

Bose

(See also matings at end of summaries, where those containing single
and double doses of 0 were separated.)

42 hiheritaiice and Evolution in Orthoptera It

The mating RF x S, (2), p. 65, could not have been other than
P/S^ X i^S which would produce

P^ SS PS^ and S8^

1111
Pea Single Walnut Rose

Another case is that of comb behaviour in experiments 33, 40 and 41,
(1), pp. 103, 106, 107. The $ in experiment 33 was well proven to
have been homozygous for the mo<lifying factor <E> which makes rose of
single (experiments 30-33) and her composition therefore should read
S^S*i> and the cT was single {SS). A pedigree arrangement of these
experiments would appear as follows :

? .S'4>S<I) X SSj (expt. 33)

20

I

h-

Same ? 8

1

SS^ X SS<P (expt. 40) ? s SS^ x SS ^ (expt. 41)

SS SS^ S#Sf*

SS ss^

79

196

68-75

206-25

Single

Bose

50

48
Single

46

48
Rose

These experiments 33, 40, and 41 of Bateson and Saunders are
exactly paralleled by matings (468) and (467) in Paratettix.

If P and S represent bhe factors for pea and single respectively and
^ the factor which modifies SS to make the two kinds of rose and PP
and PS to make the four kinds of walnut, a revision of Bateson's
16-square diagram, (2), p. 75, to correspond with the one for Paratettix
may be made :

P*

s^

s

PP

PS

pp^

PS^

PS

SS

PS*

ss^

pp^

PS^

p^p^

P*S<i>

PS^

ss^

P$S*

s^s^

R. K. Nabours 43

The essential difference between this conception and those prevailing
in most of the literature and practically all text books on the subject
is that two factors, P and S, which are paired, each with the other, and
one factor, <l>, which is not paired, unless it be with its absence, are
considered. Rose is a modified single and in no sense a character.
Walnut is a modification by the factor <I> of pea and the combination of
pea and single.

The same consideration may be given to the crossings of round
yellow with wrinkled green peas. Having in mind the behaviour of @
and any two multiple allelomorphs in ParateUix, it appears that Mendel's
experiment, (2), pp. 333, 334, consisted of one pair of factors, round and
wrinkled, each allelomorphic to the other, and a factor for yellow which
is allelomorphic only to its absence as © in Paratettix and <i> in combs
of fowls. Allowing R to represent the factor for round and W the
factor for wrinkled, the two making an allelomorphic pair, and A the
unpaired factor for yellow, we have a case of crossing RARA x WW,
with the product in Fi,R WA. These inbred produced in F2, considering
only those tested by Mendel,

RR RW

WW

RRA

RARA RWA

RAW A

WW A WAWA

35 67

30

65

38 138

60

68 28

33 66

33
30

66

33 132

66

66 33

102

301

96

99

33

297

99

Round

Wrinkled

Eound yellow

Wrinkled

green

green

yellow

Green is common to the albumen of peas. It may be that peas are
all homozygous for the factor, or factors, causing green, but the green
does not pair, as an allelomorph, with yellow. It does not matter, in
this consideration, whether the factor for yellow causes yellowness in
addition to, and to the obscurity of, gTeen, or whether it inhibits or
destroys something to prevent the green from showing.

It may be remarked, in this connection, that if all fourteen Paratettix
multiple allelomorphs, so far discovered, should be made homozygous
for ©, which could be easily accomplished, and all of them segregated,
there would be no chance of securing pure multiple allelomorphs again
from this group. As a matter of fact the forms with the double doses
of ^ would themselves become multiple allelomorphs. This may be
true now with respect to certain characters in Paratettix, e.g., the
triangular black spots on each side of the mid-pronotum which are
common to all forms. However, since all of them appear to have it

44 Inheritance and Evolution in Orthoptera II

there is no possibility of proving the matter, unless a form without it,
or only heterozygous for it, should be discovered.

There are two classes of characters, the one such SiS A, B, C, D, etc.,
in Paratettix, pea and single in combs of fowls, and round and wrinkled
in peas, and the other class such as B in Paratettix, <J> in combs of fowls
and A in colour of the albumen in peas. In the first class two factors
for any one suffice to make the whole character, and two for different
ones produce a hybrid, intermediate in fact, though the one may be
more apparent (epistatic) and the other less apparent (hypostatic).
Two factors for either A, B, C, or either of the others in that series,
suffice for a complete pattern ; one factor for one and another factor for
another make a hybrid. Two factors for pea make a complete pea
comb ; two factors for single produce a complete single comb ; while
one factor for pea and one factor for single produce the hybrid comb
(with pea epistatic, though this hybrid is said to appear somewhat
intermediate). Two factors for roundness produce the character of
roundness, two factors for wrinkledness produce a wrinkled pea; one
factor for wrinkledness and another for roundness produce an inter-
mediate in fact (with the round epistatic, though this hybrid is said to
be somewhat intermediate). In the other class (@, <1> and A) no one
can exist except in connection with, and in addition to, characters of
the first class. As already stated these factoids are allelomorphic to their
absences, never to any other character. For instance, in Pat^atettix,
when BB contains a 'single dose of @ (heterozygous for it), in the
gametogenesis, every other gamete receives a dose of @, half the gametes
formed carrying it and half without it. In neither case does the @ factor
affect the normal behaviour of the B gamete. The same conditions
prevail for hybrids, such as BG ; if @ be present in the single dose, only
every other gamete receives it, if @ be present in double dose, no gamete
is formed without it. In either case the gametogenesis in BG proceeds
as though @ were not present. (In the use of the word " dose " or
" fector " there is no intention at this time of attempting to convey any
conception of the condition or situation in the germ cells of the factor 0.)
It requires two gametes of any one multiple allelomorph to make a
homozygote for that form, or a gamete of one and a gamete of another
multiple allelomorph to make a hybrid or heterozygote. The situation
is entirely different with regard to such characters as ©, in Paratettix,
fp in combs of fowls and A in colour of peas; these are mere appur-
tenances of the gametes of the multiple allelomorphs.

This conception, if acceptable, has wide application. Besides applying

R K. Nabours 45

to colours of Paratettix, combs of fowls, and colours of the albumen of
peas, it as readily applies to all the inheritance behaviour exhibiting
the 1 : 3 (actually 1 : 2 : 1) ratios, and the 9:3:3:1 (actually
1:1:1:1:2:2:2:2:4) ratios described in the literature which I
have had time so far to examine. I suggest also that those matings
which result in the apparent 27:9:9:9:3:3:3:1 ratios actually
consist of one pair of allelomorphic characters, each allelomorphic to the
other and two non-allelomorphic or unpaired characters, of the nature
of ©, each allelomorphic only to its absence, and neither one allelo-
morphic to the other, nor to any other, character.

One multiple allelomorphic factor of the nature oi A, B, C, or others
of this series, and two non-allelomorphic factors of the nature of @, or
three factors of the nature of 0 and the factor, or factors, of the nature
of A, B, C, etc., unnoticed, or ignored, though necessarily present, proT-
duce the 1 : 1 : 1 : 1 : 2 : 2 : 2 : 2 : 4 (9 : 3 : 3 : 1) ratios. One multiple
allelomorphic factor and three non-allelomorphic factors like 0, or four
factors of the nature of 0 and the factor, or factors, like A, B, C, D, etc.,
unnoticed or ignored, will give the 27 : 9 : 9 : 9 : 3 : 3 : 3 : 1 ratios. In
no instance can more than one multiple allelomorphic factor (^, B, C, D,
etc.) be found in one gamete, or two in a zygote ; but any number of
factoi's like © can be accommodated in a gamete or zygote. The
behaviour of these factors as illustrated in the tables lends emphasis to
these distinctions which are of fundamental importance in studies of
inheritance.

During the last three years of the breeding work, and in the pre-
paration of the data and illustrations, Mr A. W. Bellamy has given me
the most valuable help. The initiation of the work at the Kansas
Experiment Station was made possible by the open-minded and helpful
consideration of Dr T. J. Headlee, now of the New Jersey Station. The
Adams fund has cared for part of the expenses, and State funds for the
balance. Director W. M. Jardine has given the most complete encourage-
ment throughout.

46 Inheritance and Evolution in Orthoptera II
EXPLANATION OF PLATES.

PLATE I.

BBQ and JSOBG, BGQ and 50(70, and GCQ and CQ)CQ were photographed in pairs,
respectively, on the same plates and were printed in pairs. All nine kinds are here easily
distinguishable.

The series of phenomena is here represented as due to the combinations of one pair of
allelomorphic characters, or factors, B and C, and a third character, or factor, 0, which is
not paired, or allelomorphic, except with its absence.

This conception is widely different from that which represents the phenomena as due
to the combinations of two pairs of allelomorphic characters or factors. According to
Mr Bateson (2), p. 65, GQCQ, or BQBQ might be considered one character of an allelo-
morphic pair.

PLATE M.

The top row (first twelve figures) represents the patterns which breed true and cannot
be further analyzed. The form AA is most abundant in nature, most of the others being
comparatively rare. Therefoi'e it is common to find the other forms hybridized with A A,
though they are sometimes found pure, or in combination with others than A A.

The eighteen figures (BE — JS) represent simple hybrids resulting from combinations
of the pure forms.

The remaining fifteen figures (BFQ — CV0) represent the pure forms and their hybrids
carrying the melanic pattern 0, heterozygously or homozygously, as indicated.

For an explanation of the three figures ISIS, BIS, and IIS see third instalment,
Journal of Genetics, Vol. vii. p. 47.

LITERATURE CITED.

1. Bateson, W., and Saunders, E. R. 1901. "Experimental Studies in the

Physiology of Heredity." Report to the Evolution Committee of the Royal
Society.

2. Bateson, W. 1909. MendeVs Principles of Heredity. Cambridge University

Press.

3. Child, C. M. 1915. Senescence and Rejuvenescence. University of Chicago

Press.

4. MoRCi AN, T. H., and others. 1915. Mechanism of Mendelian Heredity. Henry

Holt and Co.

5. Nabours, R. K. 1914. "Studiesof Inheritance and Evolution in Orthoptera. I."

Journal of Genetics, Vol. in. No. 3, pp. 141 — 170.

JOURNAL OF GENETICS, VOL VII. NO. 1

Gametes

- B c Be

PLATE I

ce

I

Bce

Bce

cce

Bce

cce

BBBe

BBCe

Bece

cece

JOURNAL OF GENETICS, VOL VII. NO. 1

PLATE II

STUDIES OF INHERITANCE AND EVOLUTION
IN ORTHOPTERA. HP.

By ROBERT K. NABOURS.

In the first instalment of this study, Journal of Genetics, Vol. ill.
pp. 141 — 170, the crossing of the hybrid (/ GE with a hybrid % BI,
Table III (b\ was reported which produced the approximate normal
expectation of BG 12, BE 11, GI 7, and EI 10, and an unexpected
individual, a male, showing the combination of the three patterns B, E,
and / {BE I). This aberrant individual was discovered at an early age
and its pattern was clearly marked at all times, and especially well soon
after moulting when patterns are always at their best. It became adult,
but was lost before any matings were made. It was suggested in the
report that perhaps the female had produced an abnormal gamete con-
taining the factors for both B and /, and that this was fertilized by a
normal gamete from the male containing the factor for E^.

Explanations of the causes of this phenomenon have been proposed
by Dexter ('14), and Bridges ('16), and comments have been made upon
it by Castle ('14), and Morgan ('14).

No other such individual was observed in any of the Paratettix
cultures until the summer of 1915, when a form combining the patterns
of B, 8, and a modified / appeared among the progeny of a mating
(472) of 18 X BS. This aberrant BI8 individual, a male, was mated to
three females, BG (mating 486), B8 (mating 488), and BB (mating
482), respectively. Reference to the accompanying tables and to the
tables in the second instalment will indicate the ancestry of the parents
in this mating (472) and also show the breeding behaviour of the
progeny and relatives. (See Plate II in second instalment.)

1 Contribution from the Zoological Laboratory of the Kansas State Agricultural College
and Experiment Station, No. 11. .

The first instalment appeared in the Journal of Geneticx, Vol. in., and the second in
Ibid. Vol. VII. pp. 1 — 46.

2 The other apparent exceptions referred to in the first instalment were those cases
involving the factor 9, which are considered in the second paper.

48 Inheritance and Evolution m Ortlioptera III

Table for Use in Making Pedigrees.

One or both of the parents of the numbers inside the parentheses came from the
number outside. This table, with the source numbers in connection with the matings,
enables one to trace the ancestry of any mating as far back, and the progeny as far
forward, as there are any records.

470 (471, 472, 474); 472(473, 479, 480, 482, 486, 488, 489); 473 (481, 490, 495, 496);
475 (476, 477, 478, 490, 504, 505, 506, 531, 532) ; 479 (494) ; 480 (481); 482 (483, 484, 485,
497, 498, 499, 502, 503, 528); 485 (487, 501, 512, 513, 536, 539) ; 486 (484, 509, 510, 515,
516, 517, 522); 488 (498); 489 (504, 505, 507, 508, 519); 491 (518); 492 (506, 508);
493 (519); 502 (518, 530) ; 503 (529, 531, 532, 533, 534, 538); 509 (525); 610 (487, 500,
514, 523, 524, 527, 529, 530, 535, 536, 540) ; 515 (500, 511, 520, 521, 523, 535) ; 522 (512,
513, 526, 537, 538, 540).^

Explanation of the Tables. Matings of the same kind are grouped in summaries.
The (? parent is on the left and 'the ? on the right of the x , except when (R) is used,
which indicates a reciprocal. Below the line, beginning at the left, (R) when used = reci-
procal ; the number in heavy type (e.g. 470) is that of the mating ; the next number, or
numbers, indicate the progeny; the last number, or numbers, in square type (e.g. 26),
indicate the sources of the parents, the <r being on the left and the ? on the right when
two numbers are given. The numbers in the sources less than (470) refer to the tables
of the second instalment. The first line of totals gives the actual numbers and the
second the expectation.

TABLES.
CS X BI CI X IS

470

1

!

-^i

1

1

T T

T

1 i

BC

BS CI

IS

CI

CS

// IS

470 10

9 3

9 407 26

474

18

20

23 26

Expect. 7-7

7-7 7-7

7-7

Expect.

21-7

21-7

21-7 21-7

BC X CC

BC CC
36 39

37-b 37-5

II X

1

II

1
II

*143

90

74

471

Expect.

470 426

475
476
477

19
29
69

143
475
475

478

28

475

IS xBS

Total

230

I
BI

472 9

473 18

BS

12

IS

7

19

SS
10
17

"I

BIS
1 470
0 472

Totals 27 20 26 27 1

Expect. 25-2 25'2 25-2 25-2 0

CE X BI

BC BE CI EI

479 12 22 20 11 292 472

Expect. 16-2 16-2 16-2 16-2

R. K. Nabours

49

IS xBI

TABLES (continued).

BS X II

I
BI

480 11

481 24

BS

9

10

II

16
12

IS

7 472
10 480 473

Totals 35
Expect. 24-7

19

24-7

28
24-7

17

24-7

BIS X BB

~i

482
483
484
485

BB
20
10
10
22

BIS

24

15

9

31

472 245
482

486 482
482

Totals
Expect.

62
70-5

79
70-5

BIS X BG

490

Expect.

r

BI
30
30-5

IS

31 473 475

30-5

CJ X CJ

I

491

Expect.

CC
0
1

CJ
3

2

JJ
1
1

110 144

492

Expect.

I

I
DJ

1

1-5

"T

BS

JS X DJ

_i

3
1-5

T
JJ

3
1-5

1

JS

0 151 149
1-5

BB BC BIS CIS

486 9 8 10 9 472 266

487 9 9 13 7 485 510

BP xBP

Totals 18
Expect. 18:5

17

18-5

23

18-5

16

18-5

493

Expect.

BB

9

10

BP
20

^0

PP
11

337

i?IS X BS

1.

BB BIS BS ISS

488 15 22 17 13 472

Expect. 16-7 16-7 16-7 16-7
CE X IS

r-T-^T-]

CI CS EI ES

489 6 8 7 5 292 472

Expect. 6-5 6:5 6:5 6:5
Journ. of Gen. vn

BC X CI

r~^

BC BI
494 4 4

Expect. 3 3

T— 1

CI EI

479

BeBO X SS

I

Bse

496 91 218 473

5jD Inheritance and Evolution in Orthoptera HI

TABLES {continued).

BSx BI

II

xDS

BB BI

T

IS

r

DI

1
1
IS

496

7 5

2

2 473

506

13

15

475 492

Expect.

4 4

4

4

Expect.

14

14

II X BB

\
BI

497 72

143

482

CS X CI

"T

1

1 1

cc

CI

cs

IS CIS

BIS X BIS

607

14

9

15

14 1 489

1

1

?

Expect.

13-2

13-2

13-2

13-2 0

1
1
BB BIS

1
ISI^

498

5 3

2

482 488

ES

X DJ

499

3 4

2

482

1

500

13 36

26

510 515

r

DE
6

"T~

T"

501

2 4

2

485

508

DS

8

EJ
5

JS

Totals

23 47

32

6 489 492

Expect.

^5-5 51

25-5

Expect.

6-2

6-2

6--a

6-2

BIS X II

r-L-

CIS

X BIS

BI

IIS

1

— T

V

— 1
1

502 21

14

482 143

BC

BIS

CIS

ISIS

503 16

17

482 143

509
510
511
512

2

8

3

18

2
15

2
11

1
10

6
15

3 486

Totals
Expect.

37
54

31

34

9 486
5 515
16 522 485

II X .

1

ES

"1

IS

513
R514

10
1

10

1

10
1

8 522 485
2 510

Totals

42

41

43

43

ET

Expect.

42-2

42-2

42-2

42-2

504 19

16

475 489

Expect.

J7-5

17-5

cTs

X BC

'

EI X

II

-i,

T

1

505 2

II

4

489 475

515

BC
2

BIS
2

CC
3

cTs

6 486 38

Expect.

3

,9

Expect,

. 3-2

3-2

3-;2

3-2

R K. Nabotjrs

51

BC

TABLES {continued).
BC X BEBOFQ

T -T~T~T-

BB BC BF CF BBQ BCQ BFB CFQ
516 12 8 10 12 16 13 7 5 486 222

Expect. 10-3 10-3 10-3 103 10-3 103 103 10-3

BC X BBFe

ISTS X ISIS

1

j

T 1

1

ISIS

BBQ

BCe BFO CFe

524 33

510

517 9

8 11

8 486 222

525 38

522 509

Expect. 9

9 9

9

526 30

527 31

522
510

Total

132

IIS X J J
IJ JIS

BB X CECQEQ

518

2 4

502 491

r"

BC

1 ■ 1"

BE Bce

1

1

Expect.

3 3

1
BEB

528

4

8 10

1 482 369

Expect.

5-7 57 5/

' 5-7

GI X BB

1 1

IIS X BC

BC BI

8 9

8-5 8-5

489 493

—r-K-

519

Expect.

1

1

BI

T T

BIS . CI

"1

CIS

529

31

30 22

24 503 510

R530

3

6 6

8 510 502

CIS X CIS

r-t

Totals 34 36 28 32
Expect. 32-5 32-5 32-5 32-5

CC CIS ISIS

520

13

24

13 515

521

25

51

17 515

522

14

21

7 486

523

10

27

12 510 515

Totals

62

123

49

Expect.

58-5

117

58-5

581

Expect.

IIS

X

II

r-

.1

-n

II

IIS

4

7

503

475

5-5

5-5

4—2

52

Inheritance and Evolution in Orthoptera III

TABLES (continited).

BI

X

II

CC X CC

1

1
.Jl.

"1

II

1
CC

1
BI

537 26 522

532

Expect.

34

33-5

31

3 a -5

503 475

IIS X CIS

-r^T-i

ciS IIS ISIS

IISx

J

IIS

503

538

Expect.

CI
31
30

533

II

4

1

IIS
5

'' TsTt

4

35 28 26 503 522

30 30 30

534

1

1

1

503

Totals

5

6

5

BB X BB

Expect.

4

8

4

1

BB

539 88 485

ISIS

X

1.

BIS

j

ISIS

BC X ISIS

BIS

r-i-n

535

32

32

510 515

r 1

536

33

21

510 485

54<

BIS CIS

Totals

65

63

5 8 6 510 522

Expect.

59

59

Expect.

7 7

The / part of the pattern in this new combination is considerably
modified. In the IS, BIS or CIS individuals the normal pigment of
the dark mahogany spot which is most characteristic of //, and stands
out sharply in the hybrids BI, IS and others, seems to be diffused or
scattered over the area of the pronotum normally occupied by the spot,
and in addition extends over all the posterior part of the pronotum.
This scattering of the pigment, normally concentrated in the spot, gives
the whole an appearance of being diluted, or diffused. The SS part of
the pattern appears to be normal. The arrangement is indicated on
Plate II.

The significant feature is the complete combination, or linkage,
apparently permanent, of the factor for S and the factor for the modified
/ which is sufficiently demonstrated in the accompanying breeding
tables. This combination, IS, becomes a new form, a new multiple
allelomorph, pairing with, and allelomorphic to, any other multiple
allelomorph with which it has been tried, including the forms (multiple

U. K. Nabours 6^

allelomorphs) SS and // from which it was itself derived. It also
breeds true, matings (524), (525), (526), and (527).

Obviously it cannot be determined in this case, mating (472),
whether the female gave an unusual gamete containing the factors for
B and S and the male gave the gamete containing a factor for the
modified /, or whether an abnormal gamete from the male carrying
the factors for S and the modified / fertilized a normal gamete from the
female carrying the factor for B. It is also not possible for me to
suggest the means by which the combination, or linkage, was accom-
plished.

In mating (507), O^S' x CI (see table), another similar combination
occurred, with C involved instead of B as in mating (472). Here an
individual combining the C, S, and modified / patterns (CIS) appeared
among the otherwise approximately normal expectation of CC 14, CI 9,
CS 15, and 7>Si 14. The / and S in this mating had come through one
generation from the mating (472) which produced the original BIS (see
table). This CIS was in all observable respects similar to the other CIS
individuals produced by descendants of the original BIS aberrant form.
It died before reaching adult stage. When this CIS nymph, mating (507),
was first observed, during its third instar, consideration was given to
the possibility of its having been introduced into the breeding jar acci-
dentally from some contemporaneous jar containing CIS, as matings
(486), (510), and others. However, careful examination of the other
breeding jars containing CIS disclosed that none of these had arrived
at the third instar stage, and, furthermore, this jar (507) had not been
near enough these jars to make it likely that any accidental exchange
might have been made. However, there remains the bare possibility of
an accident in this case.

It appears that this strain of /, or S, or both, is subject to abnormal
behaviour, and that the linkage, or modification, is permanent, thereby
making a new and true breeding pattern. It may be that some of the
numerous multiple allelomorphs in Paratettix have been developed in a
similar way. The form QQ (see plate in second instalment) which so
much resembles the form CG, may hava secured the redness of its legs
from some other fofm, just as the SS, in the experiment, mating (472),
has become greatly modified through linkage with I. (In the new
form, IS, the *Si pattern is more conspicuous than the I pattern.)

I am under obligation to Mr A. W. Bellamy (1910 — Fellow in
Zoology in the University of Chicago) for valuable help during the

54 Iiiherita^ice mid Evolution in Ortlioptera lit

progress of the experiments and arrangement of the data. The expenses
have been carried by the Adams fund and State fund of the Kansas
Experiment Station, and I have had the most open-minded and complete
encouragement from Director W. M. Jardine.

LITERATURE CITED.

Bridges, C. B. 1916. " Non-disjunction as a proof of the Chromosonae Theory
of Heredity." Genetics, Vol. i. pp. 1—52 ; 107—163.

Castle, W. E. 1914. " Nabours' Grasshoppers, Multiple Allelomorphism, Linkage,
and Misleading Terminologies in Genetics." Ainer. Nat. Vol. xlviii. pp. 383,
384; 503, 504.

Dexter, John S. 1914. " Nabours' Breeding Experiments with Grasshoppers."
Artier. Nat. Vol. xlviii. pp. 317—320.

Morgan, T. H. 1914. "The Theoretical Distinction between Multiple Allelo-
morphs and Close Linkage." Amer. Nat. Vol. xlviii. pp. 502, 503.

Nabours, R. K. 1914. " Studies of Inheritance and Evolution in Orthoptera. 1."
Journal of Genetics, Vol. in. pp. 141 — 170.

1917. " Studies of Inheritance and Evolution in Orthoptera. 11." Journal

of Genetics, Vol. vii. pp. 1 — 46.

STUDIES OF INHERITANCE AND EVOLUTION
IN ORTHOPTERA. IV\

MULTIPLE ALLELOMORPHISM AND INHERITANCE OF
COLOR PATTERNS IN TETTIGIDEA^.

By ALBERT WILLIAM BELLAMY.
(With Plate III.)

Introduction.

According to Bateson each character of an alternative pair is an
allelomorphic one. He says ('09, p. 11) "The dissociation of characters
from each other in the course of the formation of the germs, we speak
of as segregation, and the characters which segregate from each other
are described as allelomorphic, i.e. alternative to each other in the
constitution of the gametes."

According to Morgan, allelomorphic characters are characters, the
determiners for which have identical loci in homologous chromosomes.
Shull ('15, p. 55) speaks of allelomorphism as " A relation between two
characters such that the determiners of both do not enter the same
gamete, but are separated into sister gametes." If instead of a single
pair, a series of several characters exists, each of which behaves towards
another as one of an alternative pair, it is said to constitute a system of
multiple allelomorphs ; the relation of the characters to one another
being known as multiple allelomorphism.

Triple systems of allelomorphs have been described in rats, guinea-
pigs, rabbits, Drosophila, beans, snapdragons, Lychnis, et al. Quadruple

^ Studies I, II and III of this series were published by Robert K. Nabours, Journal of
Genetics, Vols. iii. and vii.

•^ Contribution from the Zoological Laboratory of the Kansas State Agricultural College
and Experiment Station, No. 12.

56 Inheritance and Evolution in Orthoptera IV

systems are equally well known, having been studied in mice, guinea-
pigs, Drosophila, and corn. Nabours ('14) published an account of his
work with Paixitettix in which he described what has since been
recognized (Dexter, '14) as a system of multiple allelomorphic charac-
ters,— the most extensive system yet reported. Nabours showed the
existence of eight colour patterns, each allelomorphic to any of the
others. Since that time, not only has the behaviour of the eight
characters been fully confirmed, but the system has been extended
to include in all, at least fourteen patterns and probably several more
(Nabours, 17).

Material and Method.

The material used in this study was collected near Houston, Texas,
by Dr Robert K. Nabours and turned over to the writer in October, 1914.
I am under deep obligation to him for aid by way of suggestion,
encouragement, and criticism.

All the specimens used in the laboratory appear to conform most
closely with Hancock's description of Tettigidea parviperinis pennata
Morse.

Occurrence and distribution. The Tettigidea are widely distributed,
frequenting the damp surfaces of more or less deeply shaded areas, where
mosses, lichens, algae, and decaying vegetation, upon which they feed,
are to be found.

Life History. That part of the life cycle from the egg to the adult
may be completed in five or six weeks, although mating and deposition
of eggs may not occur, at least in the laboratory, for several weeks after
the last moult. In one instance, e.g. series (9) of Table I, the first young
of the parental generation hatched February 9, 1915. Two females
became adult March 9, one of which was mated on March 11, to her
father, producing the second generation of young April 22, a little over
eleven weeks from the hatching of the first generation to the hatching
of the second.

Technique. In the laboratory the grasshoppers are bred in cylindrical
screen or glass jars set in pots of moist sandy loam, the surfaces of
which are covered with a thin layer of peat. The bottom of a small
three-inch pot protrudes about one-half inch above the surface of the
soil in the centre of each jar, affording a clean place upon which to

A. W. Bellamy 57

place the food. Some algae also grow upon it. The jars are of two
sizes, the mating jars being 8" x 11" and the offspring jars 9" x 15"
(cf. Nabours, 14, pp. 143, 144, and Fig. 1).

The grasshoppers are fed on various filamentous algae which grow
in abundance during the early spring, throughout the summer, and
until late fall, in small streams, live-stock watering troughs, and similar
places. In the winter, as well as in the summer, special troughs kept in
the greenhouse supply a great deal of the food. If these food sources
fail the supply is supplemented with algae and lichens that grow upon
flower pots.

Characters under observation. The characters used in this study are
the colour markings of the pronota and of the femora of the jumping
legs. For present purposes, they may be considered as polyomate
" forms " of the species parvipennis. For the sake of convenience in
reference and recording, the different patterns are represented by capital
letters. In instances where these letters are used apparently to indicate
a single factor, or gene, it is to be understood that they indicate only
the initiative reaction or impulse, or whatever it is, that ultimately
results in the character as it appears in the adult animal. The different
patterns may then be indicated as follows (see Plate III) : G = yellowish
white striped pronotum. D = white lined pronotum (bilineata), i.e. two
whitish lines extend the full length of the lateral carinae. E = slightly
fulvo-aeneous plus blackish striped pronotum, i.e. the whole pronotum
has an ill-defined blackish stripe on a pallid or slightly fulvo-aeneous
background. F= narrow banded femora ; this is subject to considerable
variation, and in some individuals the pattern approaches a circle in
outline, while in others it may appear as a white line. H (present only
in connection with and in addition to some one or two of the other
patterns) = light brownish-red pronotum and femora of the jumping legs.
M = melanic, i.e. the whole animal is a dirty-brown to almost black
individual.

Homozygous individuals, since they receive their determiners for the
character in question from two parents, are indicated by doubling the
letter representing that pattern and heterozygotes are indicated by a
combination of the letters corresponding to the patterns represented
in their constitution.

58 inheritance and Evoliitio7i in Ortlioptera IV

Experiments.

Analysis of the original material. One male of the appearance of
BF (see plate), one male of the appearance of GF, three females of the
appearance of MM, and one female of the appearance of EEH, con-
stituted the original material. They were mated as follows :

(3) GF X MM,

(5) DF' X MM,

(13) DF^ X MM,

(9) DF xEEH.

These and the subsequent matings are shown in the table of matings
which shows how all of the analyses were carried out.

It may be said that many of the matings were made, not so much
in accordance with a previously arranged schedule, but because they
represented the only available material. No doubt many of the analyses
could have been carried out to better advantage, especially if one could
have counted on having the right male and female adult at the same
time, and if the contingency of the death of a valuable specimen did not
tiave to be reckoned with.

Explanation of Table I. This table gives the entire pedigree. The
matings are arranged in serial order reading in columns down the page
and from left to right. The number in parentheses immediately above
the horizontal line is the mating number ; the number in parentheses
above the mating is the mating from which the parents came. Take
for example :

(9.3)

EF X EF

(9.16)

~1
EE EF FF

7 21 8

Expectation 9 18 9

This is mating number (9.16), the parents coming from (9.3). The
actual numbers are 7 : 21 : 8 and the expectations 9:18:9 respectively.

The original male DF was used in (5), (9), (9.1), and (13) and is
indicated thus : *DF.

^ Same male.

4

A. W. Bellamy

59

TABLE I.

Expect.

Expect.

Expect.

Expect.

Expect.

Nature
CF X MM

I

1 ^

I

(3)

CM FM

78 79

78-5 78-5

(3)
CM X CM

I (3.4)

r H n

I I I

CC CM MM

^~67 18

65 -75 ^i-^5

(3)
FM X FM

(3.5)
1

FF FM MM

68
70-5

(3)
CM X CM

26

(3.6)

CC CM MM

82
75

18

:25

(3)
FJM X FM

I (3.7)

I !

I I I

FF FM FM

110 57

lii5-25 41-75

Expect.

(3.6)

CC X CC

I
CC

9

9

(3.8)

Expect.

Expect.

Expect.

Expect.

Expect.

Expect.

(3.8)

CC X CC

1 (3.10)
CC

29

29

(3.10)

CC X CC

1 (3.11)
CC

10

10

. Nature

*DF X MM

(5)

r

DM

\

1

FM

39

50

44-5

44-5

(5)

DM X DM

1 (5-2)

r
1

MM

DD DM

63

27

67-5

;82-5

(5)

FM X DM

(5.3)

r -- -
1

y j

DF DM

FM MM

7 14

15 14

12-5 12-5

12-5 12-5

(5)

FMx FM

(5.5)

r
1
FF FM

1

MM

30

12

31-5

10-5

60 Inheritance and Evolutio7i in Orthoptera IV

TABLE I (continued).

Expect.

Expect.

(5.2)

MM X MM

I (5-6)
MM

50

50

(5.3)

DF X DF

I (5.7)

Expect.

Nature

*DFxEEH

1 (9)

DE EF DEH EFH
15 12 8 11

11-5 11-5 11-5 11-5

r~T~i

DD DF 'FF
2 5 3

2-5 5 2-5

(9)
*DF X DEH

I (9.1)

r^T"~T~T ~T""T""T~]

DB DE DF EF DDH DEH DFH EFH
9 13 13 5 14 16 8 27

Expect. 13 13 13 13 13 13 13 13

Expect.

(9)
DEH X DE

-T-V

(9.2)

T"

DD DE EE DDH DEH EEH
4 4 3 9 i6 12

6 12 6 6 12 6

Expect.

(9)
EF X EF

1 (9.3)

1

1
1

1

1
1

EE

EF

FF

35

81

42

39-5

79

39-5

(9)
EF X DE

I (9.4)

r~T "T~ I

DF DE EE EF

Expect.

38
33-5

71

67

25
33-5

(9)
EFH X DE

I (9.5)

Expect.

DE DF EE
5 5 5

7-3 7-3 7-3

T~1

EF EEH DEH DFH EFH
5 4 10 12 13

7-3 7-3 7-3 7-3 7H

1

A. W. Bellamy

61

Expect.

Expect.

Expect.

(9.4)

DE X DE

I (9.8)

TABLE I (continued).

DD DE EE
28 61 26
28-7 57-5 28-7

(9.4) (9.1)
DFxDD

(9.9)

DD

10:5

DF

13

10:5

(9.4)
DF X DF

I (9.11)

I I

DD DF

25 39

19-5 39

FF
14

19-5

Expect.

Expect.

Expect.

Expect.

(9.4) (9.3)
EE X FF

I (9.12)
EF

4

4

(9.3) (9.1)
FF X DF»

DF

18

16-2

T

(9.15)

T

FF DFH
17 19
i6'^ 16-2

FFH
11

(9.3)
EF X EF

I

7
9

I

21

18

(9.16)

— I
I
FF

8

9

(9.3)
B£ X EE

EE
2
;3

(9.17)

DD

Expect. 1'56

Expect.

Expect.

(9.4)
EE X EE

I (9.21)
EE

5

.5

(9.4)
EF X EF

EE EF
9 43
16-7 33-5

(9.22)

— I
I
FF

15

16-7

Expect.

(9.3)

(9.5)

FF X

FFH

(9.24)

EF

i<^jr

FFH

i

FFH

9

8

3

10

7:5

7-5

7-5

7-5

62 Inheritance and Evolution in Orthoptera IV

Expect.

TABLE I (continued).

(9.6)
DEH X DEH

(9.25)

r~T~T~T T ~T T T 1

DD DE EE DDH DHDH DEH DHEH EEH EHEH

2 1

0-56 1

1

0-56

1

1-69

3

3-37

1
1-69

Expect.

(9.5) (9.1>

EEH X DDH

I ^9.26)

T

DE
9

10-25

DEH
22

20-'5

DHEH
10

10-25

(9.3) (9.1)

FF X EFH

1 (9.29)

EF FF EFH FFH
10 16 24 20
Expect. 17-5 17-5 17-5 17-5

Expect.

(9.3) (9.2)
FF X DEH

I (9-27)

DF EF DFH EFH

6 18 18 8

12-5 12-5 12-5 12-5

Expect.

DF
9
16

(9.3) (9.8)

FF X DEH

I (9.30)

EF
16

16

DFH EFH

33 6

16 16

Expect.

(9.5) (9.1)

EF X DD

(9.28)

1 ~i

DE DF
6 5

5-5 5-5

Expect.

(9.8)
DE X

(9.47)

DEH

(9.37)

1

Expect.

DD DE EE DDH
3 8 4 2

4-25 8-5 4-25 4-25

(9.11) (9.18)

DD X FFH

I (9.31)

I 1

DF DFH-
0 2

DEH EEH
12 5

8-5 4-25

A. W. Bellamy

63

TABLE I (continued).

(9.6) (9.17)

FFH X EE

I (9-44)

r "1

Expect.

EF
13

15

EF»

17

15

DD

Expect. 2 -5

"T~T^T"

DE EE DDH DHDH DEH

5

2-6

(9.26)
DEH X DEH

I (9.47)

T ~T — T"

4
i-5

3
3-9

4

7-8

T

DHEH EEH EHEH

4
3-9

Expect.

(9.18) • (9.29)

DFH X FF

I (9.49)

Df jFi*' DFH FFH
7 4 3 7

5-^.5 5-;?5 5-25 5-25

Expect.

I —

!

DD
0
1

(9.11) (9.31)

DF X DFH

I (9.51)

"T~"T"~T~T^]

DF FF DDH DFH FFH
3 1112

2 112 1

Expect.

DF

2
11

EF

0
1-1

(9.26) (9.24)
D£H X FFH

(9.52)

T— T T-

DFH DHFH FFH FHFH

2
3-3

5
3-5

64

Inheritance and Evolution in Orthoptera TV

TABLE I (contimied). •

Expect.

(9.6) (9.15)
EFH X FF

I (9.53)

EF FF EFH FFH
0 4 2 3

1-75 1-75 1-75 1-75

Nature
*DF X MM
I (13)

r ]

Expect.

BM

47
46

FM
45
46

Expect.

Expect.

(13)
BM X FM

I (13.1)

r~"T T"~i

BE BM FM MM

9 8 10 15

10-5 10-5 10-5 20-5

(13)
BE X BM

1 (13.2)

r""T ~T~i

BB BM BE FM

16 8.8

16 . 8 8

(3) (5)
FM X FM

I (17)

I I

FF EM MM

Expect.

32

32-25

11
10-75

Expect.

Expect.

Expect.

(3) (13)
C3I X FM
I (18)

r^T" T""1
CF CM FM MM

11 17 13 8

12-2 12-2 12-2 12-2

Nature

3IMH X MMH

I (19)

I H -1

I I I

MM MMH MHMH

7
4-5

11
13-5

(19)

MMH X ilfilf H

I (19.1)

I 1 1

I I I

MM MMH MHMH

2

2-25

7
6-75

Expect.

(3.6) (5.2)

CM X BM
I (22)

r""T""T~i

CB CM BM MM
18 18 18 15
17-2 17-2 17-2 17-2

Expect.

GE CF EM
9 5 5

7-5 7-5 7-5

(3.6) (9.1)
CM X EFH
I (24)

T~ T~"T"T"~1
FM CEH CFH EMH FMH

10 11 7 7 6

7-5 7-5 7-5 7-5 7-5

A. W. Bellamy

65

Expect.

TABLE I (continued).

(9.1) (5.3)
DFH X DF

I (25)

r""T"~T~^~"T~~l

DD DF FF DDH DFH FFH
1.111 5 3

1-5 3 1-5 15 3 15

Expect.

(3.6) (9)

,

CC X EFH

1 —
1

•(27)

T "

1

CE

CF CEH

GFH

7

9 0

10

Expect.

6-5

6-5 6-5

6-5

(17) (9.1)

(3.6) (9.2)

Fif X DD

CC X EEH

1 (28)
DF

62

cm

1
CE CEH

52

Exp€

set.

23 26
24-5 24-5

I —

I

CC

Expect. 2

T

CE

EE

3

2

Expect.

(29)
CEH X CEH

I (29.1)

"T T T T T 1

CCH CHCH CEH CHEH EEH EHEH

3

6

(3.8) (9.26)

CC X DHEH

I (38)

i I

GDH CEH

23 20

21-5 21-5

11
12

CF

0
Expect, 0-87

Journ. of Gen. vij

"T"

EF

(9.18) (29)
FFH X CEH

"T

Cf'H

(39)

T 1

CHFH EFH EHFH

1

0-87

3
2-6

8

;?-6

66

Inheritance and Evolution in Orthoptera IV

TABLE I {continued).

(19) (9.17)

MMH X EE

I (40)

r ~i

Expect.

Expect.

Expect,

Expect.

EM
1
3

EMH
5

3

(3.8) (5.6)
CC X MM

I
- CM

29

29

(41)

(3.8) (38)

CG X GDH

I (42)

GC GD GQH GDH
4 2 4 4

3-5 3-5 3-5 3-5

(3.8) (9.47)
GG X £HEH

CEH
90
90

(43)

Expect.

Expect.

Expect.

(22) (28)

CD X DF

(44)

1

1

CD
B

T"~T 1

Ci^ DD .DF
5 4 5

4-75 4-75 4-75 4-75

(44)
CDx CD

(44.3)

— I

r

CC CD DD
8 16 9

S-i?5 16-5 8-25

(9.49) (29.1)

DjFH X CG

I (56)

r""T"~T~i

CD GF GDH GFH
5 5 3 5

4-5 4-5 4-5 4-5

Table II. Combination of matings (3.5), (3.7), (17), and (5.5).

FM X FM

1
I

FF FM MM

240 106

Expect. 259-5 86-5

hie III.

Combination of matings (3.6) and (3.4).

CM X CM

1 h i

li-

ce CM MM

149 36

Expect. 138-75 46-25

A. W. Bellamy 67

A combination of the results given in Tables II and III gives for the
first two classes 389 and for the third class, composed entirely of MM,
142 ; the expectations being 398*25 and 132*75 respectively.

Table IV. Combination of the results of the matings giving a " 3 : 1 "
ratio, viz. (3.4), (3.5), (3.6), (3.7), (5.2), (5.5), and (17). The apparent
dominance and recessiveness of certain of these patterns will be men-
tioned in another connection.

I II III

552

169

1

Expect.

540-75

180-25

Combination of

matings

giving a " ]

DD

DF

FF

2 ^

5

3

(5.7)

25

39

14

(9.11)

DD

DE

DF

28

61

26

(9.8)

EE

EF

FF

35

81

42

(9.3)

7

21

8

(9.16)

9

43

15

(9.22)

CC

CD

DD

8

16

9

(44.3)

Totals

114

266

117

Expect.

124-25

248-5

124-25

In Tables I, II, III, and IV, the results have been tabulated as though
C, D, and F were each " dominant " or epistatic, and M " recessive " or
hypostatic. This was due, in part, to the fact that when these records
were made the writer lacked experience and familiarity with the material,
which, because of the gi-eat similarity in the case of CC and CM, of the
homozygous and heterozygous forms, made it seem desirable to record
the two classes together ; and in part due to the fact that CC and CM
are not readily distinguishable until about a week after the last moult,
and even then there are a few individuals which cannot readily be dis-
tinguished. In the case of FF and FM, the pattern FF is so small that
it would be difficult to distinguish between FF and FM, even though
the hybrids were exactly intermediate between them. The case is rare,
if it ever occurs, where some detectable difference does not occur between
a homozygous form and the hybrid.

The behaviour of H. An examination of Table I shows that the female
EEH in (9) gave two sorts of gametes, viz. E and EH ; that the female
DEH in (9.1) gave four sorts of gametes, D, E, DH, EW, and that all the

68 Inheritance and Evolution in Orthoptera IV

possible combinations with the two gametes of the male give eight
classes of young. However, (9.2), DEH x i)-£', produced six classes of
young, which meets the expectation, because when all the possible
combinations have been made the classes BE and DEH will each have
occurred twice. Homozygotes, or heterozygotes containing H, are better
analyzed by mating with homozygous individuals as has been dune in
(9.15) and (9.27), FF x DFH, and a number of others.

The mating (9.18), DFH xBFH, is of interest because among the nine
possible combinations of gametes we should obtain three combinations
in which H occurs twice, viz. DHDH, DHFH, and FHFH ; i.e. these
individuals should have a double dose of H. In the sense that when
inbred, all of the gametes will receive the factor (or factors) for H, they
will be homozygous for it. It is interesting to note in this connection
that a number of individuals appear in such matings as this which have
the reddish pigmentation in an appreciably more intense condition than
other individuals containing H in the same culture or other cultures.
That such individuals may have a double dose of H is shown by matings
(38), CO X BHEH, and (43), CC x ^H^H ; (38) producing CBH 23 and
CEH 20 ; and (43) producing GEH 90.

Homozygotic individuals with a single dose of H, as regards the
characters in question, always give two sorts of gametes, while hetero-
zygotic individuals with a single dose of H always give four sorts of
gametes. Homozygotic individuals with' a double dose of H always
give one sort of gametes, while heterozygotic individuals with a double
dose of H always give -two sorts of gametes.

It is seen that while the zygotic constitution of, e.g., BFH is either
B : FH or F : Z)H, its gametic formula is always B:F:BH: FH.

All the possible combinations of such a mating as has just been
described may be obtained in the usual manner from the " 16-square"
as for an organism differing in two " independent " characters, or the^
may be derived in the following manner and the ratios be made some-
what more apparent :

D: F: DH-.FH X D:F: DH: FH

.1

1 —

1

1

T 1 ■

DD

DF

FF

DDH

DHDH

DFH

DHFH

FFH FHFH

DF

DDH

DFH
DFH
DFH

DHFH

FFH

1

2

1

2

1

4

2

2 1

1

3
3

3
3

6

3

1

9

JOURNAL OF GENETICS, VOL. VII. NO. 1

PLATE III

A. W. Bellamy (59

This ratio of 4:2:2:2:2:1:1:1:1, or, when the single and double
doses of H are counted in the same class as was done for (9.18) and
others, of 6 : 3 : 3 : 2 : 1 : 1, is actually a 9:3:3:1 ratio ; or better, as
has been emphasized by Nabours ('17), the 9:3:3:1 ratio is actually a
4:2:2:2:2:1:1:1 :l"ratio.

Discussion.

The chief purpose of this paper has been to record the inheritance
behaviour of several colour patterns in Tettigidea which constitute a
system of " multiple allelomorphs " ; and of another pigmental character-
istic (H) '^ which exists in connection with and in addition to the other
patterns" ■

The theory of multiple allelomorphism as set forth by the Morgan
school, postulates that the determiners for all the characters of a given
system of multiple allelomorphs shall have identical loci in a pair of
homologous chromosomes. There may be as many sets or groups of
characters as there are chromosomes in the matured gametes. It is
stated that the same results may be explained equally well by assuming
that there is complete linkage, i.e. that the determiners lie so close
together in the chromosomes that crossing over never takes place, and
that the end results would be the same in either case.

If one wishes to interpret these results in the light of this hypothesis, '
it may be assumed that the " determiners " for the characters C, D, E, F,
and M have identical loci in a single pair of homologous chromosomes.
In the case of H it need only be assumed that its determiner is borne by
some other chromosome. H is not sex linked and is apparently, as
stated by Nabours for his @ in Paratettix, allelomorphic only to its
absence.

This report is based upon 3,219 recorded individuals.

EXPLANATION OF PLATE III.

Five " forms " of Tettigidea parvipennis arc represented in the first five figures on the
Plate, viz. CG, DD, EE, EF, and MM. The six figures, CE, CF, CM, BE, BE, EF,
represent hybrids between certain of the preceding forms. The remaining six figures show
the result of the addition of the factor H. For fuller explanation see text, p. 57.

70 Inheritance in Evolution and Orthoptera IV

LITERATURE CITED.

Bateson, W. 1909. MendeVs I'rioiciples of Heredity, i>. W. Camb. Univ. Press.
Blatchley, W. S. 1902. " The Orthoptera of Indiana," pp. 215—232. 21th An.
Rep. Deft. Geol. Nat. Resources Indimux.

Dexter, J. S. 1914. "Nabours's Breeding Experiments with Grasshoppers."
Amer. Nat. Vol. xlviii. pp. 317—320.

Hancock, J. L. 1902. Tettigidea of North America. Chicago.

Morgan, Sturtevant, Muller, Bridges'. 1915. The Mechanism of Mendelian
Heredity. Henry Hplt and Co., N.Y.

Nabours, Robert K. 1914. "Studies of Inheritance and Evolution in Orthop-
tera. I." Journal of Genetics, Vol. lii. pp. 141 — 170.

1917. "Studies of Inheritance and Evolution in Orthoptera. II." Joicrnal

of Genetics, Vol. vii. pp. 1 — 46.

Shull, G. H. "Genetic Definitions in the Nevi^ Standard Dictionary," Amer. Nat.
Vol. XLix. pp. 52 — 59.

I

Volume VII FEBRUARY, 1918 No. 2

FERTILITY IN CIGHORIUM INTYBUS: SELF-
COMPATIBILITY AND SELF-INCOMPATIBILITY
AMONG THE OFFSPRING OF SELF-FERTILE
LINES OF DESCENT.

By a. B. STOUT.

(With Plates IV— VI.)

CONTENTS.

PACK

Introduction 71

Material, Methods, and Terminology 72

Results of the Experimental Studies 73

1. Performance of a family of Barbe de Capucin x wild white-flowered :

the {&mi\y (E3xA)-4r- • . 73

2. Performance of families descended from crosses between the wild white-

flowered plant A and plant E22 of Barbe de Capucin ... 76

(a) The family (A x E22)-4- 76

(b) The family (A x E22)-9- 78

(c) The family (E22 x A)-10- 83

3. Fertilities in various Vegetative Types or Baces .... 88

4. General Summary of Results ........ 89

Discussion .............. 94

Conclusions .............. 100

Bibliography 102

Introduction.

The results presented in this paper pertain to the variation, the
heredity and the results of selection in respect to seed production in
progenies of self-fertile plants of chicory, such self-fertile plants having
first appeared sporadically among the descendants of self-sterile parents.

It has already been noted (Stout, 1916, 1917) that the type of
sterility involved in my studies with chicory can best be ascribed to
a physiological incompatibility operating between sex organs that are

Joom. of Gen. vii 6

72 Fertility in Cichorium intybus

fully formed, anatomically perfect, potentially functional, and of simul-
taneous development. It is quite evident that the incompatibilities
are not due to anatomical incompatibility (structural differences such
as hercogamy) or to impotence (degeneration of sex organs or sex cells).
Emh^yo abortion which results from a real gametic incompatibility
that develops after fertilization may also be concerned in the decreased
seed production and in the poor germinations observed. Such a type of
abortion is to be considered as quite distinct from that which more
purely involves nutrition.

As I have already shown (Stout, 1916, 1917) this sort of sexual
incompatibility is very general in chicory. It is in evidence in the
many instances of cross-sterility and in a very pronounced self-sterility
both in wild and in cultivated varieties.

Self-fertile plants arise, however, among the progeny of self-sterile
parents, and in some of my cultures of the variety •' Red-leaved Treviso,"
such self-fertile plants first appeared after three generations of self-
sterile parentage. Thus far, ' in my cultures, the self- fertile plants
arising spontaneously have been relatively few in number, and they
have exhibited various, grades of self-compatibility as judged by seed
production. In the lines of descent grown as offspring of the self-fertile
plants, as already reported, the inheritance of self-compatibility was
very irregular ; self-incompatibility appeared in all progenies, even when
these were offspring of two generations of highly self-fertile plants.
I have now to report the results of another generation obtained in
1916.

As is pointed out in my previous papers, the behaviour of known
progenies with reference to the development of compatibilities and
incompatibilities is of special interest in its bearing on fundamental
problems of sexuality and fertilization, especially as they are seen in
the bisexual higher plants.

Material, Methods, and Terminology.

All of the plants for which data are here presented descended from
three self-sterile parents. Two of these parents were of the common
unimproved cultivated chicory (Barbe de Capucin) designated in
the records as E3 and E22\ these were crossed with a wild white-
flowered plant designated as A. There are, therefore, two main
families which may be referred to according to parentage as the A x E3
family and the A x E22 family (including reciprocals).

A. B. Stout 73

The group of sister plants grown from the same plant or from the
same cross in any one season is called a series, and in these reports such
a group bears the number of the plant that was the immediate parent,
together with the numbers in" serial order designating the previous
parentage. Thus, for example, series {A x E°22) is a generation of
plants derived by using pollen of plant E22 on pistils of plftnt A ;
series {A x E'22')-Jp- is a group of sister plants grown from self-
fertilized seed of the F-^ plant (J. x E'22) no. 4- Thus the series
{A X E2^2)-J^-3-ll-, for which data are given in Table II, has had three
generations of self-fertile ancestry, of the different series and numbers
as designated, which indicate the line of descent from the original cross
between A and E^^. Although somewhat cumbersome, this treatment
presents a complete record of pedigree, from which the performance of
individuals, of lines of descent as a whole, and of families may be ascer-
tained. All the plants of a sub-family will be given the designation of
the common ancestor, thus all the descendants of plant {A x E22) no. 4-
may be considered as family {A x E22)-4--, and all the descendants of
plant (A X E22) no. 10 will be designated as family {A xE22)-10-, both,
however, being sub-families in the main family {A x E22) but descend-
ing from two different sister plants.

The data in detail for any parents, or for any series referred to, but
which were grown previously to 1916, are given in a report already
published (Stout, 1916).

Results of the Experimental Studies.

1. Performance of 'a family of Barhe de Gapucin x wild
ivhite-fiowered : the family {E3 x A)-^-.

In 1916 two series from two generations of self-fertile ancestry were
grown in the family {ES x A)-4--. The results obtained from the self-
pollinations of these 49 plants are compiled in Table I. Of the
23 plants of the series first presented in this table, all but five were
self-fertile with percentages of fertility ranging from 4 to 48, and with
an average fertility for the self-fertile plants of 20 °/^. The percentage of
fertility, frequently referred to in this paper, is determined on the basis
of the proportion of seeds produced by the flowers involved in the con-
trolled pollinations. Of the other series, 16 plants were self- fertile, and
10 were self-sterile ; the fertilities of the self-fertile plants ranged from
2 to 60 °l^ with an average of 15 °/^.

6—2

74

Fertility in Cichorium intybus

TABLE I.

Self-compatibility and incompatibility in two series of a family of Barbe de
Capucin (E3) X wild white-Jlowered (A) ; from, two generations of self-
ferlile ancestry.

Record for heads pollinated

Plant with

Flower

Total

With no

With

Fertility

pedigree

colour

heads

seed

seed

Seed per head

per cent.

(E3 X J

[)-4-4-

B

—

—

—

—

0-43

No.

4

W

8

5

5

3, 5,5

0 09

8

B

10

10

0

—

—

12

B

11

5

6

4, 4, 5, 5, 7, 11

0-18

14

W

11

4

7

4, 4, 4, 11, 12, 14

, 18

0-23

16

'B

9

9

0

—

—

17

B

10

2

9

1, 3 + B, 4, 7, 7, \

?, 12, 15

0-30

18

W

10

1

9

1, 2, 2, 5, 6, 7, 8,

9, 13

0-29

19

W

6

0

6

2, 2, 5, 5, 5, 6

0-22

20

B

11

11

0

—

—

21

B

11

3

8

4, 4, 5, 5, 8, 8, 9,

12

0-27

22

W

7

2

5

3, 5, 6, 10, 11

0-28

23

W

13

8

5

1, 1, 8, 9, 12

0-31

26

B

11

11

0

—

—

27

W

7

4

3

1, 2, 4

0-04

28

B

10

10

0

—

—

29

W

9

0

9

1, 2, 5, 5, 9, 12, 12, 12, 16

0-48

30

B

10

0

10

2, 3, 3, 4, 5, 6, 6,

7, 13, 15

0-32

31

B

12

10

2

3, 4

0-03

32

B

9

4

5

2, 3, 4, 5, 5

0-12

34

W

10

8

2

2,5

0 07

35

B

10

4

6

1, 1, 2, 3, 4, 7

0-10

37

W

10

2

8

2, 2, 2, 4, 5, 5, 6,

9

0-19

38

W

8

5

3

1,3,8

0-09

(E3 X A)-4-

Ser. II. 20

B

_

_

_

_

0-50

No.

1

B

9

9

0

—

—

2

B

12

12

0

5

H

7

3

4

3, 3, 5, 5

012

6

B

10

2

8

1, 1, 2, 3, 3, 3, 3,

5

012

7

B

10

10

0

—

8

B

10

5

5

1,"2, 4, 7, 17

0-17

10

B

12

4

8

3, 3, 3, 3, 4, 5, 6,

7

015

11

B

6

0

6

5, 11, 11, 12, 15, :

15

0-60

14

B

9

9

0

—

16

B

10

5

5

1, 3, 5, 6, 6

0-21

17

B

12

10

2

2,3

0-02

18

B

9

5

4

1, 1, 3, 4

0-05

19

B

7

3

4

2, 2, 5, 6

Oil

21

B

9

6

3

1, 1, 2

0 03

22

B

11

11

0

—

23

B

9

0

9

4, 4, 5, 5, 7, 8, IC

), 10, 14

0-42

24

B

9

3

6

2, 2, 3, 3, 5, 17

0-19

25

B

' 8

8

0

—

—

26

B

10

7

3

1, 2, 4

0 03

27

B

11

10

1

12

0-06

29

B

10

10

0

—

—

32

B

10

10

0

—

33

B

11

9

2

1, 7

0 04

34

B

15

15

0

—

—

35

B

9

9

0

—

36

B

9

4

5

1, 1, 2, 2, 10

010

A. B. Stout 75

A glance over this table shows that a rather large proportion of
plants are self- fertile. Especially is this true of series {E3 x A)-4^-4--,
which in this respect is perhaps the most highly self-fertile of the various
series thus far grown.

The more complete summary of the record for the family {E3 y. A)
is presented in Table VII. As there indicated, the family is not a large
one. From the cross between the two self-sterile plants ES and A,
21 plants were grown in the F^, of which four were self- fertile with fer-
tilities of 2, 4, 13, and 48. Progeny were grown only from one plant
having the highest fertility. Of the 18 grown, 10 were self-fertile,
and the fertilities determined for 9 of these ranged from 3 to 50°/^,
with an average at 26. Selection for parents for the next generation was
confined to the two plants most highly self- fertile. Thus the immediate
parents of the two series were 43 and 50 % self-fertile respectively, and
the common Pj parent was 48 °/^ self-fertile. The selection has here
been continually of plants with highest fertility. The regression to
a condition of complete self-sterility and to feeble self-sterility is most
noticeable. The number of plants is perhaps not sufficient to determine
the mathematical expression for such regression with accuracy, but the
records indicate an irregular and incomplete inheritance of self-com-
patibility.

However, the proportion of self-fertile plants is higher in the series
grown from self-fertile parents than that of the self-fertile plants
appearing sporadically among the progeny of the original self-sterile
parents. The average fertility for each of the two series from two
generations of self-fertile ancestry is lower than that of the preceding
generation {E3 x A)-4.-. (A series was grown in each of two different
years from seed of the plant {E3 x A) no. If.; data for both are here
compiled as for a single series.) The range of self-fertility has, how-
ever, been extended in the case of one plant to 60 7o- This plant,
however, is of the series which, as a whole, is of lowest average fertility.
The summary for the progeny of the plant {E3 x ^)-^-, as a whole,
shows that 65 °/^ of all the plants were self-f(3rtile in some degree, with
a distribution on the basis of self-fertility that is decidedly irregular
and skew, and with an average fertility of 0*197. It has already been
suggested by the writer that there is evidence that complete self-
sterility may involve different intensities of incompatibilities If there
were some means of determining comparative values for these, the
distribution for a family and for the different series might be found
to be more in agreement with a normally fluctuating variability. As

76 Fertility in Cichorium intybus

it is, the comparison may be based on the proportion of self-fertile
plants and their fertilities, as has been done above.

2. Performance of families descended from, crosses between the wild
white-fiowered plant A and plant E22 of Barhe de Capucin.

By far the greater number of plants of the cultures were derived
from crosses between the two plants A and E22. Among the 75 F^
progeny of the reciprocal crosses .between these two self-sterile plants,
there were 8 plants self-fertile to some degree. Progenies of three of
these plants constituting three families, (A x E£2y4--, (^ x E22)-9-,
and {E^2 x A)-10-, have now been grown for three further genera-
tions. With the one exception of series {E22 x A)-10- Ser. II. 10- all
the dififerent series of these families grown in 1916 had three generations
of self-fertile parentage. The data for these various series will now be
presented together with a discussion of the results obtained for these
(sub) families.

(a) The family {A x E22)-4.-.

Data for the three series of this family grown in 1916 are presented
in Table II, and a summary of all the series of the family is given in
Table VII.

Of the series (A x E22)-4--3-6- only two plants were grown. These
were both self-sterile. The fertilities of the parental line of descent
were respectively 4, 13, and 5.

The series (A x E22)-^-3-ll- comprised 29 plants, of which 20 were
self-sterile and 9 self-fertile. Not only was the proportion of self-fertile
plants low, but the fertilities of such plants were low, ranging from
1 to 26% with the average at 8°/^. For this series the immediate
parent was of rather high fertility (32 °l^), but the ancestry previous to
this was of 4 and 13 "j^. The feeble self-fertility of the series as a whole
and of the various individuals comprising it is most noticeable, especially
in comparison with the behaviour of such a series as {E3 x A)-4--4--
reported in Table I.

Of the 16 plants of series {A x E22)-4-0-3-, eight were self-fertile
with range from 2 to 62 °l^,8ind with an average of 25 °/^. The relative
number of self-fertile plants, the range of fertilities, and the average
fertility are all higher for this series than for (A x E22)-4--3-ll-.

A consideration of the family history shows that there has been no
parent in this family with a fertility higher than 32 7o- The fertility

A. B. Stout 77

TABLE II.

Record' for three series descended fro in plant (^A x E3), No. 4. All have three
generations of self fertile parentage.

Becord for heads pollinated

Plant with

Flower

Total

With no

With

Fertility

pedigree

colour

heads

seed

seed

Seed per head

per cent

{A X E22)-

-4-3-6-

W

.

0-05

No.

1

W

4

4

0

—

5>

2

W

11

11

0

—

—

{A X E22)-

-4-3-11-

■ W

_

_

_

0-32

No.

1

W

2

2

1

6

0-16

,j

3

W

14

12

2

2, 3

0-02

,,

5

W

7

7

0

—

>)

6

W

6

C

0

7

W

10

7

3

1, 3, 4

0-04

J J

8

w

8

8

0

—

— .\

J,

9

w

5

1

4

3, 5, 8, 10

0-26

,,

10

w

7

2

5

1, 1, 3, 3, 10

0-14

,,

13

w

6

6

0

.

>>

14

w

6

6

0

—

—

,j

15

w

8

8

0

—

>>

16

w

6

6

0

—

' --

>)

17

w

8

8

0

—

>>

18

w

12

10

2

1, 1

0 01

»>

19

w

3

3

0

- — -

>>

20

w

10

10

0

__

—

99

21

w

5

3

2

4, 6

0-10

9 9

22

w

4

4

0

—

99

24

w

6

6

0

— . ■

99

25

, w

10

9

1

2

0-01

99

27

w

8

8

0

— ■

99

28

w

6

6

0

—

99

29

w

8

8

0

— " ■

—

31

w

7

7

0

—

—

9)

33

w

12

12

0

. —

■ —

34

w

7

7

0

—

—

,j

37

w

3

3

0

—

—

38

w

6

6

0

.

—

9','

41

w

9

8

1

1

001

[A X E22)-4-6-3-

w

__

^-

— •

0-31

No.

1

w

3

0

3

5, 6, 12

0-41

99

2

w

10

;o

0

—

—

3

w

4

4

0

—

—

9

w

2

2

0

—

— '

11

w

8

5

3

1, 1, 1

0 02

/

12

w

3

0 .

3

3, 8, 9

0-41

,

13

w

5

3

2

5, 8

0-16

9 9

14

w

2

2

0

—

—

15

w

7

7

0

— •

—

16

w

6

0

6

1, 1, 2, 5, 8, 8

0-21

17

w

■9

6

3

2, 4, 5

0 07

18

w

2

0

2

8, 14

0-62

20

w

2

2

0

—

—

21

w

6

4

2

2, 11

0-11 .

9)

22

w

■ 3

3

0

—

,^

23

w

10

10

0

—

• —

78 Fertility in Cichorium intybus

of the first self-fertile parent {A x E21) no. J^ was very low, being only
47^. Although 7 of the 10 plants grown from its seed were self-fertile,
the highest individual fertility was 237o> and the average was 14 7o-
The four plants selected from this series as parents had fertilities of 12,
13, 16, and 22. The numbers of plants grown from these were too
small for an adequate judgment of the various series, but from the two
most highly self-fertile plants were derived the series (J. x E22)-4--3-ll-
and {A x E22)-4--6-3- reported above. The first self- fertile parent was
very feebly self-fertile, but with one exception parents for subsequent
series were selected for highest individual fertility.

The behaviour of the various series indicates an irregular inheritance
of the characteristics of self-compatibility and the frequent and decided
regression to self-sterility. The range of the degree of self-fertility was
rather decidedly extended in the series (A x E22)-Ji.-6-3-, and the
average fertility was also high. This series was decidedly more fertile
in every comparison than the series {A x E22)~Jf.-3-l 1- : the ancestral
fertilities were only slightly higher, being 4, 22, and 31 as compared
with 4, 13, and 32.

Considered as a whole, the family (A x E'22)-4-- had 40 7o of the
total plants self-fertile. The distribution on the basis of fertilities of
the self-fertile plants is irregular and skew, the percentage of self-
fertility observed in an individual is extended to 62, and the average
percentage for all self- fertile plants is ]6*5,

{h) The family {A x E22)-9-.

As indicated in Table VII, the third filial generation in this family
consisted of six series, which were derived from as many different plants
of the second generation which had in turn descended from two plants
of the first generation. Thus the series are of the two main sub-
families {A X E22)-9-Ji,~ and {A x E22)-9-5-.

The data for the series which descended from plant {A x E22)-9- no. 4-
are given in Table III. Of the 10 plants in series (A x E22)-9-4.-4-- ,
four were self- fertile ; of 13 in series (A x E22)-9-4-i0-, seven were
self-fertile; and of the 4 plants in series (A x E22)-9-4-ll-, two were
self- fertile. The fertilities of the immediate parentage were quite alike,
the complete record of ancestry being 23, 43, and 20 for the first
mentioned, 23, 43, and 1 7 for the second, and 23, 43 and 20 for the last
named. As shown in the tables, the proportional number of self-fertile
plants varies, as do the individual fertilities. The number of plants is
perhaps insufficient for adequate comparisons of differences. It is to be

1

1

A. B. Stout

79

noted, however, that self-sterile plants appear in large proportions, and
that there are various degrees of self-fertility in each series.

TABLE TIL

Record for three series descended from plant (A x E22)—9—, No. 4.
All from three generations of self -fertile parentage.

Record for heads pollinated

Plant with
pedigree

Flower
colour

Total
heads

With no
seed

With

seed

Seed per head

Fertility
per cent.

{A X E22)-

9-4-4-

B

—

—

—

—

0-20

, No.

1

B

11

5

6

1, 2, 3, 4,

5,

7

0 12

,,

2

W

11

11

0

—

—

>>

3

W

10

10

0

—

—

1

6

W

10

10

0

—

—

,

7

w

5

0

5

1, 2, 3, 7,

8

0-28

J

8

B

7

0-

7

1, 1, 2, 4,

4,

5, 5

0-18

,

11

W

10

10

0

—

—

>

12

B

7

6

1

6

0-06

>

13

B

12

12

0

—

—

»

15

B

12

12

0

—

—

(A X E22}-9-4-10-

B*

0-17

No.

1

B

10

10

0

—

—

2

B

2

0

2

1, 5

0-20

3

B

11

11

0

—

—

4

B

14

14

0

—

—

5

B

12

11

1

3

0-02

6

B

9

5

4

1, 2, 4, 4

0-08

8

B

10

9

1

2

0 01

10

B

8

0

8

2,4,6,6,9,

10,

12,12

0-51

12

B

3

3

0

—

—

13

B

10

10

0

—

—

14

B

10

6

4

1, 3, 4, 8

0-10

15

B

10

3

7

1, 3, 4, 4, 5

,6,

14

0-25

19

B

11

11

0

—

—

{A X E22)-

9-4-11-

B

—

—

0-20

No.

1

B

9

7

2

1, 1

0-01

j»

2

W

8

6

2

1, 3

0 03

,,

3

W

10

10

0

—

—

.

>

4

W

11

11

0

—

—

Considering the three series together, 13 plants were self- fertile,
and 14 were self-sterile. In only one plant was the fertility higher
than 28 7o. and in this case the percentage was 51. The average for

* In a former publication (Stout, 1916, Table 5) this plant was by error reported
as white-flowered.

80

Fertility in Cichorium intybus

TABLE IV.

Record for three series descended from plant {A x U22)-9-, No.
All from three generations of self fertile parentage.

ftecord for heads pollinated

5.

Plant with
pedigree

Flower
colour

Total
lieada

With no
seed

With

seed

Seed per head

Fertility
per cent.

(^

X E22)-

9-5-1-

W

—

—

—

—

0-07

No.

1

W

12

10

2

1, 2

002

,,

2

W

12

12

0

—

—

,,

3

W

10

10

0

—

—

,>

4

W

' 12

12

0

—

—

M

X E22)-

■9-5-6-

W

—

—

—

—

0-46

No.

2

W

12

12

0

—

—

„

3

W

7

7

0

r—

—

)>

4

w

11

11

0

—

—

,,

5

w

10

10

0

—

,,

6

w

11

10

1

2

0-01

,,

7

w

8

3

5

2, 4, 7, 11, 14

0-29

>5

8

w

9

9

0

—

—

>>

9

w

10

10

0

—

—

,,

10

w

12

12

0

—

—

,,

11

w

11

11

0

—

—

,,

12

w

11

9

2

3, 5

0-05

,,

13

w

11

4

7

1, 2, 3, 4, 7, 10, 10

0-21

>)

14

w

6

6

0

—

—

>>

15

w

11

11

0

—

—

>S

16

w

10

10

0

—

—

,,

17

w

11

5

6

2 + 7i, 3, 6, 8, 8, 11

0-23

>>

19

w

9

6

3

1, 6, 14

. 0-15

>>

20

w

10

8

2

3, 7

0-06

,,

21

w

9

0

9

B, B, 1, 3, 4, 5, 9, 12, 13

0-42

)»

22

w

15

9

6

1, 2, 4, 4, 4, 7

0-09

>»

23

w

10

7

3

2, 10, 18

0-20

>»

24

w

11

11

0

—

—

>5

25

w

11

11

0

—

—

>>

26

w

11

11

0

—

—

,,

27

w

10

10

0

—

—

,,

28

w

13

13

0

—

—

»>

29

w

11

9

2

3, 5

0-05

JJ

30

w

13

13

0

—

—

M

31

w

6

0

6

4, 10, 13, 14, 14, 15

0-77

A. B. Stout

81

TABLE lY— [continued).

Record for heads pollinated

PUnt with
pedigree

Flower
colour

^

Total
heads

With no
seed

With
seed

Seed per head

Fertility
per cent.

(A X E2ii)-9-5-12-

W

— -

—

—

—

0-70

No. 1

w

6

0

6

5 + 7J, 5,7, 12, 13, 18

0-65

„ 2

w

11

7

4

1, 5, 5, 6

0-09

,, 3

w

12

0

12

7, 9, 9, 10, 10, lO; 11, 12,
13, 13, 14, 15

0-66

„ 4

w

13

11

2

1,5

0 03

,,- 5

w

13

13

0

—

—

„ 6

w

12

12

0

— '

—

M 7

w

10

10

0

— '

—

„ 8

w

10

10

0

—

—

„ 9

^v .

11

5

6

5, 8, 10, 12, 14, 15

0-36

„ 10

w

9

7

2

2,4

0-04

„ 11

w

13

1

12

1,4,5,5,6,6, 7, 8, 8, 9,
11, 12

0-38

» 12

w

12

6

6

1, 1, 1, 2, 2, 3

0 10

„ 13

w

9

4

5

2, 2, 4, 5, 8

0-13

., 14

w

10

0

10

2, 5, 7, 10, 10, 11, 11, 11,
14, 15

0-o6

,, 15

w

10

0

10

2,2,2,7,8,9,10,12,13,14

0-50

,, 16

w

11

11

0

—

—

„ 17

w

9

3

6

6, 6, 6, 8, 10, 15

0-34

„ 18

w

10

10

0

—

—

,, 19

w

14

0

14

1,2, 2,2,3,3,3,4,5,5,6,
6, 6, 7

0-25

„ 20

w

9

8

1

2

0-01

,, 21

w

10

10

0

—

—

„ 22 '

w

13

8

5

1, 1, 2, 3, 5

0 06

„ 23

w

10

10

0

—

—

„ 24

w

10

10

0

—

—

„ 26

w

10

4

6

2, 2, 2, 3, 5, 6

0-12

,. 27

w

12

12

0

—

—

,, 29

w

13

13

0

—

—

„ 31

w

14

14

0

—

—

„ 82

w

12

12

0

—

—

„ 33

w

11

0

11

2, 3, 8, 9, 10, 10, 10, 12,
12, 13, 15

0-54

„ 34

w

12

12

0

—

—

„ 35

'W

15

15'

0

—

—

,, 37

w

11

11

0

—

—

„ 38

w

12

12

0

—

—

82 Fertility in Oichorium intybus

all self-fertile plants was 14 7o- This sub-family therefore is one of
rather low self-fertility. The parents have all been somewhat above
the average of self- fertility, and that of the second ancestral generation
was much higher than the average.

Data for the three series of the third filial generation derived from
the plant {A x E22)-9- no. 5 are given in Table IV. The immediate
parent of one series was selected for low self-fertility ; the parents of
the other two were selected for high self-fertility, one in fact having the
highest percentage of any plant thus far utilized in the cultures as
a seed parent.

Of series {A x E2^)-9-6-l-, only four plants were grown, of which
one was very feebly self-fertile.

Twenty-nine plants were grown and tested in series {A x E22)-9-'5-6-.
Seventeen were self-sterile ; twelve were self-fertile. As further shown
in Table IV, the individual fertilities were below 30 °/„ , except for two
plants. One of these was 42 7o self- fertile, and the other gave the
percentage of 77, which is the highest thus far realized in any of the
cultures. The line of parentage has been one of rather high fertility,
being 23, 33, and 46, and the series to which the parents belonged have
given high values for average fertility (38 and 29). Except for the one
highly self-fertile plant regression has been very decided in this series.

The series {A x E£M)-9-5-12- is of special interest, for the imme-
diate parent was one of 70 °/^ self-fertility, the ancestral record being
23, 33, and 70. The series is also the largest of any thus far grown in
these families. Of the 34 plants, exactly half were self-fertile. The
fertilities range to 66 °/^ , and are somewhat more evenly distributed
than is usual. The average of the self- fertilities is 28 '^/^. A comparison
of the data for all the various series (Table VII) shows that, on the
whole, this series is one of high fertility. However, half of the plants
were self-sterile, a large number of the self-fertile plants were feebly
self-fertile, none of the self-fertile plants were more self-fertile than the
immediate parent, and. the average is below that for any one of the line
of self-fertile parentage. Still the fertilities of the two series as a whole
are decidedly greater than that of the three series (Tables III and VII)
derived from (A x E22)-9- no. If,.

Considered as a whole, this family has been grown from parents
selected for high self-fertility. With the exception of one series of four
plants, no series had any parent of less than 17 % self- fertility. The
value for the first parent in line of descent was 23 ; the values of the two
parents of the next generation were 43 and 33 : and the values for the

A. B. Stout 83

parents of the next generation ranged from 7 to 70. Self-sterile plants
appeared in every series, and after three generations of ancestry of highest
self-fertilities the proportional number of such plants was high. Of the
total of 125 plants, 61 were self-fertile. The distribution of individual
self-fertilities is decidedly skew. The average fertilities of self-fertile
plants is 0-223.

(c) The family {E22 ^ A)-10-.

In 1916, seven series were grown in this family. One of these was
of the second generation in descent. This series consisted of 15 plants,
all but one of which were self-sterile. The fertility of the one self-
fertile plant was one of feeble self-fertility. Self-sterility was almost
complete, although the plants had two generations of self-fertile parents
with fertilities of 51 and 13.

TABLE V.

Record of a series having two generations of selffertile ancestry.

Record for beads pollinated

Plant with Flower Total With no With FertUity

pedigree colour heads seed seed Seed per head per cent.

{E22xA)-10-Ser. II. 10- jr — — — — 0-13

No. 1 TF 11 11 0 — —

„ 3 IF 14 14 0 — —

„ 4 IT 10 10 0 — —

„ 5 W 7 0 7 1, 1, 3, 3, 3, 8, 8 0-26

„ 6 ir 10 10 0 — —

„ 7 • IF 11 11 0 — -

„ 8 IF 10 10 0 — —

„ 9 IF 11 11 - 0 — —

^ „ 10 IF 12 12 0 — —

,, 11 IF 10 10 0 — —

„ 12 IF 12 12 0 — —

. „ 13 IF 7 7 0 — —

„ 14 IF 12 12 0 — —

„ 16 IF 10 10 0 — —

„ 17 IF 9 9 0 — —

The data for the six series of this family, having three generations
of self-fertile ancestry, are given in Table VI. As shown in the pedi-
grees, the first two series given have the same parentage for the first
two generations. The immediate parents of these series were therefore
two sister plants, which, it may be noted, were very nearly identical in
all vegetative characters, but differed considerably in the degree of self-
fertility. From the parent of 15 ""/„ fertility, 16 plants were gi'own and
tested for self- fertility. All but two were self-sterile, and those were

84

Fertility in Cichorium iiitybus

TABLE VI.

Self-Gompatihility a^iid self-incovipatihility in six series descended froin plant
{.E22 X J), No, 10. All series from three generations of self fertile
parentage.

Record for heads pollinated

Plant with
pedigree

Flower
colour

Total
heads

With no
seed

With
seed

Seed per head

Fertility
per cent.

{E22kA)-10-

-8-U-

W

—

—

—

—

0-15

No.

1

w

9

9

0

—

—

,,

2

w

11

11

0

—

—

,,

6

w

9

9

0

—

—

,,

4

w .

10

10

0

—

—

>>

0

w

11

11

0

—

—

))

6

w

11

11

0

—

—

,,

7

w

12

12

0

—

— '

,,

8

w

10

10

0

—

—

>>

9

w

9

9

0

—

—

))

10

w

12

12

0

—

—

,,

11

w

11

11

0

— ^ _

—

,,

12

w

10

10

0

—

—

» J

13

w

11

11

0

—

—

,,

15

w

9

2

7

1, 1, 1, 1, 2, 3, 7

010

>>

16

w.

8

4

4

1, 3, 3, 6

0-10

"

18

w.

12

12

0

■ —

—

{E22 .X A)-10-8-15-

w

—

—

—

—

0-40

Nj.

1

w

14

14

0

—

—

ji

2

w

10

2

8

3,6,7,8,10,12,15,17

0-43

)»

3

w

10

5

5

1, 2, 2, 2, 7

0-08

,,

4

w

10

0

10

1,2,3,4,5,7,7,13,1)

0-36

,,

5

w

9

3

6

2, 2,3, 3,4, 5

0 12

,,

6

w

13

10

3

1, 3, 9

*0-06

))

7

w

12

12

0

—

—

,,

8

w

8

1

7

i+B, 3, 3, 3,4, 7,9.

0-24

,,

9

w

10

9

12

0-07

,,

10

w

11

11

0

—

—

,,

11

w

3

1

2

6,5

0-23

,,

12

w

7

1

6

1, 2, 3, 4, 5, 12

0-23

,,

13

w

5

1

4

3, 5, 9, 12

0-31

,,

14

w

11

8

3

5, 7, 9

0-11

,,

15

w

6

0

6

1, 3,3,4, 4, 7

0-22

,,

16

w

()

3

3

2, 2, 11

013

"

17

w

9

1

8

1, 2, 2, 2, 2, 2, 3, 9

0-15

{IL22 X A)-10-13-5-

B

r

0-29

No.

1

B

7

4

3

4, 5, 8 '

014

,,

3

B

5

5

0

—

—

,,

4

B

10

10

0

—

—

A. B. Stout 85

TABLE

VI-

-{continued).

th

Flower
cx)lour

Record for beads pollinated

Plant wil
pedigree

Total With no
heads seed

With
seed

Seed per head

FertiUty
per cent.

No.

0

W

10

10

0

—

—

,,

6

B

10

7

3

4,5, 6

0 09

>»

8

W

7

2

5

1, 7, 8, 8, 9

0-28

>i

11

B

8

8

0

—

—

,,

12

W

6

6

0

—

—

>»

13

B

10

10

0

—

—

,,

15

W

9

9

0

—

—

,,

16

W

10

10

0

—

—

,,

17

B

8

5

3

2, 5, 9

012

u

19

B

8

8

0

—

—

,,

20

B

8

8

0

—

—

1>

21

B

6

6

0

—

—

,,

22

W

10

10

0

—

—

,,

23

B

10

10

0

—

—

>»

24

B

5

5

0

—

—

)>

25

W

7

5

2

1, 1

0-02

„

26

B

10

10

0

—

—

,,

27

B

11

11

0

—

—

»'

30

W

9

8

2

1,2

0-02

,,

81

B

8

3

5

1, 2, 2, 3, 4

0-09

>»

33

B

11

11

0

—

—

,,

35

B

5

5

0

—

—

,,

37

B

10

10

0

—

—

,,

88

B

10

10

0

—

—

,,

39

W

11

4

7

1, 2, 3, 3, 4, 5, 6

013

»J

40

B

9

9

0

—

—

,,

41

B

9

7

2

1,5

0-04

; ,-. >»

44

B

5

3

2

2,2

0-05

E22 x A)-10-13-12-

- W

—

—

0-25

No.

1

w

10

10

0

—

—

2

w

14

14

0

—

—

8

w

12

12

0

—

—

4

w

8

8

0

—

—

5

w

10

10

0

—

—

6

w

10

10

0

—

—

7

w

10

8

2

1,5

0-04

8

w

4

2

2

4, 0

013

10

w

11

11

0

—

—

11

w

9

4

5

1, 2, 2, 3, 4

0-08

12

w

1)

10

0

—

—

13

w

12

12

0

—

—

14

w

11

6

5

2, 2, 5, 7, 8

0-13

15

w

10

0

10

1,1,2,3,3,6,7,9,9,13 0-33

,,

16

w

11

5

6

3, 3, 5, 5, 6, 7

0-15

jj

17

w

11

11

0

—

86

Fertility iii Cichorium intybus

TABLE VI

-{continued).
Record for heads pollinated

Plant .with
pedigree

No. 18

Flower
colour

W

(E22xA)-

Total
liearts

11

With no
seed

With
seed

11

Seed per head

3, 4, 5, 5, 5, 6, 6, 8, 9,
12, 13

Fertility
per cent.

0-42

„ 19

W

10

10

0

—

—

,, 20

W

12

12

0

—

—

„ 21

W

10

8

2

1, 5

0 03

,, 22

W

10

10

0

—

—

„ 23

W

12

12

0

—

—

„ 25

W

10

10

0

—

—

)-10-13-13-

B ,

—

—

—

—

0-56

No. 1

B

10

10

0

—

—

,, 2

B

7

3

4

5, 7, 9, 15

0-33

» 4

B

9

5

4

7, 9, 10, 14

0-28

,, 5

W

10

10

0

—

—

„ 6

B

10

9

1

5

0 03

,. 7

B

7

0

7

3, 5, 8, 10, 12, 13, 14

0-58

,. 8

B

11

8

3

6,7,9

0-13

„ 9

W

10

10

0

— .

—

„ 12

B

9

9

0

—

—

„ 15

W

12

12

0

—

—

„ 16

W

11

11

0

—

—

„ 17

B

11

11

0

—

—

„ 18

W

11

11

0

—

—

„ 19

B

12

12

0

—

—

,, 20

B

13

13

0

—

—

,. 21

W

10

6

4

7, 8, 10, 12

0-25

„ 22

B

15

3

12

1, 1,2,2,2,3, 4,5,6,
7, 8, 11

0-22

,, 24

B

7

3

4

1, 2, 3, 5

0 10

„ 26

B

9

4

5

4, 6, 7, 7, 9

0-24

„ 27

W

9

6

3

2, 2, 11

0-12

„ 28

W

11

11

0

—

—

,, 30

w

9

2

7

2, 2, 2, 4, 8, 9, 13

0-30

„ 31

B

11

3

8

1, 2, 2, 3, 5, 5, 11, 13

0-25

„ 32

B

9

4

5

1, 2, 7, 7, 9

0-18

,, 33

W

8

8

0

—

—

„ 3i

B

9

9

0

—

—

„ 35

W

12

12

0

—

—

„ 36

B

10

9

1

2

001

(E22xA)-10-14 6-
No.

4 6-

W

—

—

—

1

W

12

12

0

2

W

13

13

0

4

W

15

15

0

5

W

12

12

0

6

W

10

10

0

7

W

12

12

0

8

W

6

2

4

0-13

1, 1, 3, 5

0-10

A. B. Stout 87

rather feebly self-fertile. Of the 17 plants derived from {E22 x A)-10-8-
no. 15, all but 3 were self-fertile. The difference in self-fertility seen in
these series is most marked, especially in respect to the number of
plants self-fertile. Results in such individual cases as these seem
to indicate that selection may be effective in increasing or decreasing
the development of self-compatibilities.

The next three series recorded in Table VI were all descended from
the plant {E22 x A)-10- no. i^, which had a fertility of 38%. The
immediate parents of the three series had fertilities of 29, 25, and 56.
Of series (^".^.^ x A)-10-1S-S-, in a total of 31 plants, 21 were self-
sterile. The fertilities of the 10 self-fertile plants were low, giving an
average of 10 ''/^ and a range extending only to 28 %. Considering the
record of the line of parentage, with percents of 51, 38, and 29, and the
series in the line of descent, the fertility of this series is decidedly low.

The ancestral record for series (E2^^ x A)-10-13-l^- is quite iden-
tical to that for the series just noted. Of the 23 plants in the series,
15 were self-sterile. The fertilities of the self-fertile plants ranged to
42 °/^, but the average was 16 "Z^. This is also a low performance con-
sidering the parentage.

Of the 28 plants of series {E2'2 x A)-10-13-13- , exactly half were
self- fertile with percentages that extended to 58 and an average of 19.
There has been, perhaps, no series grown with a more highly self-fertile
ancestry. Here the selections have been from parents with percentages
of 51, 38, and 56. Yet half of the series was self-sterile, and only one
plant exhibited a percentage of self- fertility higher than 33.

The 7 plants of series {E'22 x A)-10-14--6- were derived from an
ancestry with fertilities of 51, 14, and 13. Only one plant was self-
fertile with a percentage of 10.

A summary of the ancestral records of the various lines of descent
shows that although the various parents exhibited a considerable range
in fertilities, the larger series and the greater number of plants in the
family have an ancestry of rather high fertilities. The results in
summary emphasize the irregular heredity and the continual sporadic
or ever-sporting nature of self-fertility and self-sterility in chicory. Of
the 196 plants descended from the highly self- fertile plant {E'2^2 x A)
no. 10, a total of 118 were self-sterile. The percentage of plants self-
fertile is therefore 40. The distribution of the self-fertile plants on the
basis of percentages is decidedly skew. Much the greater number of
plants are feebly self-fertile, and the average fertility of all plants self-
fertile is 0-185.

Journ. of Gen. vii . 7

88 Fertility hi Cichorium intybus

3. Fertilities in various Vegetative Types or Races.

The continued growth of line progenies by self-fertilization of parents,
as practised in my chicory cultures, has led to the isolation or segregation
of various vegetative types or races that are very distinct, not only from
each other but from the general character of the original parents crossed,
or even from the first self-fertile parents of the Fi.

Marked uniformity among the sister plants of a single series first
appeared in the Fs generation, which was the second generation
after self-fertile plants appeared. For example, the plants of series
(A X F22)-10-13- Were very uniform in habit of growth, as is very
well shown in Plate IV. (The field number of this series was 41.) The
plants were abundantly branched, making rather bushy compact indi-
viduals constituting a semi-dwarf bushy race. In respect to flower
colour, this series was variable. In the foreground of Plate IV, im-
mediately in front of field label 41, is a typical plant of series
(E22 X A)-I0-8-. The habit of growth here shown is very different
from that of the semi-dwarf bush race. It is characterized by a tall
well developed main stem with large conspicuous leaves. The branching
is somewhat sparse and is erect ; at the time the photograph was taken
the branching of the plant in question had not developed. More
mature plants of this type are shown in Plate V (see field label
no. 49).

As has been noted above, and as shown in Table VII, two plants of
(E2^ X A)-10-8- and three plants of {E'22 x A)-10-13- were selected
as parents for a further generation. A part of each series of the latter
is shown in Plate V. To the right of field label 49 is a row of
{E22 X A)-10-13-5- which, it may be said, had not fully developed
when the photograph was taken ; in front of this label are plants
of {E22 X A)-10-13-13-, and to the left are shown plants of
{E22 X A)-10-13-12-. All the plants of these three series were
quite alike in having a general habit of growth that was quite
identical to that of the parents.

The very different vegetative habit of the series {E22 x A)-10-8-
and of the next generation, series {E^^ x A)-10-8-14'- and -15-, is
shown in Plate V in the plants of field number 49.

Other equally well marked vegetative types appeared in the F^,
and bred as true in the F^. Series {A x E22)-9-6-6- is one of rather
medium but scraggly growth, and with brittle branches and stems.
(See Plate V, field number 53.)

A. B. Stout 89

Other races are shown in Plate VI. In the right foreground,
extending to field label 63, are plants of (A x E22)-4-6-3-, which
were of a very decided dwarf habit with few brittle branches, and a
very marked susceptibility to a stem-rot due to fungous or bacterial
infection in the tips of the branches. The field no. 63 designates the
15 plants of {E22 x A}-10 Ser. II. 10-, all but one of which were self-
sterile ; these, it will be noted, are of a rather tall and much-branched
habit. Field no. 57 is for series {A x E'22)-9-4-10- , which exhibited
the rather incongruous combination of large vigorous erect rosette
leaves with a weakly developed and sparsely branched main stem.
To the left of field no. 57 is a row of series {A x E2i^4-3-ll-. This
series was somewhat like the dwarf series already noted, but was of
more vigorous growth. The branches were brittle and very susceptible
to the same disease. The dying tips of the branches are quite well
shown in the reproduction of plants in the foreground.

The above rather brief description will give some idea of the various
vegetative forms that have appeared in the various lines of descent.
It is to be noted that the plants of each series of this Fi generation and
of most of the series of the F^ are very uniform among themselves in
general vigour and habit of growth, but that various vegetative types
are very different fi:om each other. The general data for flower colour are
given in the various tables, which show that for some of the series
the flower colour was quite uniform, while for others white-flowered and
blue-flowered plants were both in evidence.

The data in detail given in Tables I — VI, and the summary of Table
VII, show that self-incompatibility develops in these various races in quite
the same degree. The most dwarf and the most robust races or lines are
quite alike in performance. Plants that are widely dissimilar may be
self-fertile or self-sterile. Also the performance within the various
series indicates that plants that are quite identical in all vegetative
characters may be either self-fertile or self-sterile.

4. General Summary of Results.

In Table VII, the results obtained in 1916 are compiled together
with those obtained in previous years, thus presenting a summary for
the various series and generations of each family. Here the summary
for the Fi generation derived by crossing the self-sterile plants A and
E3 or A and E22 is presented in italics for comparison, but is not
included in the summaries of families of self-fertile lines of descent, for

7—2

<

o o o O

I ^

=8 SS

I I

<N 1 «

I

00 (M i-H (p

I <M C<J ^

rH r-l ■>* ©

I "* IM <p

rH (M lO (8

kO QO M* lO
• O t~ 'O g

o 00 ;d ^

'-I ,-{ y-t ^

00 i« o (5

00 CO CO t>

rH (jq C3 ©

e

€9

l-l

6

■o

SI-^OrHioooaoia®
iooooo 000 O

a
<

1

0

!w

^

OJ

1

Tl

^

-<+(

^^

f)

-tj

-^

a-

X

X

P-i

"SJ

=»?

*;

l»q

I Co 5

t' o o '^ o

U5 iH JJ

rH O 0
CO O -u

O C! 00 ^

C9

tooeqcocoo^ooorj

o o 00 t^ cc

<M 05 CD "O

-N r-i eo

CD -0 S^ "M 5^5

I -I .1 I -1

CO S5 so ea

•* tn ■* ■* ■* g

CD t^ "^l c^l »— I 00

0000000

A

H

6

o

CI

oi c« ei
6

I '-^ ei

(M CO (J)

<M(M IIMlNiHCOMlO

Oq O O -^ C~ (M "-I

•^ CO CO CD CO 5<i CO

CD W CO O CO Tj( -*

in

0)

IM

H

i-l

t-

*
0

05

CO

10

0000000

o

ci o 2

060

«5 W CO C*I

C<l (M c-1 ,-i i-H ,-t O

rHCTi~ICDCC'NrHC<lT*(O00-*T-(

^_ i

!M O IM -* O CO -* -^ CO '-' '~ "* CO iS
rH ,Hr-l,H<Ni-li-(|H

C0C<lC0OCC»ftl050

O t^ O t~ CO O
<M rt -M O ■* t^

• ■* ^ l-S ^ CO 3«( >;

I I I I I I I I =s

I I I I ' I ^ I I s

I >« O '3' "5 CO CO
! ^ ^ (N (M >f5 rt

i-( t-( r»( t--<

7

•o

7

33

ci

^

:>5

C)

cL

0

0

i

I I I I I I I

ctj -* J?

"-( M OS

~^i i

i-l i-l >-< g

A. B. Stout 91

the self-fertile plants in the F^ series appeared sporadically. In the
columns giving fertility of ancestry, the individual percentages are given
for the parents in line of descent and the performance of the series
to which these belonged is, of course, to be found from the pedigree.
A general summary is given in bold face type for each family as a
whole.

It is a most noticeable fact that at least some self-sterile plants
appeared in every series but one, and this was a small series of only
two plants. The proportion of self-fertile plants varies considerably.
Considering the families as a whole, it is highest in the family (£";? xA)-^.-.
In the three sub-families of J. and E22 parentage, the percentages of the
proportions are quite the same (40, 49, and 40). The distribution of
the self-fertilities is quite similar in all families with the larger number
of plants of low fertility. The range extends into somewhat higher
values in the family {A x E22)-9-. On the basis of the average
fertility of the self-fertile plants, there is a range from 0"165 to 0"223.

Considering all these data, there appear to be no very decided family
differences in regard to the heredity and variability of self-compatibilities
and incompatibilities.

It is to be recognized that the data are not sufficient to give an
adequate judgment of the performance of a family or a line of descent
constituting a considerable progeny and having an ancestral record of
feeble fertility. As I have conducted the experiments, to select con-
tinually for very feeble fertility is to greatly limit the number of the
progeny. When the pollinations are made, there is no way of knowing
with any certainty what the degree of fertility is. When this becomes
known, it is usually too late to make in that year the large number of
pollinations necessary for the production of considerable seed by feebly
self-fertile plants. It would be quite possible, however, to keep feebly
fertile plants, and by making large numbers of pollinations in succeeding
years to obtain considerable seed.

Of the families gi-own thus far (A x E22)-4.- has an ancestry of the
lowest fertility. The data for the first self-fertile ancestor (Stout, 1916,
Table III) are quite adequate to establish its low fertility. Of 26
different heads pollinated on eight different days, 20 set no seed ; in the
six heads the number of seed were 1, 3, 3, 4, 4, 4, giving a percentage of
4. Of the ten plants grown to maturity from such seed, 7 were self-
fertile, but the highest individual fertility was 23. In the next genera-
tion from four parents, 28 plants were grown. Ten plants of one series
were all self-sterile, and in another series of 8 plants, 6 were self-fertile,

02 Fertility in Cichorium intybus

the widely different results in this case being obtained from almost
identical ancestral records. After three generations of self-fertile parents
with a record of 4, 13, and 32, one rather large series of 29 plants gave
31 7o of plants self-fertile, but with ranges only to 26 %, and the very
low average of 8 "Z^. A sister series with ancestral record of 4, 22, and
31 °/q gave 50 °l^ of plants self- fertile, of a range extending to 62 "/^ and
an average of 25 °/^. Such irregular and sporadic results seem character-
istic so far as my data go.

However, on the whole, this family was one of relatively low ancestral
fertility. Only three of the members have shown a fertility above 26°/^,
and the average of 0*165 is for all offspring lower than that of any other
family.

The record of ancestry for the family {E2'2 x A)-10- is one of high
fertility. The fertility of the first self-fertile parent was 51 7o- The
23 plants of the first generation, 16 of the second, and 82 of the third
all had self- fertile parents with no fertility lower than 25 °/<,, yet of
these the percentage of self- fertile plants was 41, which was almost
identical with the record for the whole family. The most highly
self-fertile series yet obtained in regard to the percentage of self-fertile
plants was in this family (see {E^2^2 x A)-10-8-15-), but there were also
two series that were almost completely self-sterile.

In the family {A x E22}-9-, the plants selected for parents have
been, as a rule, of medium fertility, although, of course, these have been
above the average. The range of fertility is considerably higher in the
family, and the average fertility of self-fertile plants is also higher than
in any other family.

Certain aspects of the results are perhaps clear. In every family,
as a rule, the total fertility (proportion of plants self-fertile, range of
fertilities, average percentage of fertility) is increased over that which
evidently occurs in the progeny of self-sterile parents ; at least the pro-
portion of self-fertile plants is much greater. This increase is, as a
rule, marked in the first generation of offspring grown from self-fertile
parentage. Continued selection for parentage high in individual self-
fertility does not steadily and continuously increase the fertilities of the
progeny either as to individual or to average fertility. The marked
tendency to the development of self-incompatibility has not been
eliminated.

An inspection of the data presented will show that in several series
the numbers of plants that are self-fertile are relatively high. For
example, in series having 83, 78, and 77 °/„ of the plants self- fertile it

A. B. Stout 93

might seem that the fertilities have been very much increased and that
further selection might yield a race all of which would be self-fertile,
at least to some degree. In eveiy case, however, a large number of the
ofifspring grown from highly self-fertile plants selected from such series
have been self-sterile, and on the whole the record for such progeny has
not been above that of the preceding generation.

In certain cases also, high records of fertility seem to be correlated
with an ancestral record that is high. While the results in this respect
are not uniformly in agreement, a number of the most highly fertile
series do have an immediate parent of high self- fertility. Some evidence
on this point may be gained by grouping the results according to the
fertilities of the immediate parentage and without regard to the family
or generation. This has been done in the following table.

TABLE VIII.

[

Summary according to

deg

ree

of self-fertility of immediate parentage.

Record for

progeny

Total

no.

plants

Number
self-
sterile

— ? —

M umber
self-
fertile

Per-
centage
self-
fertile

Frequency distribution, Percentage fertility

Average
fertility

Fertility
of parent

1-5

-10

-15

-20 -25

-30

-35-40 -45 -50 -55

-60 -65 -70 -75 -80

of self-
fertile

0-01— 0-10
0-11— 0-20

19
101

10
70

9
31

0-49
0-31

3
6

3
5

1 2

2 3

0-12

6

5

1 1 — — 1

0-17

0-21— 0-30
0-31— 0-40

97
91

55
44

42
47 '

0-43
0-52

12
9

10
5

8

7

3 —
2 8

1
3

A 1 9 1

0'14

4 1 ^ ± —

2 4 3 1 —

1 1 1 — —

0-21

O-43-OoO

108

46

62

0-57

13

8

8

8 7

8

2 _ 3 2 —

1 — — — 1

0-19

Ool— 0-56

51

26

25

0-49

4

2

4

3 5

2

1 2 1 — —

1 _ ^ _ _

0-22

0-70

34

17

17

0-50

3

3

2

— 1

—

12—12

2 — — — —

0-22

The results show that in respect to the proportion of self-fertile
plants there were only slight differences in the progeny of parents
of high or of low fertility : the lowest percentage (31) was obtained in
the progeny of parents whose self-fertility ranged from •11-20. In
regard to the range of the self-fertilities, the offspring of parents of the
lowest fertilities (-Ol-lO) extended only to 23 7o- The highest range
is seen in offspring of the class 0*43-0'50. The number of offspring,
especially of the lowest class of parentage, is not as large as one would
wish. The evidence is quite conclusive, however, that there is strong
regression, especially in the offspring of plants of high fertility, and that
the various progenies do not differ in the degree that is decidedly
correlated with the performance of the respective parents. Still the
data are suggestive that the higher ranges and averages are to be

94 Fertility in Cichorium intybus

obtained in the offspring of parents whose self-fertility is higher
than 30%.

Discussion.

Physiological sexual compatibility in chicory is decidedly sporadic in
its heredity. That its expression in the individuals of self-fertilized
lines of descent is continually fluctuating is clearly in evidence from the
behaviour of the various self-fertilized lines of descent reported above.
The number of gener^itions and the number of plants in the main sub-
families have been, it would seem, sufficiently large to establish these
points.

Starting with self-compatible plants that arose sporadically among
the J^i progeny of self-sterile parents, lines of descent have now been
grown through three further generations {F^, t\, and ^^4) and in every
generation, and in every series (excepting a single one of only two
plants) of each family, self-sterile plants have appeared, and usually
these have been in considerable numbers. A general summary of the
different generations and families shows that about half of the plants
have been self-sterile : that is, there has been considerable regression
in each generation, in each family, and in each series to the condition of
self-incompatibility which appeared to be the rule in the original stocks
from which these plants descended. Furthermore, the average per-^
formance has been quite the same for the successive generations.
, The first self- fertile parents of these families and lines of descent
were offspring of parents whose self-sterility had been thoroughly tested
and found to be complete. As individuals they appeared to be com-
pletely self-sterile ; the races or strains to which they belonged, however,
are not to be considered as absolutely self-sterile. The self- fertile plants
used as the first parents for the cultures here reported had thus one
generation of parentage known to be self-sterile, but I have elsewhere
shown (1917) that self- fertile plants may arise after three generations
of ancestry self-sterile on both male and female sides. In the develop-
ment of self-compatibility, these plants then differed sharply from their
immediate parents, and from the greater number of their sister plants.
The extremes, self- fertility and self-sterility, it would seem, are two
quite decidedly contrasted characters. In the apparent suddenness of
the occurrence of self-fertile plants among the offspring of self-sterile
parents there is much that is suggestive of what is quite generally called
mutation.

A. B. Stout 95

When the first self-fertile plants appeared in my cultures, I was of
the opinion that the characteristic of self- fertility would be decidedly
discontinuous, and that it would be transmitted as a fixed quality. In
other words, it was thought that the occurrence of such self-fertile plants
could be interpreted as " mutation," or possibly as a recombination of
fixed heredity units which had been separated as a result of previous
crossing. However, the various self-fertile plants which first appeared
exhibited various gi-ades of compatibility which are suggestive rather
that compatibility is a highly variable q-uality. Furthermore, the self-
fertilities of the offspring of self-fertile plants in all lines of descent are also
of various gi-ades. There is obviously a series of quantitative variations
in the behaviour of the plants as wholes that grade from complete self-
incompatibility to a very decided self-compatibility. Such a variability
of expression of fertilities and such incomplete transmission of the
characteristic of self-fertility, as is revealed in all my self-fertilized lines
of descent, indicate that occurrence of self- fertility is not due to mutations
which are at once fixed, or to recombinations of hereditary units. At
least, such recombinations are decidedly not stable.

Darwin (1SG8, 1877) held that all the facts regarding the occurrence
of self-incompatibility then known in such plants as Eschscholtzia
californica and Reseda odorata show that the phenomenon is widely
distributed and is of decidedly sporadic occurrence. In his opinion,
self-sterility is due to " some change in the condition of life acting on
the plants themselves or on their parents." The causes were held to be
environmental, and the self-incompatibility was assumed to rest in too
great a uniformity or similarity of the two kinds of sexual organs pro-
duced by a plant. The characteristic of functional fertility, according
to Darwin, exhibits fluctuations and chance variation as do other
characters.

Jost's (1907) theory of individual stuffs assumes that the causes of
self-sterility (physiological incompatibility) are individual, internal, and
epigenetic in that the sex organs fail to function because they are
produced on the same plant : the sex organs have the same chemical
individual stuff, and thereby lack the differentiation assumed to be
necessary for successful fertilization. The causes were fluctuating, but
were held to be solely internal.

Morgan's (1904, 1910) studies of self-incompatibility in the animal
Ciona intestinalis, led to much the same conclusion as was reached
by Jost. The failures to function are assumed to be due to too
great similarity that involves cytoplasmic relations established in the

96 Fertility in Cichorium intybus

individual. The similarity is thus considered as independent of the
degrees of dissimilarity in the germ plasm brought about by the
crossing necessary to give fertility.

Beginning about 1910, the attention of various investigators was
especially directed to a study of the breeding performance of plants
with respect to self-sterility in the attempts to determine its heredity
and obtain clues as to the nature of the processes involved.

In 1911, Baur claimed that the self-sterility of Antirrhinum molle
was recessive to self-fertility in A. majus, giving complete self-fertility
in all plants of the F^. The F2, it was reported, was composed of a
large proportion of self-fertile plants. Compton (1912, 1913) likewise
supports the view that self- fertility is a simple dominant over self-
sterility, and further interprets breeding results in Reseda on the basis
of a simple presence and absence hypothesis, the absence of some sub-
stance, either nutritive or stimulating to the growth of pollen-tubes
giving self-sterility, while the presence of such a substance gives self-
fertility. Neither Baur nor Compton presents adequate data for his
conclusions, and evidently both assumed a Mendelian behaviour of self-
sterility and self-fertility on a priori grounds. In regard to the later
generations of these Antirrhinum hybrids, Lotsy (1913) reports that the
i^'a generations are composed of self-fertile and self-sterile plants, and
that there are various degrees of self-fertility in evidence. The state-
ment is made by both Baur and Lotsy, however, that all plants of the
species A. molle are self-sterile.

Such interpretations have an advantage of appearing definite, simple,
and conclusive. However, the performance in chicory of pedigreed
cultures of offspring of self-sterile plants does hot show any such simple
and regular behaviour. Similar methods of study may reveal quite
identical conditions and results in the above named species,

Correns, in 1912, announced the very important discovery of physio-
logical cross-incompatibility among sister plants grown in the F^ seed
progeny of a cross between two self-sterile plants of Cardamine
pratensis, a species which had previously been known as self-sterile.
Correns thus proved, for the first time, that cross-sterility may exist
within a variety among plants of seed origin which exhibit no dimorphism
or trimorphism. By a grouping of the results, Correns arrived at a
Mendelian analysis of the hereditary performance. Line stuffs were
assumed to be represented and transmitted in the germ cells by anlagen,
and it was assumed that there could be no fertilization between gametes
carrying the same line stuff. An examination of Correns' actual results

A. B. Stout 97

(Stout, 1916) shows that the inter- fertilities and sterilities do not fall
into four nearly equal classes such as Correns has grouped them.
Compton (1913) has also pointed out that if Correns' assumption holds,
one-fourth of the F^ generation which he studied should have been
self-fertile. On this particular point Correns' data are incomplete : he
seems to consider all plants self-sterile, but his report includes data for
self-pollinations of only 13 out of 60 of the F^ generation. Of these,
however, three were partly self-fertile. The interpretation that self-
sterility and cross-sterility are due to a few line stuffs that are transmitted
as single hereditary units is obviously inadequate. The conclusion,
however, has been given considerable credence, especially in Mendelian
circles.

East (1915a and h) very soon pointed out the inadequacy of Correns'
interpretation, and formulated a " near Mendelian " interpretation for
the almost complete self-sterility and the almost complete cross-fertility
which he observed in hybrids between two species of Nicotiana. While
discarding the conception of factors directly concerned with fertility
and sterility as such. East considers that these conditions arise as
indirect properties of Mendelian units ; plants are self-sterile because
the male gametophyte produced by a plant can possess no hereditary
unit not possessed by the somatic cells of the pistil. He assumes that
this degree or element of similarity between pollen-tube and pistil in
self-pollination prohibits the formation of secretions in the pistil which
are necessary for the nourishment and growth of the pollen-tubes.

As to the facts of breeding performance, we may note that neither
East nor Correns gives adequate data as to the fluctuations in the degree
of fertility, or in the behaviour of pedigreed lines of descent from self-
sterile parents of a variety or a species, and there have been no data
published regarding the behaviour of pedigreed lines of descent from
self-fertile individuals which originated sporadically from self-sterile
parentage. A few such plants were in evidence in the ^i crop studied
by Correns quite as 1 found them in chicory.

It is especially to be noted that there are no published data
regarding the performance with reference to sterility and fertility of
cultures of the so-called self-sterile species Antirrhinum niolle and
Nicotiana Forgetiana. Detailed studies of the performance of these,
as well as of other species reported self-sterile, are greatly to be desired.

Moore (1917) recently reports that: "The species of Tradescantia,
alsike clover, alfalfa, and Shirley poppy showed different degrees of self-
sterility. Tradescantia was completely self-sterile ; in alsike clover about

98 Fei^tility in Cichorium intybus

2 °/^ of the flowers set seed when self-pollinated ; in alfalfa 27 °/g of the
flowers were fertilized with self-pollen, and when Shirley poppies were
self-pollinated 39 "/^ of the flowers set seed." The performance of in-
dividual plants is not indicated in these results, so it is impossible to
judge of the variability in fertility that occurs in the various individuals
involved. Evidently some individuals are self-sterile (except in Trade-
scantia ?) and some are self-fertile.

At this point one may venture to recognize that most of our mis-
understanding (and assumed understanding as well) of the transmission
of characters and of the nature of variation of all sorts is, no doubt, due
to attempts to analyze all sorts of characters in terms of hereditary
units. There has been a tendency to ascribe all sorts of characters,
superficial, fundamental, all sorts of pattern effects in pigment dis-
tribution, minutely qualitative or quantitative differences of highly
specialized organs, and general qualities of an organism as a whole to
factors which, it would seem, are mostly thought of as corpuscular units
serially arranged in the germ plasm. The inadequacies of the attempts
to analyze self-sterility on this basis are quite apparent both as to
methods and results.

To speak of the occasional appearance of self-compatible individuals
in an ordinarily self-sterile race as sporadic, and to refer the processes
determining the possibility of fertilization to variable interactions
between tissues and cells as such, may to many seem less definite than
an interpretation on the basis of assumed hereditary units. But the
irregular behaviour of compatibility and incompatibility both in ontogeny
and heredity in chicory is clear. Neither compatibility nor incompati-
bility are fixed and unchanging characters in transmission and in
expression, and are not to be considered as directly represented in the
germ plasm by hereditary elements.

In general it has been held that functional sex-vigor is congenital,
and that fertility in the sense of ability to produce large numbers of
offspring is hereditary. In many hermaphrodite plants, perhaps the
majority, self-fertility appears complete; within many species cross-
compatibility is perhaps complete ; the functional compatibility between
the sexes is so general in the plant and animal kingdoms that it has
been held to be congenital.

The presence of sexual incompatibility, therefore, between individuals
of a single race or variety, or even single line of descent, as it is found
in chicory, strikes one at first as a decided anomaly, and it seems still
more .an anomaly that the sex organs produced on the same plant, and

A. B. Stout 99

even in the same flower, may be as incompatible in function as though
they were produced by plants of unrelated genera.

There has been much speculation as to the nature and operation of
the physiological processes operating in such incompatibilities as are
seen in physiological self- and cross-sterility. In many cases of self-
incompatibility it has been reported that there is a limited or restricted
growth of pollen-tubes. These facts have led to views that the de-
termining factors in compatibility and incompatibility are limited to
the relations between pollen-tubes and pistils alone. Jost (1907) con-
siders that the poor growth of pollen-tubes in such cases is due to
the action of individual stuff which inhibits growth of pollen-tubes
having the same stuff. Compton (1912, 1913) believes that self-
sterility is due to the absence of a stimulating stuff, the presence of
which gives fertility. East (1915 a) attributes self-sterility to absence of
food stufis which are not secreted because the pollen-tubes involved do
not possess any hereditary element not possessed by the diploid cells
of the pistil. Moore (1917) considers that the limited growth in length
of pollen-tubes observed in self-sterility in Tradescantia is really due
to the presence of too much food.

Some of these views appear to regard the determining factors as
conditions of the pistil alone ; others consider that the conditions arise
through a reciprocal reaction between tubes and pistil. All of them
fail to recognize that a critical period in the growth of the pollen-tube
may result from secretions of the egg, and that the diflferent qualities
of the pistil may be due to the diffusion of hormones from the gameto-
phytes. As I have earlier pointed out, there is some evidence that
some cases of embryo abortion may be due to incompatibility, expressing
itself after fertilization and during the development of the embryo.
This may be true in some- cases in chicory. Further studies are in
progress on this point.

Cross-sterility (within a species) without self-sterility might be
explained as are isoprecipitation phenomena on the basis of an intra-
specific specificity of individuals, or groups of individuals, as such.
Self-compatibility, however, shows that an equally- marked differential
specificity may develop in sex organs and gametes produced by a single
individual : such specificity is not characteristic of the sporophytic
individual as a whole, but of the pollen-tubes, pistils, embryo sacs and
eggs as such.

100 Fertility in Cichorium intybus

Conclusions.

1. Self- and cross-incompatibilities in chicory develop independently
of either (a) anatomical incompatibility with its marked structural
differences and adaptations for cross-pollination ; or (b) embryo abortion,
at least of the sort that is due to the malnutrition of embryos by the
parent plant, and which in many fruit-bearing plants involves various
conditions of seed and flesh formation.

2. There is some evidence that both impotence and embryo abortion
are also present in some degree in chicory.

3. The experiments with chicory already reported (1916, 1917)
indicate (a) that self- and cross-incompatibilities are strongly in
evidence ; (b) that a few self-compatible plants may arise sporadically
from parents that are self-sterile, even after three generations of self-
sterile parentage ; (c) that the progeny of such self-fertile plants do not
breed true as to this character ; (d) that the degree of self-compatibility
varies; (e) and that selection for increased self-fertility after two
generations was not effective in isolating a completely self-fertile strain.

4. The new data reported in this paper are fully in accord with the
results obtained in previous years. A new generation, the third having
self-fertile ancestry, exhibited quite the same irregular heredity and
sporadic development of self-sterility (or regression to the condition
of self-sterility) as was seen in the former generation. In this respect
every family, every line, and every series were in close agreement.

5. Self- compatibility is entirely independent of differences in vege-
tative vigor. The various series of the crops grown in 1915 and 1916
showed widely different types in respect to vegetative vigor. Self-
sterility appeared in all these races with practically equal frequency.
Sister plants of the least vigorous dwarf race or sister plants of the
most vigorous vegetative race were either self-fertile or self-sterile
indiscriminately.

6. Self-compatibility and self-incompatibility operate independently
of potential sex-vigor. The total production of flowers varied greatly
among the various series. Plants with large numbers of sex organs
were either self-sterile or self-fertile, as were plants with the fewest
number.

7. Self-compatibility and self-incompatibility operate independently
of the purely nutritive relations Of the embryos to their parent plants.

A. B. Stout 101

Ten flower heads self-pollinated on a completely self-sterile plant will
set no seed, while ten heads on the same plant pollinated on the same
day with pollen from a highly cross-compatible plant will set abundant
seed. The fruits are rather small achenes having no endosperm, and are
practically composed only of the embryo : provided the pollination is
compatible, they develop equally well throughout the season (Data, see
1916, Tables XV and XVI).

8. Self-Gornpatibility and self-incompatihiiity appear independently
of any combination of germ plasm elements in so far as these can be
judged by the expression of characters. Each operates alike between
gametes that are similar or those that are dissimilar in respect to
hereditary units of genetic analysis. Plants widely different in such
qualities as colour of flowers, type of branching, shape of leaves, etc.
are either self-fertile or self-sterile, and plants of a sister series quite
similar in all respects are either self-fertile or self-sterile. When an F^
plant of hybrid origin is self-fertile in any degree the evidence indicates
that any of the sex cells may function in any recombination; on the
other hand in self-sterile sister plants whose sex cells must, it would
seem, be of much the same diversity none are compatible. Also all the
sex cells of an F^ plant which must have much the same germ-plasm
constitution may fail to function together, while those of a sister plant
may be highly functional. Two self-sterile plants, sisters of an F^ cross
or sisters of any generation, may be cross-fertile or cross-sterile quite
indiscri minately.

9. The development of either self compatibility or self-incompati-
bility occurs in both cross-bred and inbred races, the latter often being
highly constant races for vegetative characters. Both self -fertile and
self-sterile plants occurred among sister plants that were Fi hybrids of
rather wide crosses (Stout, 191G, 1917); they also appeared among
inbred strains derived by crossing self-sterile parents for as many as
three generations (1917), and they occurred, as here reported, among
the progeny of self- fertile plants, even after three generations of self-
fertile parentage. The positive evidence at hand, however, makes it
clear that self-compatibilities do not decrease in self-fertilized lines
of descent which are so uniform that they constitute decidedly pure
races.

10. The results obtained in the cultures of chicory make it clear
that self-incompatibility and self compatibility are here not to be described
as dominant and recessive characters, or paired allelomorphs, and that
there is no simple Mendelian formula that fits the results. The evidence

102 Fet'tility in Cichorium intybus

at hand for the behaviour of similar phenomena in other species is also
quite in agreement with this conclusion.

11. The conditions controlling sex-fusions, ']Mdigedi by the behaviour
of compatibilities and incompatibilities in such species as Cichorium
Intybus, arise in connection with the development of the sex organs and
sex cells as such. In this sense the controlling factors are of epigenetic
and individual development.

12. The factors which determine or prohibit successful fertilization
in chicory, whatever their essential nature may be, are highly variable as
to degree, specificity, and transmission in heredity.

New York Botanical Garden,
Jfai/ 10, 1917.

EXPLANATION OF PLATES.

PLATE IV.

View of series {A x E22)-10-13- of the semi-dwarf bushy race showing very decided
uniformity in general vegetative habit. Ten of this series were self-sterile and six
were self-fertile. In the foreground is a plant of series (A x E22)-10-8- showing a
tall sparsely branched habit of growth. These plants are from two generations of
self-fertile ancestry.

PLATE V.
View of chicory in eicperimental plot. Crop of 1916. Several races are shown.

PLATE VI.

Another view in experimental plot. Crop of 1916. Dwarf and semi-dwarf races are
especially prominent. Plants shown in Plates V and VI are from three generations
of self-fertile ancestry. Plants here shown were self-compatible or self-incompatible
quite indiscriminately.

BIBLIOGRAPHY.

Bauk, E. 1911. Einfiihrung in die experimentelle Vererhungslehre.

CoMPTON, R. H. 1912. "Preliminary note on the inheritance of sterility in Reseda

odorata." Proc. Cambridge Phil. Soc. x\u}.
1913. " Phenomena and problems of self-sterility." Neto Phytologist,

Vol. XII. pp. 197—206.
CoRRENS, C. 1912. " Selbststerilitat und Individiialstoffe." Festschr. d. Med. Nat.

Oes. z. 84. Versamml. Deutsch. Naturf. u. Arzte.

*1

JOURNAL OF GENETICS, VOL VII. NO. 2

PLATE IV

JOURNAL OF GENETICS, VOL. VM. NO. 2

PLATE V

JOURNAL OF GENETICS, VOL. VM. NO. 2

PLATE VI

A. B. Stotjt 103

Darwin, C. 1868. Animals and Plants under domestication, Vol. ii. Edition
by Orange Judd Co., New York.

1877. Cross and self-fertilization in the vegetable kingdom. Edition by

D. Appleton Co., New York.

East, E. M. 1915a. "The phenomenon of self-sterility." Amer. Nat. Vol. xlix.
pp. 77—87.

19155. "An interpretation of sterility in certain plants." Proc. Amer.

Phil. Soc. Vol. Liv. pp. 70—72.

JosT, L. 1907. " Ueber die Selbststerilitat einiger Bliiten." Bof. Zeit. Vol. Lxv.
pp. 77—117.

LoTSY, J. P. 1913. " Hybrides entre especes &' Antirrhinum.^^ IV" Conference
Internationale de Genetique, pp. 416 — 428.

Moore, C. W. 1917. " Self-sterility." Jour, of Hereditij, Vol. viii. pp. 203—207.
Morgan, T. H. 1904. "Some further experiments on self-fertilization in Ciona.
Biol. Bull. Vol. VIII. pp. 313—330.

1910. "Cross and self-fertilization in Ciona intestinalis.^' Arch. Entwicke-

lungsmech. Organ. Bd. xxx^. pp. 206—234.

Stout, A. B. 1916 "Self- and cross-pollinations in Cichorium Intyhus with reference
to sterility. Mem. N. Y. Bot. Oard. Vol. vi. pp. 333—454. PI. 30.

1917. "Fertility in Cichorium Intyhus: The sporadic appearance of self-
fertile plants among the progeny of self-sterile plants." Amer. Jour. Bot.
Vol. IV. pp. 375 — 395 (in press).

Journ. of Gen. vii

RACIAL STUDIES IN FISHES.

I. STATISTICAL INVESTIGATIONS WITH ZOARCES
VIVIPARUS L

By JOHS. SCHMIDT, D.Sc.

Director of the Carlsberg Physiological Laboratory,
Copenhagen, Denmark,

(With Plate VII and seven text-figures.)

During the past few decades, variation-statistical investigations have
been carried out on a large scale with several species of the food fishes
having their habitat in our northern seas ; a classical example is the
great work by Heincke on the Races of the Herring {Clupea harengus).
This appeared in 1898, and was of great importance, both in methodical
respects and also by reason of the results arrived at,

Heincke's work, with that of several others dealing with the same
question^ showed that the herrings of North and West Europe do not
make up a single coherent and homogeneous shoal. They are on the
contrary divided up into numerous more or less highly localised " com-
munities " or " populations," each leading an isolated existence, and each
to be characterised as distinct from other populations by average struc-
tural conditions, spawning time, etc.

Similar, more or less marked differences have been found among
practically all the species of fish which have been sufficiently investi-
gated in detail. A characteristic exception from this rule, however, is
the common freshwater eel (Anguilla vulgaris) which will be referred
to later on.

^ Eeaders wishing for further informatiou on the subject of Herring investigations may
refer to the recent paper by H. Chas. Williamson : "A short resume of the researches
into the European Races of Herrings and the method of investigation " (Fishery Board for
Scotland, Scientific Investigations, 1914, No. I. Edinburgh, 1914).

8—2

106 Racial Studies in Fishes

In trying to characterise these populations, various quantitative
features are employed, both such as are determinable by counting (e.g.
the number of vertebrae or of fin rays), and such as involve measure-
ment (as for instance the size and shape of the head). We have thus
in the former case to deal with integrated, in the latter with graduated
variates.

Where two populations have been found, by statistical examination
of a great number of specimens, to differ- in the mean of one or more
characters, it is customary, in fishery biology, to say that each belongs
to its own race of the species in question. The word " race " is, how-
ever, not employed by all writers ; some prefer the term " local forms "
or " families," while others again use all three indiscriminately.

This uncertainty in the terminology itself serves to indicate the in-
completeness of our knowledge as to the true nature of races in fishes.
We do not even know whether they are genotypically determined or
merely phenotypical phenomena. And a genetic analysis is still wanting
in this sphere. The reason for this lack of knowledge must be sought,
partly in the great experimental difficulties attending investigation of
our marine food fishes, partly also in the fact that fishery biologists most
frequently regard their problems as solved when once they have shown
that two or more populations actually are distinct and biologically inde-
pendent.

The question then is still before us : What is the cause of the racial
difierences found ? Why, for instance, should the race of herrings living
on the east coast of Scotland have a higher average number of vertebrae
than that — or those — of the Baltic, and why have the Baltic herrings
again a higher figure for this character than the herrings of the White
Sea ? Or how is it that the plaice (Pleuronectes platessa) in the North
Sea manage to develop a greater number of rays in the anal fin than the
plaice of the Baltic waters ?

The generally accepted view among fishery biologists is that racial
differences arise from the influence of differing external conditions under
which the races in question live, such as for instance salinity of the
water, its temperature, etc. It is thought that the average values for
the differing qualities, i.e. the racial characters, are actually dependent
upon. this or that temperature, salinity, or the like, which happens to be
peculiar to the water in question. Certain writers, such as Heincke,
imagine these external conditions as exerting a directly determinative
effect ; others are more inclined to regard them as acting indirectly,
through a process of selection. In this latter case, the variates not

JoHS. Schmidt 107

suited to the surroundings would be rejected, whereby the average
values for the different characters would of course be indirectly altered.

The object which I had set before me was to elucidate as far as
possible, by means of suitable material, some of the conditions respon-
sible for the racial differences found among our marine fish species in
nature. From the outset, I had a clear appreciation of two points;
first, that the investigation would have to be of an even more detailed
character than any of those previously carried out (though these might
seem detailed enough, for instance, in the case of the herring and plaice
investigations !). And secondly, that experimental aids would be neces-
sary, as without such, it would hardly be possible to arrive at any decisive
result.

The necessity of experimental investigations again forced me to
abandon the two species in which the question of race has been best
investigated (statistically), viz. the herring and the plaice, since experi-
ments with these seemed out of the question at present, at any rate with
the means at my disposal. In addition, I wished to have a species even
more " variable " and more local than the herring or the plaice. Both
of these are, as we know, pelagic for a more or less considerable part of
their life, and consequently subject to important passive or active dis-
locations which may lead to an intermingling of the individuals from
different populations.

A species which appears to fulfil the required conditions is the com-
mon Viviparous Blmny {Zoarces vivipamis, L.) with which I have now
been working since the autumn of 1914. I have made both statistical
and experimental investigations with this species ; only the former,
however, have up to the present been brought to a conclusion, and
only these are therefore so far advanced as to be suitable for publi-
cation.

In course of time, a great number of specimens, over 25,000, have
been examined, with regard to several characters. A detailed account,
with the figures pertaining thereto, appears in Vol. xiii of the Comptes
rendus des travaux du Laboratoire de Garlsherg, Copenhagen, and I must
here restrict myself to mentioning some of the principal results.

Zoarces viviparus is an extremely common fish in our Danish waters,
where it plays a far greater part than in the British. It lives as a rule
in quite shallow water, inside the 10 metre curve. Its distribution
in Eurojie has a marked north-easter.ly character. It is found from the

108 Racial Studies in Fishes

White Sea in the north to the English Channel in the south, is common
on the east coasts of Scotland and England but rare on the west, and is
not known with certainty from Ireland. Eastward, however, it pene-
trates into the innermost waters of the Baltic, to the base of the Gulf
of Bothnia and Gulf of Finland. A striking feature is the fact that it
is not found at the Faroes or Iceland.

One of the most prominent points in the natural history of Zoarces is
that it is viviparous. Pairing takes place late in the summer, and by
the close of the winter a considerable number of young, up to 400, are
brought into the world. These are at birth 4 — 5 cm. long, exactly
resembling their parents in all main features, and, like these, keeping
to the bottom from the very first. The quality of being viviparous is
almost unique among our northern teleosteans, and this is also one of
the essential reasons for my taking Zoarces for the purposes of investi-
gation, as we have here the great advantage of being able to examine
the progeny of a single female and compare the qualities with those of
the mother. For the young have, before they are born, all the qualities
with which we are here concerned fully developed in numerical respects.

Besides a great number of characters which have not been investi-
gated throughout in all samples, we have devoted particular attention
to the following four: (1) number of vertebrae ("vert"), (2) of rays in
the right pectoral fin (" Pd "), (3) of hard rays in the dorsal fin (" D^ "),
and (4) of pigment spots on the dorsal fin (" Pigm. D^ "). All these
qualities are determined by counting, so that we have only to deal with
integrated variates. This is an advantage, as the qualities in question
are not altered from long before the birth of the individual until its
death.

The characters investigated are of essential physiognomical impor-
tance, the number of pectoral rays, however, to a lesser degree than the
others. This will be seen from the three sketches (PI. VII, figs. 1 — 3).
The specimen shown in Fig. 1 has a large number of vertebrae, viz. 121,
and thus appears much more slender than those in Figs. 2 and 3, with
105 and 109 vertebrae respectively.

Highly characteristic of Zoarces is the fact that the dorsal fin has an
incurvation extending for a more or less considerable length down its
posterior part. This portion of the fin contains only hard rays, the
number of which will thus determine the length of the incurvation.
In the specimen Fig. 1, where there are no less than 12 hard rays, the
incurvate part is long ; in the specimen Fig. 2, where the number is
only 5, it is much shorter. Finally, the dorsal fin in the specimen

Jons. Schmidt 109

Fig. 3, where hard rays are altogether lacking, exhibits no incurvation
at all.

The number of pigment spots is also, as will be seen from the figures,
of physiognomical importance.

The number of vertebrae has been found to vary from 101 to 126
inclusive, a range of variation which is very considerable, and greater
than in any other species of fish hitherto investigated. The other
characters also have a wide range of variation. Thus we find, for the
hard rays 0—17, for the pigment spots 7 — 21, and for the pectoral
rays 16 — 22. For the sake of brevity I will here restrict myself mainly
to the mention of a single character, viz. the number of vertebrae.

Preliminary investigations soon showed that there was no difference
between the sexes in respect of the characters here in question. It was
further found, on repeated analyses of the same population, that most of
the characters exhibited a high degree of constancy from year to year,
as far as the scope of the investigations extends up to now. An instance
of such constancy is shown in "the graphs of Fig. 4, representing the
number of vertebrae in the years 1914, 1915 and 1916 for the same
population (St. 31, Ise Fjord, Sealand, Denmark). The number of
vertebrae is noted in figures along the vertical lines, each dot denoting
a single specimen of Zoarces. It will be seen that the three graphs
agree excellently well together, and therewith also the average values
for the three years, viz. 11342 (±0-49), 113-23 (±0-43) and 11340
(±0-44).

Altogether about 80 population analyses were made, embracing
material from the greater part of the area of distribution of Zoarces.
The separate population samples consisted as a rule of about 200
specimens, which permit of a very good degree of accuracy in the
determinations. For vertebrae, the highest mean found was 119-44,
and the lowest 107-98 ; similar results were arrived at in the case of
the other characters, least in the case of the pectoral rays, where there
are also, it must be noted, very rarely more than four variates.

The mean values for the various characters may appear in highly
differing combinations in the different populations. Disregarding the
pectoral rays, where the range of variation is but slight, and using
A, B and C to denote a high value, a, b and c a low mean value for
number of vertebrae, hard rays and pigment spots respectively, we find,
out of the eight possible combinations, six were realised in our 80
population analyses: ABC, ABc, AbC, Abe, abC and abc. Up to the
present, however, we have encountered no instance of aBc and aBC,

110

Racial Studies in Fishes

and of the six combinations noted above as found, not all were of the
same frequency.

An attempt has been made to characterise the various larger regions
of sea within the Zoarces area by means of the average qualities of the

118
116
114

0 •

' . .'

112
110
108

1914

•

Population

•
•

a = 113-42 (±0-49).

118
116
114

• •
•

112
110
108

'

Population 1915

•• •
•

a = Ji5-^5(±0-43).

118

116
114

112
110

108

Population 1916

a = ii5-40(± 0-44).

Fig. 4. Zoarces vivipants, L. Number of vertebrae. Population analyses for three
successive years from Station 31, Ise Fjord, Sealand, Denmark.

EASTERN PART OF

(1=117-37 (±0-48).

Outside Fjord

Mouth of Fjord

122
120
118
116
114

Population

• • • •

126
124
122
120
118

Population

******* ••••-•

116
114

112

• t •

112
110
108

••

• •
••

• •
•

a:^ 117-10 (±0-62).

Fig. 5. Zoarces viviparus, L. Number of vertebrae. Four population analyses from th

from the moi

ISE FJORD WITH R(

Mouth of Fjord

120
118
116
114

!..,. Population

Popula

•••••••<•••.•.•••••••••••••••

114

• «

*

112
110

108

112
110
108
106
104

• • • •

•'

a = llP01 (±0-i6). . a = 109-it6(±0-4

Fig. 6. Zoarces vivipanis, L. Number of vertebrae. Three population analyses from I«e Fjord, Roskih

LM FJORD, DENMARK.

Inner

part of Fjord

Population

Population

10

.

18

•

10

114

•

14

12

112
110
108

1(1

08

m

104

•• •
• •

a=113-ll{±0-o9).

a = 109-69 (±0-io).

ein part of Liiin Fjord, Jutland, Denmark. Showing how the number of vertebrae decreases
the fjord inwards.

ILDE FJORD, DENMARK.

Inner part of Fjord

Population

«

112

110

108

106

104
102

• •

a =107-98 {±0-36).
3rd, Sealand, Denmark. Showing how the number of vertebrae decreases from the mouth of the fjord inwards

JoHS. Schmidt 111

populations there found. Dividing up the area into four parts, and
utilising the same symbols as above, we obtain the follovv^ing : (1) western
North Sea, Abe; (2) eastern North Sea, abc; (3) west Baltic region,
ABC ; (4) east Baltic region, ABc. On taking together three popula-
tions, typical as far as possible, from each of the four regions, we obtain
the following mean values :

TABLE I.

Average value for Vert. Pd iJj Pigm. Di

I. Noith Sea 1. Western part 116-2 18 71 7-2 12-7

„ 2. Eastern part 111-2 18-48 6-0 12-3

II. Baltic 3. Western part 1177 19*35 8-0 14-3

4. Eastern part 117-2 18-66 11-1 123

On going into details, we find many discrepancies and irregularities
in the geographical distribution of the mean values. A certain regu-
larity, on the other hand, and also highly peculiar conditions are noted
in the shallow Danish and other fjords of the western Baltic. The
rule here is that the average number of vertebrae, hard rays and pig-
ment spots will be essentially lower up in the ^ords than outside, and
this applies even where the distance between the two populations only
amounts to some few miles. The graphs in Figs. 5, 6 show some in-
stances of this as regards the number of vertebrae. Table II, giving
the values for Manager Fjord, east coast of Jutland, Denmark, shows
the same also for the other characters.

TABLE 11.

Mariager Fjord, Denmark.

Average value for Vert. D.^

Stat, 14 li7-37 9-21

r„ 15 115-43 8-74

J „ 16 110-99 7-21

{_ „ 17 ■ 110-18 6-87

„ 19 109-30 6-40

St. 14 lies outside the jQord ; the remaining stations up in the ^ord ;
St. 19 at its base.

The lower number of vertebrae in the Zoarces of the ^ords gives
them a physiognomy differing greatly from that of their relatives on
the coasts, the former being of short and stumpy build, the latter
slender and elongated in shape.

112 Racial Studies in Fishes

Our population analyses have thus shown that Zoarces is split up
into numerous " races," often highly localised, and differing widely one
from another. In this respect, however, there is no difference of prin-
ciple between Zoarces and the other well investigated fish species such
as herring and plaice. The difference is only one of degree, depending
upon the fact that Zoarces is an even more " variable " and even less
migratory species than either of the other two.

One species of fish, however, to which I have also devoted particular
attention, viz. the common freshwater eel {Anguilla vulgaris), does differ
in principle, and a comparison with this species, which has about the
same number of vertebrae, etc. as Zoai'ces, is therefore highly instructive.

On going through a large amount of material, it was found that in
contrast to Zoarces, which, as we have seen, is split up into numerous
local races, all the eel populations of Europe are identical (see graphs.
Fig. 7). The explanation of this most remarkable fact must be sought
in the great biological difference between the two species; Zoarces
spends the whole of its life in the same very restricted area, whereas
all the freshwater eels of Europe undertake migrations of thousands
of miles out into the Atlantic to' spawn, the young thereafter journeying
the same way back. Long before the young, or rather larvae {Lepto-
cephalus hrevirostris), reach the coasts of Europe, they have developed
their full complement of vertebrae, etc., and we have here a natural
explanation of the fact that all the eel populations of Europe are
identical, or belong to one and the same " race."

As mentioned above, Zoarces gives birth to a great number of
young, which have already, long before they are born, developed their
full number of vertebrae, fin rays, pigment spots, etc. This renders it
possible for us to carry out individual offspring analyses with this
species, the offspring of each gravid female being separately investigated.

Examples of such individual offspring analyses will be found in the
graphs (Fig. 8). The material is from the same population (St. 31) as
the three population analyses for 1914, 1915 and 1916 in Fig. 4. On
comparing Fig. 4 with Fig. 8 we find that the average number of
vertebrae in various offspring samples can exhibit considerable variation
among themselves, and differ greatly from the average for the popula-
tion. Thus in the case of the three offspring samples in Fig. 8; the
mean values for number of vertebrae were about 117-83, 112-36 and
109-48 respectively, whereas the number for the population was in all
the three years 1914-16 between about 113-2 and 113-4.

JoHS. Schmidt

Iceland

113

118
116

•

Population

-"

114
112

.

a^ll4-73 (±0-32).
Copenhagen

118
116
W
112

Population

Bayonne

ll8
116
114
112

Population

•

a = 114-67 (±0-21).
Azores

116
114
112

Population

.

118
116

liT

112

...

Population

1

.....

a = 114-77 (±0-31).
Fig. 7. Common freshwater eel {Anguilla vulgaris). Number of vertebrae. Five population
analyses of material from Iceland, Copenhagen, Bayonne (France), Azores, Comacchio
(Adriatic). Showing that the eel populations of Europe are identical.

114

Racial Studies in Fishes

a = n7-83 {±0-o5).

a = 112 -36 (±0-S2).

/

114

$108

•

«..*.....>..■• 110

>in..»n...i_.

110

108

(continued from above)

106

• • •

a=109-48.(±0-i5).

Fig. 8. Zoarces viviparu.o, L. Number of vertebrae. Three individual offspring analyses
of gravid females from the 1914 population at St. 31, Ise Fjord, Sealand, Denmark ;
the same population as shown in Fig. 4. On comparing with P'ig. 4 it will be seen that
the number of vertebrae in individual offspring samples may differ greatly from that
of the population. For each of the graphs of offspring samples, the number of vertebrae
in the mother is noted.

n

JoHS. Schmidt 115

My own interpretation of the offspring analyses, which is supported
by the results arrived at in 1903 by W. Johannsen in his studies on bean
communities, is roughly as follows. It was found possible by population
analyses to subdivide Zoarces viviparus into numerous local " races."
The individual offspring analyses showed that the smallest unit hitherto
considered, the " race," may be resolved into still smaller elements, ex-
pressed by the means of the offspring samples. These smaller elements
may differ widely one from another, and likewise from the average of
the population (" race "), but when added together they reproduce the
picture of the race itself as the latter is expressed in the results of the
population analyses.

As already mentioned above, it is a generally accepted view among
fishery biologists that racial differences are due to the differing external
conditions under which the various races live. The salinity of the water
is here as a rule considered of great importance, and as this factor is also
the best elucidated up to date, I have endeavoured to ascertain how far
any agreement could be found to exist between the salinity and the
average values characterising the races of Zoarces. A few examples
may be quoted. Of the populations we have analysed, that of St. 57
in the North Sea (Anstruther, Scotland) is the one living in saltest
water, St. 52 in the Gulf of Bothnia in the Baltic (Hudiksvall, Sweden)
that with lowest salinity. The difference between these two places is
very great, the salinity at the Scottish station being about 34 °/^^, and
in the Gulf of Bothnia about 5 °/„^ only. The population at St. 57 thus
lives in a salt solution more than six times the strength of the water at
St. 52. The graphs in Fig. 9 show the values for number of vertebrae
in these two populations. It will be noticed that there is no difference
to speak of between the two graphs, either as regards course or average
values, the latter being about 1160 and about 116-4 respectively. This
example thus seems to suggest that the salinity is not of any real im-
portance.

Another example is illustrated in Fig. 10. The two populations
here shown are from Roskilde Fjord, Sealand, Denmark, and Kjelds
Nor, Langeland, Denmark. At these two places the salinity is very
nearly the same and rather low, viz. about 12°/^^. But it is at once
noticeable that the values for number of vertebrae exhibit an enormous
difference as between the two populations, amounting to no less than
about 11-5 mean. So great is the dissimilarity indeed that the two
populations have not a single common variate. The picture presented

116

Racial Studies m Fishes

is such that the uninitiated might well take it as representing two
distinct species, and yet as a matter of fact we have here only to deal
with two populations living within the same area of sea, and, as far as
we are aware, under uniform external conditions.

St. 52. Gulf of Bothnia (Sweden).

120

1 1 o

• •

Population

1 lO

116
114

• • e o

St. 57.

a = ii6'40(±0-47).

Scotland (Anstruther).

120
118

• •

• • •

Population

116

114

• • •

• •

112

rt = ii6-iO(±0-45).

Fig. 9. Zoarces viviparus, L. Number of vertebrae. Two population analyses from
St. 57 (Anstruther, Scotland) and St. 52 (Hudiksvall, Gulf of Bothnia). At St. 57
the salinity of the water is about 31 °l^^; at St. 52 only about 5 %o.

Our investigations thus by no means support the hypothesis that
the racial characters are determined exclusively by environment. On
the contrary, they seem rather to indicate that differences of environ-
ment are not sufficient to explain the structural differences between
the races, and that the importance of the salinity especially has doubt-
less been greatly over-estimated.

JoHS. Schmidt

117

On the other hand, the very distinct gradation of average qualities
found in the ^ord populations seems to suggest that the surroundings
may be of importance, either directly or indirectly, but what factors here
come into play we cannot as yet determine.

My view then, with regard to the nature of " races " in fishes, as
characterised by our population analyses, is briefly this: A fish "race"

124
122
120

118
116
114

•

' —

• • ••

•

Population

112
110
108
106
104

•
• • •

•

Population

••

• • ••

• ••

a = 119-M (±0-75)

a = 108-06 {±0-69)

Fig. 10. Zoarces viviparus, L. Number of vertebrae. Two population analyses from
Kjelds Nor, Langeland, Denmark (above), and Koskilde Fjord, Sealand, Denmark
(below). At both places the salinity is about 12 7oo- The boundary between the
populations indicated by a dotted line.

is largely a statistical conception. It implies a mixing of different
genotypes, and the average values characterising the " race " arc pri-
marily dependent upon the quantitative proportion between these;
only secondarily on the environment.

It should nevertheless be pointed out that we cannot expect to
arrive at any final decision before the results of the experimental

118 Racial Studies in Fishes

analysis are available. I hope therefore to revert to the question on
a subsequent occasion, after conclusion of the experiments, of which the
individual offspring analyses briefly referred to above form a part.

DESCRIPTION OF PLATE VII.

Fig. 1. Zoarces viviparus, L. Sketch of a specimen about 26 cm. in length from the

Sound. Number of vertebrae 121, of hard rays 12, and of pigment spots on dorsal 16.
Fig. 2. Zoarces viviparus, L. Sketch of a specimen about 25 cm. in length from

Roskilde Fjord, Sealand, Denmark. Number of vertebrae 105, of hard rays 5, and of

pigment spots on dorsal 11.

Fig. 3. Zoarces viviparus, L. Sketch of a specimen about 33 cm. in length, from Gullmar
Fjord, Sweden. Hard rays lacking. Number of vertebrae 109.

In all three figures the foremost pigment spot is situated on the front margin of the
dorsal fin.

Note. In calculating the probable fluctuation of the average number of vertebrae etc
; following formula was employed, viz. =*= -p x 0-67449 x 5. 1
in brackets after each average value of the character in question

the following formula was employed, viz. =*= -p x 0-67449 x 5. This value is always given

JOURNAL OF GENETICS, VOL VII. NO. !2

Fig. 2.

Fig. 3.

w

^^r ^9^^

^

PLATE VII

S^''

-■^.

A NOTE ON THE INHERITANCE OF COLOUR
IN ONE BREED OF PIGEONS— AN ATTEMPT
TO DEMONSTRATE A MENDELIAN TYPE OF
TRANSMISSION.

By J. S. W. NUTTALL, MB.

The inheritance of feather colour in domestic pigeons having proved
so difficult to fit into any scheme, I determined, in 1914, to investigate
the matter afresh, confining my energies to one breed and to a small
series of colours. I am strongly of opinion that when distinct types or
breeds are crossed, the results are complicated by the occurrence of
reversion. The matter is obviously complex ; a simple beginning was
indicated. The work done completes the preliminary stage only; the
onset of war, of necessity, seriously interrupted further experiments.
The results are given in bare outline, but it was thought advisable to
publish some details of what has been done.

The breed of pigeons used is known to fanciers as the " Racing
Pigeon," a variety which is being used in the present war in the Over-
seas Pigeon Service, and is doing useful work. This breed has been
firmly established in this country, and more firmly in Belgium, for
many years. Various colours are to be found, but the vast majority
of the hundreds of thousands bred yearly are of the following colours —
blue, blue chequered, red chequered and mealy. A short description of
these colours is necessary.

The blue birds are similar in colouring to the Columba livia, except
that the ground colour is, as a rule, brighter or " cleaner " in appearance,
though distinctly slaty-blue. They present, in common with Columba
livia, the typical wing-bars, the white (albescent) or blue croup and the
tail-bar. The ground colour of the primary and secondary wing-quills,
and the tail-quills, is dark blue. The blue chequered birds are similar
to the blues with the addition of chequering. The wing-bars of the
blue are mainly due to spots on the outer sides of the secondary wing
feathers ; the chequering arises from an extension of these spots to the

Journ. of Gen. vii 9

120 Colour Inheritance in Pigeons

smaller wing coverts and other feathers. The wing-bars and the tail-
bar are present as in the blue.

The colour of the so-called mealy birds is difficult to describe. The
ground colour is somewhat like that of fine oat-meal ; the wing-bars are
reddish — approaching the colour of damp sand. The mealy birds differ
in two salient points from the blues — the wing- and tail-quills are, as a
rule, pale in colour, and the tail-bar is absent.

The red chequered birds stand in the same relation to the mealies
as the blue chequers do to the blues, i.e., they are mealies with the
addition of chequering. The wing-quills and tail-quills are generally
pale in colour ; there is no tail-bar.

The colour of the birds used being of unknown composition, the
results have been based on group, rather than individual, matings. The
calculated results have been arrived at by considering all the types of
gametic combinations theoretically possible, and by assuming that each
pair of birds produces 16 offspring.

The experiments have led to the following conclusions :

1. Red (of red chequer or mealy) is dominant to blue.

2. Presence of chequering is dominant to absence of chequering.

3. Red chequers may be homozygous or heterozygous for colour or
chequering.

4. Mealies may be homozygous or heterozygous for colour.

5. Blue chequers may be heterozygous or homozygous for chequer-
ing, but are homozygous for colour.

6. Blues are homozygous.

Two pairs of allelomorphic factors may therefore be considered :
R (domt.), presence of red. r (rec.) absence of red.

C (domt.), presence of chequering. c (rec.) absence of chequering.

Red chequer may then be represented by EC, blue chequer by rC,
mealy by Re, and blue by ro.

Group 1. Red Chequer x Red Chequer.

Composition of parents, RGRC, RCRc, RCrC, RCrc, RcrC.
Types of mating possible, 15.
Number of offspring (16 from each mating), 240.
Calculated appearance :

(240) 196 Red cheq. : 20 Blue cheq. : 20 Mealy : 4 Blue.
(70-8) 57-9 : 5-9 : 5-9 : 11.

J. S. W. NUTTALL 121

Observed appearance of 71 birds produced in this group :
(71) 50 Red cheq. : 11 Blue cheq. : 8 Mealy : 2 Blue.

Group 2. Red Chequer x Blue Chequer.

Composition of parents — five red chequers as in group 1, two blue
chequers rOrC and rCrc.

Types of mating possible, 10.

Number of offspring, 160.

Calculated appearance :

(160) 100 Red cheq. : 44 Blue cheq. : 8 Mealy : 8 Blue.
(94-8) 59-3 : 261 : 4-7 : 47.

Observed appearance of 95 birds produced in this group :
(95) 41 Red cheq. : 50 Blue cheq. : 2 Mealy : 2 Blue.

Group 3. Red Chequer x Mealy.

Composition of parents, red chequers as in group 1, mealies RcRc
and Rci'c.

Types of mating possible, 10.

Number of offspring, 160.

Calculated appearance :

(160) 104 Red cheq. : 46 Mealy : 8 Blue cheq. : 2 Blue.
(44-9) 29-2 : 129 : 2'2 : 06.

Observed appearance of 45 birds produced in this group :
(45) 28 Red cheq. : 11 Mealy : 4 Blue cheq. : 2 Blue.
Group 4. Red Chequer x Blue.

Composition of parents, five types of red chequers as in group 1,
a single type of blue, rcrc.

Types of mating possible, 5. *"

Number of offspring, 80.
Calculated appearance :

(80) 40 Red cheq. : 16 Mealy : 16 Blue cheq. : 8 Blue.
(40) 20 : 8 : 8 : 4.

Observed appearance of 40 birds produced in this group :

(40) 17 Red cheq. : 7 Mealy : 11 Blue cheq. : 5 Blue.
Group 5. Mealy x Mealy.
Composition of parents, RcRc and Rcrc.
Types of mating possible, 3.
Number of offspring, 48.

9—2

122 Colour Inheritance in Pigeons

Calculated appearance :

(48) 44 Mealy : 4 Blue.
(33) 30-2 : 2-8.

Observed appearance of 33 birds produced in this group :

(33) 28 Mealy : 5 Blue.
Group 6. Blue Chequer x Blue Chequer.
Composition of parents, rCrC and rcrC.
Types of mating possible, 3.
Number of ofifspring, 48.
Calculated appearance :

(48) 44 Blue cheq. : 4 Blue.

(72) 66 : 6

Observed appearance of 72 birds produced in this group :

(72) 67 Blue cheq. : 5 Blue.
Group 7. Blue Chequer x Blue.

Composition of parents, rCrG, rcrG, rcrc.
Types of mating possible, 2.
Number of offspring, 32.
Calculated appearance :

(32) 24 Blue cheq. : 8 Blue.
Observed appearance of 32 birds produced in this group :

(32) 20 Blue cheq. : 10 Blue : 1 Red cheq. : 1 Mealy.
Group 8. Blue x Blue.

Composition of parents, rcrc.

Types of mating possible, 1.

Number of offspring, 16.

Calculated appearance :

(16) 16 Blue.

Observed appearance of 32 birds produced in this group :

(32) 32 Blue.

Group 1. In this group the observed red chequer and mealy figures
are lower, the blue chequer and blue higher, than the estimated figures.
To my mind there is a simple explanation of this. Fanciers generally
object to this type of mating, prefeTring to " mix the colours" ; it thus
naturally follows that the majority of red chequers are heterozygous for
colour, and when mated red to red, they will produce less than the

J. S. W. NUTTALL 123

expected number of red chequers and more than the expected number
of blue chequers. The experiments having been for the present almost
suspended, the testing of the extracted colours is incomplete. As far
as this has been carried out, the anticipated results have been obtained.

The blue chequers have proved to be homozygous for colour, some
being heterozygous for chequering ; the mealies homozygous for absence
of chequering, some giving rise to blues ; the blues have proved to be
homozygous for colour and absence of chequering. (See comments on
group 6 for occurrence of white in these extracted colours.)

Group 2. The heterozygous composition of red chequers may be
expected to have a greater effect in this type of mating than in group 1
type. The blue chequers being homozygous, and probably the majority
of the red chequers heterozygous for colour, it follows that the observed
red chequer figure will be low and that for blue chequer high. It is
worthy of mention that the colour of red chequers bred from dissimilarly
coloured parents (red chequer and blue chequer, or red chequer and blue)
is, as a rule, richer than that of birds bred from two reds. This probably
accounts for the popular aversion to red chequer and red chequer mating.
Red chequer cock birds from red chequer x blue chequer almost in-
variably possess black ticks. I have not found an example of these black
ticks in hen birds ; brown ticks may however sometimes be met with.

All the red chequered cocks which have come under my notice,
having one parent blue or blue chequered, present some degree of
ticking. A proportion of those from two red chequers and of those
from red chequer x mealy, are free fi'om ticks and do not appear to
develop any with age. On this may rest the possibility of separating
the homozygous and heterozygous red chequers of the male sex.

Group 3. As, in this group, red may come from either parent, the
effect of the heterozygous composition of red chequer is modified and
the observed approaches closely to the estimated result. The extracted
blue chequers and blues have so far proved to be pure for colour.

Group 4. The results obtained in this group and in group 3,
suggest that chequering depends on a single factor, and that it may
produce its full effect even when contributed by one parent. The
eleven blue chequers in this group were typical of the chequered type.

Group 5. A correlation was observed between paleness of plumage
(ground colour especially) and light coloured beaks and claws in some
of the mealies in this group. Of the blues, however, two in particular
were exceptionally sound in colour with dark beaks and claws. Blues

124 Colour Inheritance in Pigeons

from this group have been tested at greater length than most of the
extracted colours and have been found to be pure.

Group 6. The results from this group do not support the generally
accepted view of fanciers that almost any colour may arise from blue
chequer x blue chequer. I have examined the results of many breeders,
and where cross-mating can be definitely excluded, the results agree
with my own. The majority of the birds used in this group had at
least one red parent. This fact is strong evidence of their recessive
and homozygous nature.

Some blue chequers from two reds have a reddish brown tinge in
some of the wing-quills especially. I expected to obtain reds from these,
but so far have failed. This type of blue chequer appears to give an
increased proportion of reds when mated to reds.

Pied types. The behaviour of white is difficult to follow. Two main
types are met with, the " gay " pied and the type with a few white
feathers. The majority of gay pied birds follow a fairly uniform pattern
in the distribution of white. This " pattern " type is probably dominant
to self colour. On the other hand the type with a few white feathers is
recessive. In testing the extracted colours several examples of this
recessive type arose. No pied birds were used in any of the experi-
ments.

Group 7. The red chequer and mealy in this group arose from the
same pair and in the same nest. Their occurrence is disconcerting as
the likelihood of cross-mating was no greater in their particular case
than in any other. However, that cross-.mating is the explanation,
I feel assured of by the subsequent offspring of the parents of these two
exceptions. Thus far 14 birds have been produced from this one pair
in complete isolation. The seventh pair of young now (Mar. 15th,
1917) three weeks old provides no single atypically coloured specimen.

The appearance of this single family is :

(14) 8 Blue cheq. : 6 Blue.

It may be allowed that the composition of the blue parent is rcrc
and that of the blue chequer rCrc.

Group 8. In this group one of the birds had two white primary
wing-quills. Of the remaining 31 birds, 30 were typical blues, one had
grizzle primaries and a general colouring approaching more closely to
grizzle than blue. Grizzle being a macroscopic admixture of blue and
white, all the birds in this group have been considered to be substantially
blue.

THE INHERITANCE OF GLUME LENGTH IN
TRITICUM POLONICUM.

A CASE OF ZYGOTIC INHIBITION.

By W. O. backhouse,

Economic Botanist to the Argentine Government.

(With Chart.)

The sub-species of THticum known as T. polonicum is characterised
by long glumes which, in extreme cases, can attain a length of 40 mm.
whereas that of an ordinary wheat is in the neighbourhood of 10 mm.
only. There is a large number of varieties of T. polonicum known,
varying considerably in minor characters such as colour of leaf, colour
and shape of grain, degree of felting, etc., also in glume length itself,
some having an average length of about 19 mm., others as high as
28 mm. The sub-species T. polonicum hybridises easily with both
T. durum and turgidum and shows \ by the total lack of sterile indi-
viduals in F2 when crossed with varieties of the former, that it might
be considered, genetically speaking, merely an aberrant form of the
sub-species T. durum. The result of hybridising the long and the
short glume lengths is a first generation intermediate in this respect,
splitting in the second into long, intermediate and short in the ordinary
1:2:1 ratio, but not, in a manner possible to classify by eye and
necessitating the plotting of a curve to show the segregation.

When at Verrieres, in 1911, through the kindness of M. Ph. de
Vilmorin, the writer was able to examine a collection of varieties of
T. polonicum grown there and was struck by the fact that there were
none with perfectly smooth glumes and, furthermore, that the shorter
the glume of the variety, the more felted did it seem to be. The longest
glumed varieties, being only faintly pubescent, would roughly speaking

1 R. H. Biffen, Journal of Genetics, Vol. v. p. 225,

126 Glume Length in Triticum Polonicum

be classed as smooth. With the object of investigating this, in 1912 a
cross was made between a variety of T. polonicum, with an average glume
length of 29 mm. and very faintly pubescent, and a variety of T. durum,
which will be referred to in this paper as Kubanka. This last is a
smooth and otherwise typical example of T. durum, with an average
glume length of 12 mm. The first generation was a hybrid of inter-
mediate length — actually of an average of 18 or 19 mm. It was, how-
ever, remarkable in that it was distinctly pubescent — very much more
so than the polonicum parent. The second generation was surprising,
for it was soon observed that there was a proportion of plants bearing
fully pubescent ears— pubescence, be it remembered, is a dominant
character — yet the variety of Polish wheat used as a parent would have
been classed as smooth in comparison with such a wheat as Rivet or
Essex Rough Chaff.

At harvest time, a middle glume in the ear of each plant was mea-
sured and a curve plotted of the number of plants of each glume length
in millimetres (Charts Fig. la). The plants were also classified into
pubescent, intermediate and smooth, by means of a hand lens. The
polonicum parent would have fallen among the intermediates in this
classification, while the smooths were, as far as could be seen, perfectly
glabrous. The numbers observed, considering glume length alone, were
as follows :

Long and
Intermediate

Short

172

55

170-25

56-75

Expectation

A glance at the curve of this family will show that there is no dividing
line between the longs and the intermediates but, actually, the shorts
can be distinguished by eye-^that is to say, an extra short glumed
heterozygote. A starved plant, for example, which might fall in the
14 or 15 mm. lengths class, has an indefinable something about it which
points to its really belonging to the heterozygote class.

The carrying of large numbers of plants into the F^ generation
showed that, while only two mistakes were made at the short end of
the curves, it had been impossible to pick out any but the extreme
longs with the certainty that they would be pure to their particular
length. Considering those with a glume length varying between 10
and 14 mm. as being pure short segregates and those between 15 and
31 mm. as including both the longs and the heterozygotes, a count of
the proportion of pubescent individuals shows that in the short glumed
^ See pp. 130, 131 and explanation on p. 133.

W. O. Backhouse 127

class, the roughs predominate in the ordinary 3 : 1 proportion (Chart,
Fig. 1 b).

Felted Glume Smooth

40 15

Expectation ... 41-25 13-75

Examining 56 individuals, the theoretical expectation of homozygous
longs, and beginning at the extreme long end of the curve, it was found
that there were, among them, no individuals which could be called felted,
though with a lens a short velvety pubescence was seen on most. Among
the individuals with glume length between 15 mm. and 22 mm., which
may roughly be said to comprise the heterozygotes, the proportion was
85 felted to 31 smooth; but here there were only 15 individuals which
could confidently be called felted — in the majority of cases the closest
scrutiny was needed to determine to which category they belonged.

Finally the long class were examined critically — with the help of a
lens and the individuals selected which appeared to be absolutely smooth
— as smooth as the original short glumed Kuhanka. These plants were
grown the succeeding year and found to be all pure longs except two,
and, what is more important, a careful inspection with a lens showed
them to be also breeding true to this smoothness ; a smoothness which,
however, turned out to be only apparent in some cases.

Test crosses were made between these smooth lines and the original
Kuhanka, also other durums, to see whether the presence of the pubes-
cence in any way affected the segregation of glume length. The second
generations from these test crosses were surprising for, while some were
all smooth, others behaved in the same way as the original cross (giving
a 3 to 1 proportion of roughs and smooths among the short glumed
class), differing only in that the pubescence was of a minor degree, as
exemplified in the Canadian variety Prelude. From this it will be seen
that the long glume was able to inhibit the expression of a dominant
character and, furthermore, that there was a direct relation between
the length of the glume and the degree of felting — the greater the
glume, length, the less being the pubescence, even among the variable
heterozygotes. This will be better illustrated in the following case.

The- same variety of T. polonicum used in the first experiment was
crossed with a felted, black glumed variety of T. turgidum, not unlike
Rivet wheat of which it is, in fact, a descendant. The average glume
length of the turgidum is 11 mm. and of the polonicum 2S or 29 mm.
The first generation was intermediate in glume length — vaiying between
14 mm. and 17 mm. It was fairly felted and in colour white or faintly

128 Glume Length in Triticum Polonicum

tinged. In this experiment the polonicum will be considered smooth,
as indeed it is in comparison with the other ; the classification was done
by eye, unaided by a lens. The second generation was plotted as a curve
(Chart, Figs. 2, 3, 4 and 5) in the. same way as the Kubanka Polish
cross. Here again it was impossible to separate the pure longs from
the heterozygotes and statistically there is no sharp dividing line be-
tween the shorts and the heterozygotes (Chart, Fig. 5). However it
is fairly safe to say that the pure shorts are comprised among those with
a glume length between 9 and 13 mm. — though undoubtedly several of
13 mm. are poorly grown heterozygotes.

The ratio of short glume to long is :

Long and
Intermediate Short

614 178

Expectation ... 519 173

There were 692 plants in this family, besides 39 which were not noted,
being too green at the time of harvesting to determine the colour.

The second curve (Chart, Fig. 4) shows the total analysed into
those individuals which were felted like the short glumed parent and
those (shown by the dotted line) which were practically smooth, like
the Polish parent. Here again, it will be seen, the length of the glume
has acted as an inhibitor of pubescence. A study of the colour shows
this inhibiting nature even more clearly (Chart, Figs. 2 and 3). With
the single exception of one plant of 16 mm. glume length, all the fully
coloured individuals are between 8 and 13 mm. — among the shorts, in
fact. The proportion is :

Tinged and White Coloured

J29 49

Expectation ... 133-5 44-5

It was impossible to draw a really satisfactory distinction between the
heterozygous tinged or faintly coloured, and the colourless, as the faintly
tinged individuals were easily confused with stained whites, but the
pure blacks were easily classified. Nevertheless Chart, Fig. 3, shows
a curve of those individuals which were considered to be tinged.

The point of interest which attaches to this curve is the distribution
of the tinged individuals ; it will be seen (Fig. 3) that they are not quite
evenly distributed among those with glume length varying between
13 and 20 mm,, but that they occur with greater frequency at the
short end of the heterozygote curve. The colour seemed to be quite
independent of the pubescence. To ascertain whether, among the

W. O. Backhouse 129

longs, there existed some which, though they could not show it, were,
in fact, homozygous for colour, five long glumed individuals were crossed
with the short glumed Kubanka and the first generation plants grown
this year (1916). There is no need to wait for the second generation
for the results. Two individuals gave all tinged, one gave all white,
and two gave a mixture of tinged and whites. There can be no doubt,
therefore, that when the second generation is groAvn and true shorts
appear, there will also appear fully coloured individuals.

Only two families of long x short have been described but they are
typical of no less than seven second generations grown, all of which
show that the long glume in- wheat behaves as an inhibitor which, in
extreme cases, is as complete as though it were a case of genetic repul-
sion. There remains only one thing more to note — that from such long
and short glumed crosses it is possible to isolate a number of pure lines,
each with a different average length and breeding perfectly true to its
particular length.

It is possible that the gi-eater variation in length of glume among
the polonicums is merely due to the effect of magnified small differences,
these escaping observation in the ordinary short glumed wheats. If a
curve be plotted of the variation in short glumed wheats, this is always
steep and acute ; long glumed varieties, on the other hand, however often
reduced to single plant cultures and so purified, always give a long low
curve. The heterozygote curve is more or less intermediate in shape
between the two parents'.

Should it be the case that the apparent greater variation in glume
length is only due to a magnifying effect of the extra long glume, then
it seems possible to regard T. polonicum, which has always been con-
sidered a good sub-species in wheat, as merely a number of variations of
T. durum, differing from the short glumed type in one single unit
character which makes the long glume. Thus, had the existence of
black and fully pubescent polonicums been possible, a separate sub-
species would possibly never have been created.

Further experiments with T. polonicum throw some light on the
strange behaviour of Polish Rivet crosses 2.

I have been able to examine some of Professor R. H. Biflfen's material
and am familiar with the behaviour of this cross in England. When
leaving for this country, I took with me, in the form of grain, hybrids

1 R. H. BiffeD, Journal of Agricultural Science, Vol. i. Part 1.

^ " Suppression of Characters on Crossing," R. H. Biffen, Journal of Genetics, Vol. v.
Part 4.

130

Glume Length in Triticum Polonicum

1 1 1 r

cc

txX)

g:

CO

CO

00

o

CO

o

CN

o

o

O

O

O

o

O

o

o

CO

<N

CO

CN

>—

<N

W. O. Backhouse

131

o

o

o

o

o

o

o

O

O

o

o

o

O

o

o

o

O

00

r>-

CO

in

-"t

CO

CN

T*-

00

r>

CD

lO

-<*

CO

04

132 Glume Length m Triticum Polonicum

already made with the same strains of Rivet and of Polish wheats as
used by Biffen. The first generation was grown at Pergamino, in 1913,
and was notable from the start as being decidedly tinged. The second
generation was divided and grown, in 1914, in three different places,
viz, in the north, centre and south of the wheat producing area of the
Argentine Republic. In the north, all the individuals were colourless,
as in England. In the centre, at the latitude of Buenos Aires, some of
the short glumed individuals were tinged. In the south, in the Pampa,
the coloured ones were fairly clearly defined and could be classified, giving
the following proportions :

Rivet X Polish F^. Guatrache, Pampa.

Long and Intermediate

glume length Short glume length

72 30

Coloured "White Coloured White

0 72 7 23

Three doubtful shorts grown in 1915, to test, bred true to short
glume, but not to colour, and gave a total of 7 white to 13 coloured.
The proportion is peculiar but it is always difficult to distinguish the
homozygous coloured individuals from the heterozygous tinged ones. *

Polish X Rivet t\. [Reciprocal of the other cross.)

Long and Intermediate

glume length Short glume length

25 9

Coloured White Coloured White

0 25 4 5

It will be observed that here again the coloured individuals are only
found among the short glumed category. The coloured segregates of
these crosses are never quite so deeply coloured as Pedigree Rivet wheat
itself and grade almost imperceptibly from coloured to tinged and tinged
to colourless.

The interest of the experiment, however, lies in the fact that, whereas
in England the colour disappears and does not return in any subsequent
generation, the result of growing ^^s, obtained in identically the same
way, in the Argentine, is to prove that at any rate the colour is there
and, given suitable climatic conditions, will show itself. Rivet wheat,
grown for comparison, had the same peculiar mouse-grey colour as in
England and was no darker in this climate.

W. O. Backhouse 133

The cause of the suppression of colour in this particular cross must
be sought for in the shape of an inhibitor, brought in, either by Polish
wheat and meeting something in Rivet to release it, as it were, or vice
versa ; for this particular strain of polonicum crossed with coloured
varieties other than Rivet gives coloured descendants, in climatic con-
ditions under which, crossed with Rivet, they are colourless.

EXPLANATION OF CHART (on pp. 130, 131).

Fig. la. Family 74/14. Kubankax Polonicum.

Curve of glume length plotted for 227 individuals, showing that, whereas the short
glumed individuals, namely those whose glume length varies between 9 and 14 mm.,
are easily distinguished from the rest, there is no discontinuity in the curve between
the heterozygotes (of, say 15 mm. to 22 mm.) and the bulk of the homozygous longs.

Fig. lb. Curve of those individuals which were completely felted. It will be seen that
there are few of heterozygote glume length and no longs.

Fig. 2. Family 104/16. Polonicum x Turgidum 179.

The fully coloured individuals. These, with one single exception, are among the
short glumed plants.

Fig. 3. Those individuals which were not fully coloured and classed as tinged. Owing to
gradation in colour from plainly tinged to faintly tinged or stained white, this curve
is only of relative value, but serves to show that among the real longs there were only
three or four with a trace of colour.

Fig. 4. The same curve as in Fig. 5, resolved into felted individuals — shown by the whole
line — and those classified, without the aid of a hand lens, as smooth — shown by the
dotted line.

Fig. 5. Curve of the whole family of 692 plants, plotted for glume length alone.

STUDIES IN INHERITANCE IN THE HYBRID
PHILOSAMIA (ATTACUS) RICINI (BOISD) $
xPHILOSAMIA CYNTHIA (DRURY) ?.

By Mrs ONERA A. MERRITT HAWKES, M.Sc, B.Sc.

Number I. ON LARVAL CHARACTERS.
(With Plate VIII and two text-figures.)

Section I. On the Inheritance of Spots in the Larva.

The larvae of P. ricini and P. cynthia are alike in general body colour
but differ from one another as regards the ribmber of black spots present
in the last instar.

When an adult larva, P. cynthia has seven longitudinal series of black
spots, consisting, in all segments except the anal and cephalic, of two
mid-dorsal, two upper lateral, two or three spiracular and one or two
lower lateral spots. The upper lateral, spiracular and lower lateral are
arranged bisymmetricall3^ The spots of the lower lateral series occur
above the thoracic legs, on the abdominal legs and at a low level on the
legless segments. In the thoracic segments there is an extra spot on
each side at the base of the thoracic legs (Fig. 1). I have not been able
to breed P. cynthia (Drury) from Ning-po, so give this description from
the verbal statement of Mr J. W, Watson. Packard (9) gives a description
and figure of P. cynthia advena (Walkeri), the nearly related American
species, but does not mention the moth from Ning-po. Packard's descrip-
tion of P. cynthia advena would do perfectly for the Ning-po variety,
not only as regards spots but also in other details, but as I have
found the scales of the Ning-po cynthia considerably different fi'om
those of advena, and as they may differ considerably in genetic consti-
tution, it is possible that further observation may show an increasing
number of distinguishing points between the two.

Jonrn. of Gen. vii 10

136

Inheritance in Philosamia Hybrids

F. ricini differs from P. cynthia in the absence of the mid-dorsal,
upper lateral and spiracular series of spots. I bred a number of the
wild P. ricini and found that the lower lateral spots varied considerably —
there was every condition between two spots in each segment and none.
(Text-figure 1.) I did not inbreed the ricini, hence do not know whether,

M.D.Sp.

Fig. 1. Diagrammatic drawing of the second and third thoracic segments in P. cynthia.

D.r. = Dorsal tubercle.

I.)S?p.r. =Infra-spiracular tubercle.

L.L.Sp. = IjOwei lateral spots.

L.iS. =Leg spots.

M.D.Sp. = Mid-dorsa,l spots.

Sp. S"!). -- Spiracular spots.
T/i.r.^ Thoracic tubercle.
{7.L..<S'jj. = Upper lateral spots.
U.L. T. = Upper lateral tubercle.

as regards this variation, there is or is not segregation within the
species. Lefroy(8) when describing P. ricini (Eri worm) gives two
types of adult larvae, spotted and plain. Mr Watson however believes
this is an impure race due to a former mixture with P. canningi (Hutton).
The description of the spotless larva suggests that it has no spots in
any position, but the statement made in reference to this point is not

O. A. Mbrritt Hawkbs 137

definite. The drawing on Plate V (8), which is not sufficiently clear to
be authoritative, shows no spots.

The difference between the two species therefore is, that whilst
cynthia has seven complete series of black spots, five of these are entirely
lacking in ricini and the remaining two are generally partially present
but are variable. The condition in cynthia is therefore called spotted
and denoted by S, whereas that in ricini, which superficially looks plain,
in spite of the lower lateral series, is called plain and is represented by
P. (Plate VIII, figures B 2, C 3, C 4.)

Method of Rearing.

The larvae were reared at first in glass-lidded boxes, and later in
zinc cages covered with black net. The room used for breeding pur-
poses was heated by steam pipes and kept damp by water boxes on the
pipes and damp cloths over the open cages. The minimum of handling
was resorted to — I soon dropped the usual method of moving the young
larvae by means of a camel's hair brush, as it seemed unsuitable to a
larva so well provided with setae, preferring the more .tedious method of
waiting until the larvae had crawled on to a leaf or stem which could
be moved by forceps. The greatest care was taken to keep the boxes
clean ; when changed they were sterilised with boiling water and in the
later stages, when disease appeared, by boiling water plus carbolic.
Periodically the larvae were all turned out of the room and the room
was sterilised with a formalin apparatus, but as the experiment and the
War progressed coincidentally, it became impossible to obtain the
formalin and finally, after nearly two years, there was much disease.
The larvae were doubtless very susceptible to disease as, owing to
inadequate warmth, they were eating little and growing slowly.

Cynthia and ricini were crossed by Mr J. W. Watson at Manchester.
He divided the eggs between a number of entomologists, giving me
180 eggs from which 149 adult larvae were reared. Unfortunately, the
other breeders reared only a very small proportion of their larvae and
kept no record of the characteristics.

The Fi generation.

All the 149 larvae of this generation had spots on the upper part
of the body ; most had the full complement of spots characteristic of
cynthia and were therefore spotted (S), but some had only a portion of

10—2

138 Liheritance in Philosamia Hybrids

these spots and were called "reduced spots" {R8). The two types were
distributed among the 110 imagines as follows :

RS

38 28 8 36

Totals 66 44

This table shows a preponderance of males among the spotted forms,
but further investigation showed that the character was not sex-limited.

TABLE I.

Table to show results of various matings in the F^ generation. -

(?? (?? <?? (J?

Types of matings... S<->^ -^fxR-S . SxBS RS ^ RS

Types of larvae ... P S RS P S RS P S RS PS RS

M 5.6.7.12 ... 76.194.56 _ _ _

58. 125. 29 — — —

Mi.ll — 40.138.36 — —

— 19. 63.18 — ■ —
M 4 — — 18. 44. 23 —

_ — ,14. 25. 13 —

1/3.9.10 _ _ — 64.123.49

— • — —29. 51. 86
Ratio of P to SxRS 1 : 3-28 1 : 4-35 1 : 3-72 1 : 3-18

Ditto l:2-65 1:4-26 1:2-71 1 : 3-0

The first column contains the numbers given to the matings and used throughout the
records ; thus M 5 means mating five.

The ordinary numerals refer to the larvae and the italics to the imagines.

The Fi generation.

Twelve matings produced twelve F^ families, which consisted of 8,
RS and P types of larvae in the proportions 3 (S x RS) to 1 P. Every
variety of mating (see Table I) was made to test whether S and RS had
different hereditary values, but these few matings gave no indication of
any difference. It is equally clear that no mating took place between
homozygous forms of either /if or RS. This generation only showed that
some type of spotted form was dominant. The particular S and RS
parents used in these particular'matings appear to have the same inherit-
ance value, as judged by their offspring.

It was observed that there appeared a regularity in the disappear-
ance of the spots, for example, when only one spot was lacking, it was

O. A. Merritt Hawkes . 139

usually from the mid-dorsal region of the second abdominal segment.
The mid-dorsal spots had usually entirely disappeared before the upper
lateral series were affected, although there might be a simultaneous
diminution in the size of the spots all over the body. This suggested
that one was not dealing with spottedness (S) as a simple character but
as a compound of a large number of characters, each of which was related
primarily to a definite segment and subsequently to a segment as a part
of a longitudinal series.

The above statement, " every variety of mating," must be qualified—
the RS condition represents the loss of anything between 1 and 140
(approx.) spots. On account of the small space and limited amount of
food at my disposal, I have had to experiment as if the S condition acted
as a heritable unit ; ideally, the matings should have been made between
each different type of reduction. But as the S and P types occur
naturally, the method of experiment used may be considered as a useful
and justifiable introduction to further detailed and more analytical ex-
periments. Also, as the lower lateral row of spots occurs in riciiii and
the upper rows are absent, may it not be that there is a connecting link
between the gens for the upper rows of spots, which has, somewhere
in the past, made them act as a unit, in appearance or disappearance,
independent of the lowest rows of spots ?

Table I shows that the proportions among the larvae and imagines
are approximately the same, in spite of the numbers, which appear so
small when compared with those of the Japanese workers. This means
that, under domesticated conditions, there is no selective death rate as
regards spotted and plain larvae in the pupae : this in an important point
in deducing results from an animal which appears in three stages in its
life history and where, in artificial conditions, the death rate is so great.
It should not be forgotten however that, even in its native haunts, the
death rate must be very high. Thus Crampton(2) states in reference
to Philosamia cyrdhia (advena) — " the perfect imagines constituted only
166°/o of the whole number of individuals which entered the cocoons,
from which we may gain an idea of the severity of the conditions under
which the quiescent pupa exists."

In the early stages the majority of deaths are due to the strain at
ecdysis, but in the later stages, deaths are mostly due to disease. An
attempt was made to find if there was a differential death rate for the
S, RS and P types during the first instars, but it was unsuccessful.
The larvae are all spotted when they emerge and become RS and P at
later ecdyses ; but whereas these changes take place at regular and

140 . Inheritance in Philosamia Hybrids

known stages in the pure-bred larva, in the hybrid the reductions took
place at various stages in the same and in different families. Although
Table I shows that there is no differential death rate in the pupal stage,
one cannot make any deduction from it as regards the larval death
rate.

The plain (P) larvae usually appeared in the third and fourth
instars, but in one family the change took place in the second instar.
Lefroy(8), p. 21, writes: "From the second stage onwards, the black
spots are retained by some larvae, while in others they entirely dis-
appear in the second stage : in a few, a few black spots may be present
faintly till the third stage and then disappear." This irregularity sup-
ports Mr Watson's contention that the commercial Eri worm is a hybrid
and hence makes it imperative that scientific work should be done only
with the wild ricini, collected from the forests.

The R8 condition is particularly difficult, as close observation showed
that during the last stage itself spots may disappear. Entomologists
usually consider that any larval change in colour and marking only takes
place at an ecdysis, but that is certainly not the case in this hybrid. It
is this which has made the work of separating the S from the RS very
difficult and has allowed possibilities of error to creep in. Thus a larva
labelled 8 may, between examination and spinning-up, have lost some
spots and should therefore be called RS, or again, a larva labelled RS
may, during the same period, have lost all its few spots and hence be a
P. The latter occurrence is much rarer than the former. The daily
examination took place in the morning, on account of the light, and the
spinning usually began in the late afternoon or evening ; as the changes
noted took place between the daily examinations, they may have taken
place without my cognisance before spinning up.

Healthy larvae are covered with a white powder which does not hide
the black spots ; nevertheless, in all intermediate (RS) conditions the
powder was brushed off before the final decision was made.

The Fa generation.

Nineteen matings were made to produce this generation, seven were
between plain parents and the others were as follows :

S X S—H matings (if 27, 15, 17, 42, 47, 41, 25, 30).
S X S—2 matings {M 24, 22).
RS X RS~2 matings (M 49, 52).

O. A. Merritt Hawkes 14r

No "matings produced only full spotted (S) larvae, but three raatings
(30, 47, 17) produced 8 and R8 larvae but no plain ones — -the results
of these matings are as follows :

Types of larvae

P

s

RS

Mating 30 ...

0

34

14

„ 47 ...

0

19

4

„ 17 ...

•

0

6

6

Totals ... — 59 24

The history of mating 30 is interesting, for it had a series of
descendants,, all of which were 8 and RS, none being F, until in F^ there
appeared a family consisting entirely of .Si larvae (see Table II). Evi-

TABLE II.

Table to show the Inheritance of a Pure S Family.

P (ricini) x S {cijnthia)

I
!— ^ *

r

I I

Fi SI (67) X .9 1 (68) and RS larvae

, ^-r

II
i?2 8 6 (22) X S 6 (19) and RS and P larvae

I " r-

RS 30 (15) and S and RS but no P larvae x Sf 42 (7) and RS and P larvae

I I

S 75 (12) X S 75 (11) and RS but no P larvae

I

S 93 (1) X S 93 (2) and RS but no P larvae

. j^^ The offspring of mating 107 which consisted entirely of S larvae

The formulae used are to be interpreted thus : for example, RS 30 (15) and S and RS
but no P larvae means, the individual was a larva with reduced (RS) spots, the offspring
of mating 30, of which it was the fifteenth member, and in its fraternity were produced S
and RS, but no P larvae.

dently the parents of this mating (30) had all the reproductive cells
supplied with some factors or gens for spots, but it was only in the .^5
generation that two parents were chosen, which had all the gens which
are necessary for a complete (>S^) or cynthia type of larva. It ma}'^ be
that the results would have been reached more quickly if there had been

142

Inheritance in Philosamia Hybrids

more brother and sister matings, but these were not always possible, as
in each family the males tended to emerge some days before the females
and as the time during which the males will mate is only a few days
even when retarded by dark and cold, no matings could be effected ;
secondly, I did not wish at first to go contrary to the advice of ex-
perienced entomological breeders who were sure that inbreeding was
disastrous for moths. ,

The F^ generation.

Ten families were reared but only three produced only S and R8
larvae, the results being as follows :

Types of larvae ... P S RS
Mating QO RSxRS ... 0 8 5

64 SxRS
76 RSxS

12
15

Totals

36

12

The proportion of ^S* to RS is very different in this generation from that
in F3, but the numbers here are excessively small and the constitution is
doubtless different.

Mating 64 had no fertile descendants and so the family came to an
end. Mating 75 is the descendant of mating 30 mentioned above.
Mating 60 is interesting as a demonstration that the RS form may be
either homo- or heterozygous as regards the character of full spots.

The probability is that full spots are due to a series of unit characters,
which are linked to one another rather more closely than they are to
other characters. To demonstrate this would have required much more
space than was at my disposal.

Mating 60 was produced by mating ^^ 25 (3) RS with $ 22 (7) RS;
a large number of eggs were laid, but there was great mortality at the
first three ecdyses. No P larvae were produced. Only 13 larvae spun
up, eight S and five RS, and of these few only ten moths emerged, viz.
1 S,S RS. If RS is a constantly heterozygous form, some P larvae would
be expected in the family. But as there was such a large mortality
all the P larvae might have died, especially if there was an extra large
death rate for the P type. Two members of this family, 60(5) and
60 (6), mated (mating 88, F^ generation), and produced all types of off-
spring, viz. S, RS and P, hence one of the grandparents was a hetero-
zygote. But, when another member of the family, 60 ( 1 0), mated with the
heterozygote 79 (23) S, in mating 90, only S and RS larvae were pro-
duced. This suggests that the other RS grandparent was a homozygous

O. A. Merritt Hawkes 143^^

form for full spots, otherwise P larvae would have been found since 79(23)
is a heterozygous form although an 8 larva, being produced by the
union of an R8 and a P parent. The heterozygous nature of this R8
parent is shown by the number of the offspring — 40 P to 44 iS + R8,
approximately equal numbers of the recessive and the other types. The
heterozygous nature of these 8 types was proved by the P, 8 and R8
larvae which appeared in the subsequent generation (matings 92, 91,
84).

One of the disadvantages in breeding animals with a short life is
that they are dead long before one knows their genetic content, as
judged by their progeny — on the other hand the families should be large
and hence certain judgments can be made as a result of a study of the
fraternity.

The Fs generation:

In this generation, for the first time, I did a large amount of fraternal
inbreeding, but the results, as far as the numbers were concerned, were
very poor, thirteen matings producing only 171 imagines. A much
smaller number of eggs were laid — in one case as few as 25 — but
this may have been due to the fact that I was also doing selective breeding
for a smaller type of moth. There was also a much reduced percentage
in the number of eggs hatched, which was as low as 25% and never
higher than 45 "Z^. In this generation only one mating (90) was not
fraternal, but that also was non-productive ; only 45 eggs were laid of
which 40 hatched, but in the end only one imago resulted.

The fraternal matings (90, 93) produced only 8 and R8 offspring.
Three matings between brother and sister, both 8 (84, 85, 91), produced
all types of larvae.

The Fs generation.

This generation has been a failure for several reasons ; mice ate more
than two-thirds of the F^ cocoons, the moths that did emerge could not
be kept, on account of war conditions, at an adequate temperature, very
few eggs were laid, and finally the larvae grew so slowly in the cool
breeding room that many succumbed to disease. This generation pro-
duced one family (107), whose adult larvae all had the full complement
of spots. The larvae, numbering 23, were about to cocoon when they
died of cold. As the number of larvae which hatched was only 50,
the 23 which reached maturity were probably a good average lot, so
that one is justified in supposing that at last two parents with all the

10—5

144 Inhentance in Philosamia Hyhrids

gens for full spots were isolated and mated. It is clear that forms
which are homo- and heterozygous for the full number of spots may
appear either as aS or as RS larvae. It is possible that P. cynthia
may have two forms, one with the full range of spots and another or
others in which certain spots are missing; no picture or description
indicates that this is the case, but as entomologists seek " type" such
variations may well pass unnoticed.

No families have been bred which produced only R8 forms. Eleven
matings have been made between R8 parents in F2, F^, F^,a,nd F^ genera-
tions, nine of which produced all types of larvae, the proportion being
as follows : S \Q2 : US ^1 : P 90, or SxRS:P= 249 : 90, approximately
the same proportions as were found in the F^ generation.

Revieio of literature.

Lefroy at Pusa(8) evidently made experiments to segregate spotted
and unspotted larvae, for he writes, p. 22 : " The offspring of either has
been seen to be spotted and unspotted mixed. If only unspotted ones
are bred, the majority tend to be unspotted. On the other hand, if
spotted ones are bred, the majority tend to be spotted... it is possible to
eliminate either spotted or unspotted worms wholly... of worms from the
eggs of these moths (spotted), more than 50 "/^ (but not all) were black
spotted." These observations were made from mass matings, not from
individual matings as in this present work, hence the somewhat confused
statements. It would be surprising to find that Lefroy was right in his
statement that spotted larvae appear in the progeny of unspotted forms.
He makes no distinction between completely and partly spotted larvae.

Kellogg(7) crossed white (plain larvae) and inoricaud types of silk-
worm, and as a result of his work he says, p. 16 : " White is regularly
recessive to all the other larval colour pattern types. And white larvae
mated with white never produce any but white larvae." He does not
distinguish between a homo- and a heterozygous type and indeed, con-
sidering the nature of the markings, it does not seem possible to do so,
except by breeding and judgment from the offspring. Whilst Kellogg
found that white cocoon colour might be dominant or recessive, according
to race, the white colour of the larva was always recessive. But
Tanaka(ll) states, p. 24, "the quail factor is partially dominant and
partially recessive to the plain. These facts show that the relation of
dominance and recessiveness is more complex than is generally sup-
posed." The Japanese workers have made the following general state-

O. A. Merritt Hawkes 145

ment(12), p. 148 : " In the first generation of the crossing between white
and common marking worms, all the worms were of the common marking
and in the second generation two kinds appeared ; of the common
marking 77°/^ and of the white 23°/^. In the fourth generation the
worms produced by the mating of the white worms were all white." This
looks like a case of simple dominance, but later work of Tanaka(ll),p. 211,
states: "There is great variation of pigment intensity in the normal...
type. In this we find almost every gradation from the lightest to the
darkest, apparently presenting a continuous variation.... Provisionally I
have divided the normal, according to the heritable characteristics into
four sub-types; namely normal 1, 2,3, 4." At the bottom of the series
is a form " hardly distinguishable from the pure plain " and he further
adds in a footnote, p. 211, "I assume the existence of different genes
respectively for the different sub-types of marking. and colour," Now
this is what I also am inclined to do as regards the series of spots,
although it is stated already that there have not been enough experi-
ments done to prove the point, but it is strongly suggested by the fact
that the spots disappear in an orderly and continuous manner. The
experiments in Japan by Tanaka, Toyama, etc. and those above recorded
all confirm ttie statement of Kellogg, p. 68 : " In larval colour-pattern
characters, the inheritance behaviour is rigorously alternative and
Mendelian, dominance always being consistent in relation to a given
colour pattern as related to another."

Experiments have been made in crossing certain allied Saturnians ;
thus Joutel(6) and Pollard crossed Philosamia cynthia (advena) and
Callosaniia promethia, but as this cross was made for the sake of the
imago rather than the larva, the results with regard to the latter are
not stated in detail, but it seems clear that there is a sex-linked inherit-
ance as the reciprocal crosses gave different results. Soule(lO) also
crossed these species, but her findings are different from those of Joutel.
Her results are shown in Plate LXXI of Packard (9), which shows plain
larvae as a result of the cross between a plain (P) and a spotted (*S)
parent. This indicates that although P. cynthia advena and its Chinese
relative P. cynthia (Ning-po) appear so alike, genetically they are very
different — if these results are confirmed it will be an interesting point
in insect heredity.

146 Inheritance in Philosamia Hybrids

Section II. The Inheritance of the Plain Larvae.

The plain larvae, as stated in Section I, appeared in the F^ generation
in the proportion of a simple Mendelian recessive, viz. one of the reces-
sive to three of the other types.

The matings of plain to plain produced only plain larvae ; of such
matings, there were six in the F^ generation, one in the F^, seven in the
F-a, and six in the F^ generation.

These results agree with those of Kellogg and the Japanese workers
(see p. 144).

Mr J. W. Watson crossed P. ricini (^ (plain larva) with P pryeri %
(spotted larva) and produced a hybrid which he called Rothschildi (13).
These F^ larvae were all spotted, there being both S and RS forms. He
inbred the Fi generation and gave me some of the eggs. These &g^^
produced an F.2 generation which consisted of P, >S^ and RS larvae in the
proportion of 1 P to '6Sx RS. Only one mating was made between
P parents, but that produced only P larvae. In this mating with pryeri,
ricini behaved genetically exactly as in the mating with cynthia.-

Section III. On Larvae with reduced Tubercles (Scoli).

P. cynthia and P. ricini have throughout their larval life six com-
plete longitudinal series of tubercles, three on each side, the upper rows
being the best developed ; the lowest row is shorter than the others. In
the three thoracic segments there is also present a fourth bi-lateral series
of tubercles, situated at the base of the thoracic legs ; these are very
short, especially the first pair, which may be reduced to mere knobs.
The setae borne by the tubercles vary in proportionate length in the
various stages, being fewest and shortest in the last. (Text-fig. 1.)

. The tubercles of the two species are alike in all anatomical features,
but differ in colour, those of cynthia (Ning-po) having a slightly pink tip.

The tubercles in P, were normal in appearance, but among the twelve
families of Pg five matings produced a few larvae with abnormal tubercles.
These larvae were very variable, the tubercles being entirely absent,
very short or varying in length on the various segments of the body.
(Plate VIII.)

Before proceeding to a statement of the breeding experiments, it
will be as well to define the term tubercle, which is often used very
vaguely, and hence denotes a variety of larval appendages. Fracker(3),
p. 44, states that the term " tubercle has been used to mean any cuticular

O. A. Merritt Hawkes 147

projection of the body wall, from a minute papilla to a conspicuous
prominence," but he considers that the word tubercle should be used,
p. 16, "as a general term to indicate the location of a seta or a group of
setae or a process of the body wall bearing such a group." Fracker has
given the name scoli to the particular tubercles which occur in Satur-
nian and Nymphalidian larvae, scoli being defined as thorny processes
bearing spine-like setae.

The three upper rows of scoli of both cyvthia and ncini are, to use
Fracker's words, " well developed, conspicuous, cylindrical higher than
wide," but the infra-spiracular series are short and the thoracic are
dumpy.

The larvae with reduced tubercles appeared after the last ecdysis and
seemed, within their individual lives, to reverse the process by means of
which the typical Saturnian scolus has been developed within the
species. Fracker(3), p. 44, states that the scolus has been evolved from
papillae, i.e. " setae surrounded at the base by a small chitinized ring."
The larvae which had the most perfectly reduced tubercles showed in
some segments, on careful examination, traces of the chitinized ring,
but the setae had disappeared, just as they have disappeared on the short
infra-spiracular papillae of the normal larvae.

The number of abnormal larvae in F2 was small, only 47 occurring
out of a total number of approximately 900 adult larvae. Only the
adult larvae can be counted as this condition appeared in the last instar.
These 47 larvae were S, RS, and P forms, of both sexes and gave rise
to both dark and light moths ; this particular phenomenon was therefore
not correlated with any of the other characters studied in this series of
papers.

In order to know the anatomical value of the scoli, Dr A. D. Imms,
of Manchester University, very kindly cut sections for me, as I did not
have the adequate apparatus at my disposal. The scoli (Fig. 2) consist
of two parts, a basal portion with a greatly thickened cuticle, pre-
sumably for support, and a distal portion over which the cuticle is
thinner and quite smooth. This cuticle is evidently different from that
on the basal portion and the general body surface, as it stains much less
deeply. The hypodermis consists of columnar cells, among which are
the large trichogenous cells with their large flattened nuclei — this layer
is at places invaginated to form what are apparently glands, although
they are not connected with any opening in the cuticle.

The general body cavity extends into the distal portion of the
tubercles, through a somewhat narrow passage, which perhaps acts as a

148

Inheritance in Philosamia Hybrids

B. C. = Body cavity.
7i.C'e. = Blood cells.
C.C. = Corrugated cuticle.

C, S. = Smooth cuticle.
C. Se. = Cell (trichogenous) which pro-
duces a seta.

J?.J5.C.= Extension of body cavity
into the distal part of the tubercle.

^.7^. = Fat-body.

G. = Gland of unknown nature.

Jf?/. = Hypodermjs.

L.M. = Section through muscle.

^. = Nerve.

N. T. = Narrow passage between gene-
ral body cavity and that of the
distal part of the tubercle.

Se.^Seta.

T. = Tracheae.

T.M. = Transverse sectionofaniudole.

B.C.
Fig. 2.

Semi-diagrammatic drawing of a well developed tubercle (Scolus).

1

O. A. Merritt Hawkes 149

valve in the erection of the tubercles. The tubercles are well supplied
by tracheae and nerves. The fat-body and muscles extend into the
basal, but not into the distal portion of the tubercles. The tubercles
are similar to those of the Tortoise-shell butterfly {Vanessa urticae),
described by Berlese(l). This being their anatomy the tubercles are
not arsimple character like colour, etc., but they have a certain anatomical
importance, and must be represented in the chromosome by a large
number of correlated factors.

An attempt was made to mate these abnormal forms, but two were
never fit for mating at the same time and it was not until the following
generation {F-^ that resort was finally made to a mating between a
normal and an abnormal larva. The abnormal form used in the suc-
cessful mating (m. 79) was not one with perfectly reduced tubercles, but
had them represented only by stumps (papillae) except in the infra-
spiracular row, where the position of each tubercle was indicated by a
chitinous ring. In this family 86 adult larvae were reared (see
Table III). These larvae all had normal tubercles, the abnormal form
had disappeared. These were then inbred, ten families being reared,
and amongst them, larvae occurred with long, with short, and with no
tubercles — an obvious Mendelian segregation. The individuals with
some or all tubercles absent were separated from those that had tubercles,
but the long and short tubercled forms were not separated from one
another — partly because of lack of space and partly because I had no
satisfactory standard of size by which to effect a satisfactory separation.
The tubercles vary in distention and also in individuals. Unfortunately
the completely non-tubercled forms failed to mate with one another.
A large number of the long and short tubercled forms were eaten by
mice, and those that appeared as imagines did not mate well and laid
very few eggs. These results, as stated in the earlier section, may have
been due to inbreeding for five generations or simply to the unsatis-
factory heat of the breeding room. As a result of breeding in F<i four
small families were reared (98, 96, 100, 103). In two of the families, 100
and 103, the larvae all had short tubercles, but no tubercles were absent.
In another family (96) most of the larvae were normal, but a few had short
tubercles. In mating 98 one, of the parents (83*7) was fi-om the cage
which contained a mixture of larvae with both long and short tubercles,
and the other was from a larva (81*21) with reduced tubercles; this
larva was as follows : the two rows of dorsal tubercles were very short,
those on the seventh and eleventh segments being represented only by
minute protrusions, the other rows of tubercles were not quite as short

150 Inheritance in Philosamia Hybrids

as those just described, but on the left side of the seventh segment there
was no infra-spiracular tubercle. The offspring from this mating had
all the tubercles present, but the upper lateral and the infra-spiracular
were only stumps. Again there is an indication of Mendelian segrega-
tion, but of course further breeding is necessary. The weather conditions
were so severe early in 1917, that privet, the food on which the larvae
lived, could not be obtained and the families died out.

TABLE 111.

Normal larva x Abnormal larva

I

Fi 86 Normal larvae F\

\

Fi 10 families which had normal tubercles, short tubercles and no tubercles F^

\

F3 m. 96 (nearly all normal) ; m. 100 and m. 103 (short tubercles); m. 98 Fq

(reduced tubercles)

There is no recorded direct ancestor of the normal or abnormal parents used in this parti-
cular mating which had reduced tubercles in the three proceeding generations reared. But
they both had ancestors in the five original families which produced the tubercled forms.

The generations on the left side refer to the breeding for the absence of tubercles,
whilst those on the right side are the total generations bred.

Not only from the general heredity view-point but also in relation
to the question of the origin of certain unusual forms in the Saturnidae,
these experiments are interesting.

Most of the Saturnidae have larvae with tubercles, but a few, Cal-
losamia angulifera, Brahmaea japonica, RJiodia fugax, etc., have no
tubercles in the last instar. Apparently the smooth- form has arisen as
a mutation from the tubercled forms within the group. What is the
relation of the abnormal forms with reduced tubercles which occurred in
the F^ generation to the smooth forms in the group of Saturniids ? Did
they arise because there was in the parent cynthia or in the parent
ricini a dormant gen for smoothness which appeared when inbreeding
took place in F^ or was the original cynthia or ricini from a larva which
had mutated and then passed on the character ? Presumably the gen
was not present on both sides or some non- tubercled forms would have
appeared in the F^ generation. As soon as material can be obtained the
cross will be repeated, when it will be known if one is dealing with a
mutation or not. Among the silkworms, normally with caudal horns,
with which Kellogg experimented, were found a number without a horn ;
these he mated, but he states(7), p. 58, " that the character is not

%

I

O. A. Mbrritt Hawkes 151

heritable ; it does not behave Mendelianly and cannot be fixed."
Packard(9), p. 152, writes: "The single median tubercle on the eighth
abdominal segment of the specialised Saturnian larvae represents the
' caudal horn ' of Bomhyx mori and is evidently the fusian before the end
of embryonic life of what was originally two separate tubercles." The
tubercles of this cynilna x ricini cross and the caudal horn being thus
equivalent structures, it will be interesting to compare their inheritance
behaviour when it is fully worked out, although the indication so far is
that the smooth form in this hybrid does behave as a Mendelian reces-
sive, I do not know if the structures are anatomically the same, but
even if they were it would not necessarily follow that they would behave
in the same way in a crossing. Thus one may instance the variability
of the dominance or recessiveness of white cocoon colour, and again,
whilst I find that the black spots of the cynihia larva are dominant over
the plain condition of the ricini larva, Soule(lO) found the black spots
of the nearly related P. cynthia advena were recessive when crossed with
the plain Callosamia promethea.

Tanaka(ll), p. 197, mentions a knobbed race oi Bomhyx mori. The
knobs, he writes, " are large paired evaginations of the skin formed in
the sub-dorsal line."..." No permanent peculiarity is seen in the larvae
until they have passed through the third moult, when the knobs come
forth for the first time, and their full development is attained only in the
last stage of larval life." He found segi'egation of knobbed and smooth
fi-om the hybrid, but does not say whether the extracted dominants and
recessives remain pure.

This character appears to be very different from the tubercles of
Philosamia or the caudal horns of Bomhyx mori, as it is an addition
which originates late in larval life, and Tanaka continues, p. 197, " the
remnants of certain knobs are still visible in the pupal and imaginal
stages." The tubercle belongs to an early stage of Saturnian evolution,
but this knob is a mutation added so late that it is even ceasing to be a
purely larval structure and is carried over into the imago. There may
also be a fundamental difference between the two, for whilst Tanaka
writes of the knob as an " evagination of the skin," the tubercle is an
evagination of the body wall related to a definite group of setae.

Packard (9), p. 151, writes: "The larval characters of this gi-oup
tend to show that the family has originated fi-om a spiny group and most
probably, when we take into account the transformations of Aglia tau,
fi-om the Ceratocampidae....In the European Saturni...we have the
perhaps most generalised and primitive members of the gi-oup."..."It is

152 Inheritance in Philosamia Hybrids

iiot unreasonable to suppose that all the more specialised genera must
have been derived from a Saturnia-like ancestral form, i.e. a larva in
which all the tubercles, whether the thoracic or the abdominal, were of
the same shape, size and colour on all the segments ; those on the
eighth abdominal segment being of the same number (6) as on the
segments in front"; p. 19, "The larva o( Aglia after its last moult
becomes like a Saturnian, possibly of the type of Telea and especially
like a smooth form of the larva of Attacus hetis (Walker) (also called
Rothschildia), It is clear that the- Saturnian larva, although well
tubercled in the earlier stages, has, as a whole, a less complicated tubercle
in the last stage than the Ceratocampidae and that it is only in the
Saturniidae that a number of forms occur which have much reduced
tubercles in the last stage."

The complete reduction of tubercles (except on one segment) has
reached its limit in two forms, one belonging to the Ceratocampidae,
viz. Cercophana frauenfeldii (venusta variety) and one of the Satur-
niidae, viz. Rhodia fugax. In appearance these two are remarkably
alike and so abnormal in appearance that they must be regarded as
highly specialised forms which have approached one another perhaps
because of a similar environment. These forms being so highly
specialised it is necessary to seek the relation of the present non-
tubercled hybrid, if any relation exists, among the more ordinary forms
which occur in the Saturniae in which the tubercles are much reduced
rather than entirely absent, such as Telea polyphemea, Gallosamia
promethea, Tropea luna, Aglia tau, etc.

Apparently a form with simplified tubercles, similar to Aglia, arose
as a mutation within the Ceratocampidae, and from this form some of
the Saturniidae were certainly descended — it is difficult when a study is
made of the larvae, to suppose that all the Saturnians are derived from
the same ancestor since they are divided into two natural groups by
the condition of the tubercles on the eighth abdominal segment.

Has there been a repetition de novo in this hybrid of the same
phenomenon which produced Aglia tau, or through the ages since the
latter originated, have the gens for a reduced tubercular condition in
the last instar, been waiting in ricini or in cynthia ?

My thanks are due to the Natural History Society of Birmingham
and to the British Association for grants in aid of this research.

JOURNAL OF GENETICS, VOL VII. NO. 2

PLATE VIII

Fig. Al.

Fig. A. 2.

iP^^^^^^^^^-

Fig. CI.

Fig. C2.

Fig. A3.

Fig. A 4.

Fig. C3.

Fig. C4.

Fig. B 1.

Fig. B 2.

O. A. Merritt Hawkes 153

• Summary.

1. The plain or non-spotted condition of the larva of P. ricini is
recessive to the spotted condition of the larva of P. cyntlivi (Ning-po
variety).

;: 2. The dominance in the F^ generation is incomplet.e ; all the larvae
are spotted but not all have the full complement of. spots. ■

3. " The spots are probably represented, not by one, but by a group
of related gens in the chromosome.

• 4. The dominant homozygous and the heterozygous forms can bo
distinguished only by breeding, as both may be either full- or partly-
.spotted.

5. A number of abnormal larvae with reduced tubercles occurred
in the F^ generation; when mated with normal larvae, this character
behaved as a recessive; whether it was a mutation which occurred de
novo as a result of this particular cross or whether it is a result always
to be expected from this hybrid, has not yet been determined. This non-
tubercled mutation (?) is interesting in relation to the few non-tubercled
species among the Saturniids.

EXPLANATION OF PLATE VIII.

figures A 1, A 2, A3, A 4 are larvae of F-i-

A 1 — \ natural size. Normal tubercles. Full spots (S).

A 2 — f natural size. No trace of tubercles. Full spots (S).

A 3 — I natural size. Irregular tubercles. Full spots {S).

A 4 — \ natural size. Slight vestiges ol tubercles. Reduced spots {RS). .

Figures B 1, B2 are larvae of Fi and F.^ of the cross between a normal and an abnormal
larva. (See Table III.)

B 1. — Natural size. Normal tubercles. Full spots (S).

B 2. — Upper larva — 5 natural size. Traces of tubercles as chitinous rings. No spots (P).
Lower lafVa — J natural size. Normal tubercles. No spots (P).
i- igures C 1, C 2, C 3, C 4 are laivae of F-t, . They were blown by Mr J. W. Watson, f natural
size.

C 1, C 2 are lateral and dorsal views of a nou-tubercled, full-spotted larva (S).
C3, C4 are lateral and dorsal views of a tubercled, plain (P) larva. This specimen
shows some spots of the lower lateral series on the abdominal legs..; The large
spots at the side are the spiracles.

154 Inhei'itance in Philosamia Hybrids

LITERATURE.

(1) Bkhlehe, a. Gli Insetti, Vol. i. p. 477, fig. 529, 1909.

(2) Crampton, H. E. " Experimental and Statistical studies upon Lepidoptei-a.

I. Variation and elimination in Philosamia cynthia." Bioinetrika, ill. pp.
113 — 130. " II. Variation and Selection in Saturnid Lepidoptera. Philo-
samia cynthia." Biol. Bull. vi. pp. 310, 311.

(3) Frackee, Stanley Black. "The Classification of Lepidopterous Larvae."

Illinois Biological Monographs, No. l. Vol. ii. July 1915. Published at
Urbana, 111.

(4) Frings. " Saturnia Hybrids (Zurich). Casparii." M. Soc. Entomol. Zurich,

Vol. XXI. pp. 25—27, 1906.

(5) Hawkes, O. a. Merritt. "The Effect of Moisture upon the Silk of the Hy-

brid, Philosamia (Attacus) ricini (Boisd.) (^ x Philosamia cynthia (Drury) $ .
Journ. Exp. Zoo. Vol. xxi. No. i. July 1916, pp. 51—60.

(6) JouTBL, L. H. " Philosamia cynthia x Callosamia promethea crosses," pp. 101 —

103. Journ. New York Ent. Soc. Vol. xv. 1907.

(7) Kellogg, Vernon L. Inheritance in Silkworms. I. Stanford University,

California, 1908.

(8) Lefroy, Maxwell H., and Ghosh, C. C. " Memoirs of the Department of

Agriculture in India." Eri Silk Entomological Series, Vol. iv. No. 1, May
1912.

(9) Packard, A. S. (A) " Monograj^h of the Bombycine Moths of North America."

Part III. First Memoir of Vol. xii. of the Memoirs of the National Academy
of Sciences, Washington, 1914.
(B) Ditto, Part i. issued as Vol. vii. in 1895 contains "Hints on the Evolution
of the Bristles, Spines and Tubercles of Notodotian and other Caterpillars,"
p. 11, which is reprinted with some alterations from an article in the Proc.
Boston Society of Nat. History, Vol. xxiv. 1890, pp. 482—515, 556—559.

(10) SouLE, C. G. (A) " Some experiments on Hybrids." Psyche, Boston, Vol. ix.

pp. 411—413. 1902.
(B) Ditto. Ps?/cAe, Vol. xiv. pp. 116, 117. 1907.

(11) Tanaka, YoSKlMARO. "Genetic Studies in the Silkworm." Journal of the

College of Agriculture. Toboku Imperial University, Sapporo, Japan, June
1916. In this is found complete references to all the Japanese work on
Silkworms.

(12) . A general Report of Sericidtural Investigation. The Imperial Seri-

cultural Institute at Tokyo, 1910.

(13) Watson, J. W. " New Saturnidse." Transactions and Annual Report of the

Manx:hester Entomological Society, 1913. Plate I.

f

^

I

Volume VII MAY, 1918 No. 3

ON THE STERILITY OF HYBRIDS BETWEEN THE
PHEASANT AND THE GOLD CAMPINE FOWL.

By D. W. CUTLER, M.A. (Cantab.),
Assistant Lecturer in Zoology, Manchester University.

(With Plate IX.)

Our knowledge of the changes that occur in the constituents of the
nucleus during the maturation of the reproductive cells of animals and
plants, renders the cytological study of the gonads of hybrids of great-
interest and importance. However, it is only within comparatively
recent times that such investigations have been undertaken. Guyer's
work on the gonads of hybrid pigeons was one of the first made, and this
has been followed by the publication of the results of research on other
hybrid material.

In 1915 Dr Doncaster and I published the result of investigation of
the testes of a sterile tortoiseshell cat. The condition of the organ was
so interesting in comparison with that found in other cases of sterility,
that I very gladly undertook to study the gonads of some hybrid birds
kindly given to me by Dr Doncaster. The progress of the research has
been slow owing to the pressure of other work and some of the problems
connected with it still remain unsolved. I have, however, been able to
discover the cause of sterility, which seems to justify the publication of
this paper.

Material.

The hybrids were bred by the Rev. Lewis Jones, who crossed pheasant
males with hens of the Gold Campine breed. About a dozen chicks were
obtained, which were of two kinds :

(a) dark variety, (6) light variety.

Joum. of tren. vii - 11

156 Hybrids between Pheasant and Fowl

All the chicks, however, were males. There is a possibility that the
absence of females was accidental and that the males are rifiore viable
than the females, which died in the shell : indeed in support of this
Mr Lewis Jones says that he thinks he was able to identify one of the
offspring, which died soon after hatching, as a female.

The well-marked colour varieties present a very complicated, but
interesting problem. There are two breeds of Campine fowl, silver and
gold, which differ from one another in the fact that the silver race possesses
a dominant sex limited (sex linked) character that is absent in the gold
race. The silver female transmits therefore this dominant factor for
silver only to her sons, while the gold female transmits a recessive gold
factor to the offspring of both sexes. Since the female parent of the
hybrid chicks was a Gold Campine, the existence of the two distinct
colour varieties cannot be explained by assuming the presence of the
silver factor in some, but not in others of the offspring. If, however,
the fact that all the hybrids reared were males is not due to the mor-
tality of the female embryos, but is caused by the development into males
of eggs which, if normally fertilised by sperms of their own species
would have been females ; then it is possible that the two colour classes
of hybrids may be due to the presence of some other sex limited factor
derived from the Gold Campine female parent. Such a factor would
normally be transmitted to the male offspring only and not to the female ;
but if the potentially female eggs are transformed into males by the
pheasant spermatozoa, then in this way two colour classes might arise
among the hybrids.

Methods.

The tissues were fixed in Bouin's formal-picro-acetic mixture and in
corrosive sublimate. Both these fixatives gave very satisfactory results,
but the former seemed to be a little the better of the two. A large
variety of stains were employed, but Heidenhain's Iron Haematoxylin
was much superior to them all.

At an early stage in the work it became obvious that it would be
necessary to study the spermatogenesis of the parents used in the cross.
This I was able to do owing to the kindness of Mrs Haig Thomas and
Mr Lewis Jones, who supplied me with pheasant and campine material.

D. W. Cutler 157

Spermatogenesis of the Pheasant.

The spermatogenesis of the pheasant is exceedingly difficult to eluci-
date, owing to the tendency of the chromosomes to stick together, thus
making it very hard to separate one from another, also the linin of the
nucleus takes up the stain, though not to the same extent as the
chromosomes, making clear figures very difficult to find.

I have examined many sections of the testes, but in no case has it
been possible to determine the exact number of chromosomes in the
spermatogonial cells. It has frequently happened that the chromosomes
were all separated one from another, with the exception of one or two,
which were so closely associated that it was impossible to decide the
exact number representing the complex. However, from the examination
of many spermatogonial division figures I conclude that the chromosome
number is from twenty to twenty-two. Fig. 1 shows very well the
appearance which most of the spermatogonial cells have when seen in
polar view. When viewed from the side they appear quite typical in
character.

Primary Spermatocytes. The growth of the spermatogonia into
primary spermatocytes is initiated by the nucleus increasing a little in
size ; the chromatin is scattered through it in the form of small granules
lying on a fine network (Fig. 2). The next stage is the formation of a
thin spireme, which gradually contracts to form the synizesis stage in
Avhich the threads pass into a tangled condition. It is very characteristic
of this stage to find the chromatin threads aggregated together at one
side of the nucleus (Fig. 3). I have not been able to follow the details
of this process or to determine in what manner the bivalent chromosomes
arise. The next stage, which immediately follows synizesis, is shown in
Fig. 4 where the threads fill the whole of the nucleus and are thickened.
The question of the number of double chromosomes that arise from these
threads is again exceedingly difficult to decide. In most cases the
number appears to be either 10 or 11, but even in the best of the figures
one cannot be absolutely certain. The arrangement of the chromosomes
is, however, characteristic. In polar view there is seen a ring of chromo-
somes with one or two lying in the centre (Figs. 5, and 6). Also it is
common to find one of the chromosomes protruded far outside the others,
making it a very conspicuous element. Fig. 6 exhibits this well and
corresponds very closely with some of the figures published by Guyer in
his recent paper on the chromosomes of the fowl.

11—2

158 Hyhrids between Pheasant and Fowl

The division stages of the primary spermatocytes, when seen from
the side, do not present any features of interest, except that one of
the chromosomes is often sharply differentiated from the rest (Fig. 8)
I have not been able to find any trace of a lagging element, though many
division stages have been carefully examined.

Secondary Spermatocytes. These cells are usually a little smaller
than are those of the primary spermatocytes. There is, however, at this
stage, a great tendency for the chromosomes to undergo secondary fusion,
which is in many cases in pairs. As a result of this, in the majority of
cells, five or six clumps of chromatin are seen (Fig. 9). This fusion in
pairs, however, is by no means a regular occurrence as the number of
chromatin masses in the secondary spermatocytes varies from 1 to 7 or 8
(Fig. 10). I have not observed a secondary spermatocyte with more than
eight chromatin elements so that it appears as though a certain amount
of pairing always tak^s place.

This phenomenon is not of rare occurrence and has been recorded in
the spermatogenesis of insects and mammals. The spermatids and sper-
matozoa are of the normal type and do not require any special description.

Spermatogenesis of Gold Campine Cock.

The phases in the development of the spermatozoa are very like those
described for the male pheasant. Guyer has recently published the
results of research on the spermatogenesis of the fowl and arrives at the
following conclusions :

1. The somatic chromosome number is 18 of which two are sex
chromosomes.

2. During synapsis pairing of the chromosomes takes place so that
nine bivalent chromosomes appear in the primary spermatocytes. The
two sex chromosomes pair together, but during the division of the cell
they behave as a single bivalent, so that two kinds of spermatocytes are
formed, one with 8 + 2X chromosomes, the other with 8 chromosomes.

3. Secondary fusion occurs in the secondary spermatocytes so that
the majority of the cells contain four or five clumps of chromatin.

4. Spermatocytes with four chromatin masses frequently show
abnormal characters, which leads to the conclusion that only one class
of spermatid, that with the X chromosomes, develops into functional
spermatozoa.

D. W. Cutler 159

I am at present unable either to confirm or dispute these results as
I have not obtained figures which are sufficiently clear to make an
accurate count of the chromosomes possible. Spermatogonial mitoses
in which an accurate count can be made are exceedingly rare, but from
the few figures I have observed I should conclude that the number is
from 18 to 20.

In the primary spermatocytes the chromosomes are arranged in a
ring with one element projecting beyond the others, as described by'
Guyer, and the number is from 8 to 10.

The aggregation of the chromosomes into clumps during the growth
of the secondary spermatocytes is very obvious in the sections, but I have
not been able to find any trace of degenerating spermatids, though I
have made a special search for them.

Spermatogenesis of Hybrids.

A careful examination of the testes of the two colour classes of hybrid
birds failed to show any differences whatsoever between them : the two
types will therefore be treated under the same heading. The testicular
tubules do not differ from those of the parent species, nor is there any
increase in the amount of interstitial tissue as was found in the sterile
tortoiseshell tom cat investigated by Doncaster and myself

The spermatogonial cells are arranged round the margin of the
tubules and many are found in the dividing condition, which is in sharp
contrast with that of the parent species where dividing spermatogonia
are rare. The divisions are normal, the chromosomes arrange themselves
in a regular way on the spindle and pass to each pole without any trace
of lagging (Figs. 11, 12, 13). When seen in side view the impression is
given that good countable figures would be obtained from a face view ;
this, however, is not the case for the chromosomes stick together, as in
the parent birds, rendering it impossible to make an accurate estimation
of the number. It is probable, however, that it is from 18 to 20 (Figs. 14,
15). It is not possible to distinguish between the maternal and paternal
chromosomes.

The early development of the primary spermatocytes is normal; a
granular spireme is formed (Fig. 16) which retracts to one side of the
nucleus (Fig. 18), producing a closely packed mass of threads (Figs. 17,
19). The next stage is the one of interest in connection with the
problem of sterility for it is now that spermatogenesis becomes abnormal.
The synaptic threads, instead of breaking into bivalent chromosomes,

160 Hybrids between Pheasant and Fowl

form a varying number of irregularly shaped clumps of chromatin.
There is no uniformity as regards the number produced and a series can
easily be formed in which the number varies from two to three large
masses (Fig. 20), up to the condition where the chromosomes are almost
all bivalent, or where many have separated as univalents (Fig. 21).
This failure of the synaptic threads to form bivalent chromosomes is
evidently the cause of the sterility in the hybrids. The conditions are
•seen in Figs. 20 to 27.

This is the final stage in spermatogenesis, for divisions of the primary
spermatocytes do not occur, nor is there any trace of abnormal mitoses.
The formation of giant spermatids recorded for hybrid pigeons by
Geoffrey Smith does not take place, nor is there any evidence of the
production of multiple spindles, as found by Guyer in hybrid pigeons
and by Wodsedalek in the testes of the mule.

Discussion.

As was pointed out at the beginning of the paper, our knowledge of
the cytology of the gonads of hybrids is still small ; the results, however,
of investigations which have been made, are extremely interesting as
regards the behaviour of the chromosomes during the maturation of the
germ cells. In both animals and plants the offspring of a cross may be
completely sterile, partially so, or entirely fertile, and the evidence which
we have points to the conclusion that one or the other of the results is
largely dependent upon whether the paternal and maternal germ plasms
are able to co-operate one with another during maturation.

Guyer concluded, from investigations on hybrid pigeons, that there
was something repellent in the two germ plasms, which had contributed
to the cross, so that the chromosomes were unable to pair, causing an
irregularity in synapsis.

Geoffrey Smith obtained three sterile male birds from the cross
domestic dove female with Magpie pigeon male. Synapsis was disturbed,
but not to the same extent as with Guyer 's birds, for abnormal spermatozoa
were produced. The reductive divisions, however, were abnormal for
no bivalent chromosomes were produced. The second division was
entirely suppressed and the spermatocytes proceeded without further
development to form giant spermatids from which spermatozoa of ab-
normal size developed.

In a further paper two years later this observer and Mrs Haig Thomas
published the results of experiments on the hybridisation of pheasants.

D. W. Cutler 161

From the cross P. i^eevesii $ with P.formosanus </ a sterile male was pro-
duced, and a second one from the cross P. reevesii $ with P. versicolor ^ .

On examination of the testes of these hybrids it was found that
the spermatogonial cells were normal, but all subsequent stages were
abnormal, owing to the failure of the two sets of chromosomes to pair
during synapsis.

On another occasion from the same crosses three sterile females were
obtained. The ovaries of two of these were investigated, but there was
no trace of sexual cells, the whole organ being composed of interstitial
cells mixed with stroma and fibrous tissue.

The sterility of the male hybrids is therefore due, as in other cases
enumerated above, to the inability of the chromosomes derived from the
two parents to co-operate during maturation.

Although in the female hybrids degeneration of the ovaries occurred
before maturation, it is probable that here also sterility is assignable to
the same cause as in the males : for as G. Smith points out, in his paper
on pigeon hybrids, there is evidence that a precocious temporary synapsis
takes place in the female germ cells. Probably it was at this period that
the chromosomes failed to pair, thus leading to degeneration of the gonads.

A paper by Wodsedalek dealing with the sterility of the mule has
appeared recently. He finds that the maturation divisions are further
suppressed than in the above cases, for there is no trace of normal
synapsis, and the spireme condition which usually follows it is entirely
lacking. The conflict between the paternal and maternal chromosomes
begins in the primary spermatocytes leading to the production of abnormal
cells. The spermatogonial chromosomes are 51 in number, but in the
spermatocytes the number varies from 34 — 49. Thus it appears that
there is some attempt on the part of a few chromosomes to pair. Wodse-
dalek found, however, that the most pronounced signs of decay were to
be seen in the cells where pairing had been greatest.

Finally there is the case of sterile moths, hybrids of the genus Bis-
toninae, which has been investigated by Harrison and Doncaster. The
parents of the hybrids were Bistort hirtaria and Nyssia zonaria. Recip-
rocal crosses were made but the resulting offspring were all sterile. An
interesting change in the sex ratio was produced according as to the
manner in which the cross was made; this will be dealt with later in the
paper. In both spermatocytes and oocytes the first stages in synapsis are
normal, but there is no production of a thick coiled spireme. Complete
pairing of the chromosomes does not occur and only a few are able to find
mates. It was estimated that in the cross zonaria female with hirtaria

162 Hybrids hetiveen Pheasant and Foml

male about 10 chromosomes paired, but in the converse cross the number
was greater.

The reduced number in zonaria is 56, while that of hirtama is 14.
It will be noticed that the number in the latter species is four times
that of the former, and Doncaster suggests that in hirtaria the chromo-
somes are compound and the zonaria units are here united together,
but in a different order. If this is so the further suggestion is made
that sterility is due to the chromosomes not being able to pair normally
because of the different grouping of the units.

In connection with the foregoing, the results of Federley's work are
of interest. Pygaera curtula male and Pygaera anachoreta female were
mated and the offspring, which were fertile, showed certain characters
of one species dominant and other characters intermediate. On crossing
the hybrids back to one of the parent species segregation of a few
characters occurred.

Cytological investigation showed that during maturation of the
original hybrids there was practically no pairing of the chromosomes;
one or two did mate, but at division they segregated, while the remainder
divided equationally, giving rise to apparently normal spermatocytes.

In the gonads of a hybrid formed by mating a male of the original
hybrids back to a pure anachoreta female, there was a triple set of chromo-
somes composed of a double set of anachoreta and a single set of curtula.
Pairing of the anachoreta chromosomes occurred at maturation, while the
curtula ones divided equationally. The gametes produced, therefore, had
a haploid set of anachoreta and curtula chromosomes.

The fertility of these hybrids may then be due to the almost entire
lack of pairing of the chromosomes in the first hybrids produced.

Reviewing the cases enumerated above it is apparent that sterility
is primarily due to the incompatibility of the maternal and paternal
germ plasms; this is sometimes manifested by the inability of the chromo-
somes to pair and separate in a normal way during synapsis, while in
other cases the repellent nature of the germ plasms seems to be so great
that pairing occurs only to a slight degree. In other hybrids, however,
synapsis takes place, but bivalent chromosomes are not formed, the
chromatin threads breaking into irregular masses.

It is evident, however, that whatever may be the cause of the
incompatibility of the chromosomes, it is not sufficient to prevent the
functioning of the rest of the cells of the body, which appear to perform
their several duties in a normal manner. Further we saw that in the
Pygaera hybrids, which were quite fertile, there was little or no pairing

D. W. Cutler 163

of the chromosomes. Thus it seems that although the failure of the
chromosomes to pair and separate is not the direct cause of sterility, yet
it is often the manifestation of the antagonistic nature of the constituents
of the germ plasm.

Up to this I have dealt entirely with sterility of hybrids: there is, how-
ever, a second type where sterility is found in a pure species of animal.
Two such examples were investigated by Doncaster and myself. The first
was that of a sterile tortoiseshell tom cat. These animals are exceedingly
rare and, when produced, are usually sterile, or fertile to only a slight
degree. The colour is normally found only in the female and can be pro-
duced by mating a yellow male with a black female. The male offspring
are then black. The yellovv colour is sex-limited in inheritance and goes
only with female producing gametes. On occasion, however, this yellow
factor goes with male producing gametes and tortoiseshell toms result.
Such a cat we had in our possession and though repeatedly mated no
offspring were developed. Investigation of the testes demonstrated that
the sterility was caused by the complete failure of the spermatogonia to
form. Sterility produced in this way was also found in a hen-feathered
cock, extracted from the cross Sebright hen with Hamburgh cock. Hen-
feathering is undoubtedly an inherited character, so that these two cases
suggest that, when a factor normally confined to the female, is received by
the male animal, there is a tendency for sterility to result. Though there
appears at present to be no connection between these cases and those of
sterile hybrids it is possible that there is a common basis to both. That
many of the factors which behave in a Mendelian- way are chemical ones
seems to be extremely probable. If then a factor common to the constitu-
tion of the female is on rare occasions transmitted to the male, the general
metabolism of the sex cells may be so upset that they are unable to
perform their usual functions. In the same way in hybrids the com-
position of the germ cells of the parents may be so divergent, that when
the two sets of chromosomes are brought together in the nuclei of the
gonads, they are unable to co-operate during the complicated phases of
maturation but are still quite capable of performing the activities
characteristic of the nuclei of the various somatic cells, and also of some
germ cells'solongas they are not brought into intimate contact with one
another.

It is by no means uncommon to find that the sex ratio of hybrids
is unequal, as was discovered by Guyer and Geoffrey Smith in hybrid
pigeons, by Mrs Haig Thomas in pheasant hybrids, and by Mr Lewis
Jones in the hybrids which I have been able to investigate.

164 Hybrids between Pheasant and Fowl

In the Bistort hybrids Harrison and Doncaster found that reciprocal
crosses gave very different results ; thus from hirtaria female and zonaria
male both sexes were formed, but from the converse cross only males
developed. An explanation of this was offered which assumed that in
the zonaria egg a certain amount of male determining substance must
be present if fertilisation was to produce a male. The hirtaria, sperm,
however, is supposed to possess a much greater amount of this substance
than does the zonaria sperm, and thus from any zonaria egg which is
fertilised by a hirtaria sperm, male offspring will develop. This assump-
tion is supported by the fact that the hirtaria chromosomes are four
times as large as the zonaria ones, and if the male determining substance
is carried by a sex chromosome it will contain four times as much as
does any single zonaria sex chromosome.

It is possible that the high male production in the hybrids I have
investigated is due to a similar cause, and that the pheasant sperm is
preponderately male producing when it fertilises the egg of the fowl.
No definite evidence of this has as yet been obtained, but I hope to
conduct research on these lines in the future.

I wish to express my thanks to Mrs Haig .Thomas and Mr Lewis
Jones for supplying me with material, and especially am I grateful to
Dr Doncaster, who has read over this manuscript and offered me many
valuable suggestions.

EXPLANATION OF PLATE IX.

All the figures were drawn to the same scale. The outlines were drawn by means of a
Zeiss-Greil drawing apparatus, and the details filled in free hand.

Figures 1 — 10. Spermatogenesis of male pheasant.

Fig. 1. Equatorial plates of three spermatogonial cells.

Fig. '2. Young primary spermatocyte.

Fig. 3. Early synapsis stage : fine threads are withdrawn to one side of the nucleus.

Fig. 4. Later synapsis stage with thick coiled threads.

Fig. 5. Primary spermatocyte, equatorial plate probably with 10 chromosomes.

Fig. 6. Primary spermatocyte, equatorial plate probably with 11 chromosomes, one

projecting beyond the others.
Figs. 7, 8. Division of primary spermatocytes : in Fig. 8 one of the chromosomes

is differentiated from the others.
Figs. 9, 10. Secondary spermatocytes, equatorial plates.

JOURNAL OF GENETICS, VOL VII. NO. 3

PLATE IX

•//!

24

10

?»

D. W. CUTLKR 165

Figures 11 — 27. Spermatogenesis of hybrids.

Fig. 11. Spermatogonia! division of dark coloured hybrids.

Figs. 12, 13. Spermatogonia! divisions of !ight coloured hybrids.

Figs. 14, 14 a. Spermatogonia! equatorial plates of darli coloured hybrids showing in
Fig. 14 probably 20 chromosomes, and in Fig. 14a 19 or 20.

Fig. 15. Spermatogonia! equatorial plates of light coloured hybrids, showing pro-
bably 19 chromosomes on the left, and 20 on the right hand figure.

Fig. 16. Spireme stage of darlc coloured hybrids.

Figs. 17, 18, 19. Stages in synapsis of dark coloured hybrids, showing the concen-
tration of the threads to one side of the nucleus.

Pig. 20. Three large chromatin masses in the primary spermatocytes of light coloured
hybrids.

Figs. 21, 22, 23. Irregular number of chromatin masses in the primary spermato-
cytes of light coloured hybrids.

Figs. 24 — 27. Irregular number of chromatin masses in the primary spermatocytes
of dark coloured hybrids.

REFERENCES.

Cutler, D. W. and Doncaster, L. " On the Sterility of the Tortoise-shell Tom

Cat." Joiirn. of Genetics^ Vol. v. 1915.
Federley, H. "Das Verhalten der Chromosomen bei der Speruaatogenese von

Pygaera." Zeitschr. f. indukt. Abstamm., Vol. ix. 1913.
Guyer, M. F. " Spermatogenesis of Normal and Hybrid Pigeons." Thesis, University

of Chicago, 1900.

"Sex of Hybrid Birds." Biol. Bull, Vol. xvi. 1909.

" Studies on the Chromosomes of the Common Fowl as seen in Testes and

in Embryos." Biol. Bull., Vol. xxxi. 1916.

Harrison, J. W. H. and Doncaster, L. " On Hybrids l>etween Moths of the

Geometrid sub-family Bistoninae, etc." Journ. of Genetics, Vol. lir. 1914.
Smith, G. " Studies in the Experimental Analysis of Sex, Part 9. Spermatogenesis

and the Formation of Giant Spermatozoa in Hybrid Pigeons." Quart. Journ.

Micr. Sci., Vol. LVlii. 1912.
Smith, G. and Thomas, R. Haig. "Sterile and Hybrid Pheasants." Journ. of

Genetics, Vol. ili. 1913-14.
Thomas, R. Haig. " Experimental Pheasant-breeding." Proc. Zool. Soc., Vol. iii.

1912.
Wodsedalek, J. E. "Causes of Sterility in the Mule." Biol. Bull., Vol.xxx. 1916.

A FUETHEE NOTE ON THE GENETICS
OF FRAG ARIA.

By C. W. RICHARDSON.

My work on Fragaria has been somewhat intermittent in character
for the last three years owing to other calls on my time, accordingly
many experiments I hoped to continue, when I wrote my first " note "
to this Journal (Vol. ill. No. 3, Feb. 1914), I have been obliged to leave
on one side for a future date ; also a considerable amount of seed gathered
from plants in 1914 has either failed to germinate or germinated badly.
But there is an ever growing bundle of facts which time might tend to
render stale or to submerge under fresh detail, and it is this accumu-
lation I would add to the record.

Once again I express my gratitude to the John Innes Horticultural
Institution for the facilities they have afforded me in carrying on my
work, at a time when labour is very short and there are few to answer
the numerous calls of a garden " in being."

Flower Colour. It is difficult to distinguish between very light
pinks and pure whites, so much so that I have found it impossible to
draw a clear line between them after 5 p.m. in June, when, as a rule, in
England the light tends to become red.

As stated, in my previous paper, the cross pink flowering vesca x white
tiowering vesca produced pink flowering F^h. These selfed produced
20 Pink, 57 Pale pink, 10 White or very nearly white (of the latter at
least 3 were absolutely white).

Expectation 1 to 15 — 8157 Pink to 5-43 White.

Double jioivering. My original double vesca parents were hardly so
consistent in their double quality as one might desire, but they were
quite as double as any I have seen elsewhere or obtained in my F.,
generations, the leading flowers in a truss are generally the most double
and are frequently perfect.

168 Further' Note on Fragaria

The cross double flowering vesca x single produced in the Fi gene-
ration single flowers with occasional extra petals. The Fi's selfed pro-
duced

60 Single 24 Double

37 Single 13 Double

Total 97 Single 37 Double.

Expectation (3 to 1) lOO'S Single 33"5 Double.
The cross Single x Double produced /^/s single.
The Fi generation selfed produced

58 Single 25 Double.

Expectation (3 to 1) 62'25 Single 20-75 Double.

It is worth noting that some dark pink doubles were very beautiful
flowers : my experiment was not on a large enough scale to make certain
of my figures but they pointed to a 1 to 15 ratio, the white double being
the 1.

Hairy leaf stems and front of leaf . In reality all, stems are to some
extent hairy, but some appear to be " hairy " because the hairs stand
out from the stems and are more numerous. On the other hand, the
front of the leaf may be quite glabrous.

In a cross Virgihiana x Chiloensis the F-^h had " hairy " stems and
the front of the Ifeaves was hairy. The i^i's selfed produced

32 " Hairy " stems 10 not " hairy "

29-1-3 (marked ? " consider hairy ") Hairy front of leaf 10 Glabrous.

Clearly a 3 to 1 in each case.

An F., generation, from a selfed F^ of a cross Chiloensis x Virginiana,
have yielded plants showing marked segregation of numerous characters
— too numerous for me to go into carefully in the short time at my
disposal.

Leaving these simple matters I pass to the much more difficult
problem — Sex.

Chine nsis $ x Chiloensis ^ gave

26 $ 21 c/ or ? .

Chinensis $ x Virginiana </ gave

18 ? U^ or ^.

0. W. Richardson 169

Virginiima % x Mexicana </ gave

20 % 15 d" or ? .

Virginiana % x Virginiana ,^ gave

17 $ 16 c/ or ? .

The total of my recorded figures now stands
183 % 155 cT or 5 .

The noteworthy fact is the persistent dominance of the females which
seems to fit a 9 to 7 ratio. The chief difficulty in this line of work is
classification — and I fear to a large extent this must remain a matter
of individual preference. At present I find it simpler to group all sterile
flowers with the sex to which they appear to belong, and to consider
males and hermaphrodites as one sex, rather than two, as it is almost
impossible to say that a given apparently male plant (e.g. a Virginiana)
will not set an occasional well-developed fruit or an occasional seed.

Up to the present I have no recorded case of a female flowering plant
becoming male or hermaphrodite — once a female always a female.

I have failed to produce by crossing species any fruit markedly larger
than either of its original parents, but I have not received the results of
Hautbois crosses.

James Barnet in a description of the plants in the Society's garden
{Tr. Hurt. Soc. vi.) makes special mention of " Hudson Bay" and other
Canadian plants as giving large fruit — it is possible the key may lie in
these fruits slowly developed in the long light of the far North summer ;
when opportunity for continuous work comes once more I hope to make
use of some Canadian species.

From a cross Virginiana x vesca I have now 5 F^ plants from selfed
F^ hermaphrodite plants, they are still too young to draw conclusions
from except that they have the leaf-colour of Virginiana and only one
has normal leaves, the remainder having five or six malformed. It is
not surprising that this cross has been generally considered sterile. Out
of some 200 plants fiowering freely in the open I found 4 females which
set one or two seeds on each plant and 3 hermaphrodites which behaved
in the same manner. Last year from 4 large free flowering runners,
obtained from the most fertile of these 3 hermaphrodites, I obtained 13
seeds five of which were fertile. I counted 260 flowers on one of these
plants ! No flower set more than two seeds, and any one might have set
from 80—150.

From 12 chance seeds gathered from F^'s in the open, with garden
varieties in rows next to them, I obtained four plants, which produced

170 Further Note on Fragaria

malformed hermaphrodite or male flowers of no set shape and up to the
present sterile.

From crosses made with Vesca x Daltoniana, garden varieties and
Chiloensis, I have never obtained free fruiting plants, and am now very
strongly inclined to believe that Vesca has nothing to do with the present
garden varieties.

From my own experience in crossing Fragaria, I may perhaps be
allowed to add the following note which may be useful to others engaged
on similar work.

In gathering strawberries for seed it is essential to gather them when
the seed is quite ripe. I find the best plan is — gather dry, i.e. not wet,
ripe fruit without handling it — place in a strong paper seed packet —
gum up — and keep in a fairly dry room. The fruit becomes nearly
desiccated, most of the liquid passing into the air and some into the
paper. There may be some mildew produced, but it does not matter.
Seeds when wanted can be rubbed from the dried fruit. I have found
most seed fertile after three years, but, when the fruit has been very
small, and the seed, through lack of pulp, become very dry, the fertility
is largely lost.

There is nothing to be gained by sowing in July-October, but some
gain is made by sowing during November in a hot-house. When seed
is valuable or hard to obtain it is an excellent plan to place it between
two thin folds of wet cotton wool and keep in a warm house (about 70° F.).

On the least sign of germination the seeds should be placed in pans
in a cool house, where they can remain till planted out or potted up ;
the first seeds generally germinate within 12 days, the last may take
months. Wood lice take kindly to very young strawberry seedlings, so
it is wise to place pans or boxes on glass jars until plants develop two ■
leaves.

Vesca or vesca-like plants or crosses should be renewed by runners
or subdivision, as they wear out after two years. On the whole the only
real disadvantage the strawberry presents as a subject for study is that
it requires a considerable amount of space all the year round, and it is
extremely doubtful policy to grow catch crops between the rows of
plants.

GYNANDRY IN ARACHNIDA.
By the Rev. J. E. HULL.

(With one Text-figure.)

i. I WRITE under the general heading of Arachnida though actual
cases of gynandry are known in one order only — Araneae. In the con-
cluding section of this paper, however, I shall have something to say
concerning the other orders : meanwhile I proceed to discuss the spiders.

It may be well in the first place to review the general sexual
characters, beginning with the external structures of the genitalia. These
lie on the median line of the epigaster — the anterior segment of the
venter — flanked on either hand by the anterior spiracles. In the female
this sexual area is a more or less specialised epigynium, sometimes simple,
sometimes elaborately sculptured (affording excellent specific characters),
in or under which the vulvar apertures are situated. In the male there
is no special epiandrium ; the paired apertures open on the epigastral
margin, and are practically invisible under ordinary powers of magnifi-
cation. There is no penis ; the copulatory apparatus is a special
modification of the terminal article of the palp.

The outstanding secondary sexual characters are the following :

1. Size. The female is almost invariably larger than the male,
sometimes considerably larger : but none of the . known instances of
gynandry have occurred where the difference in size is unusually great ;
all the records are of species in which the total length of the female does
not exceed that of the male by more than 15 per cent, of the latter. The
difference is always sufficient to cause asymmetry of the body in the
gynandromorph, though in some cases it is not very conspicuous.

2. Palpi. As the tarsus in the male is expanded and hollowed out
beneath to accommodate the highly specialised copulatory apparatus, it
becomes an organ of the greatest importance, affording, like the epigyne
of the female, excellent specific characters. The ' genital bulb,' as it is

Journ. of Gen. vii 12

172 Gynandry in Arachnida

generally called, may briefly be described as a more or less elaborated
syringe, capable of sucking up the seminal fluid ^ and expelling it again
gradually in the act of coition. The corresponding article of the female
is a plain cylindrical joint more or less acuminate at the apex.

3. Other characters. So far as the species enumerated below
are concerned these are all associated with the cephalic region, including
the pair of prae-oral appendages (falces or mandibles). Up to the time
of the penultimate moult these differences do not show (nor any other,
as a matter of fact) ; but as the final moult approaches, the tarsus of the
male palp enlarges rapidly, and in the case of the Linyphiidae (to which
family belong nearly all species now to be dealt with) there is often a
special development of the male caput, and of the form and armature
of the falces.

ii. Gynandromorphs of three species have been figured and de-
scribed :

Oedothorax fuscus Bl. {sub Erigone fusca). Kulczynski, PoiworeA;

Obojnakowy Pajaka, Cracow, 1885.
Maso sundevallii Westr. Falconer, Naturalist, 1910, p. 229.
Lophomma herbigradum Bl. Hull, Trans. Nat. Hist. Society of

Northumberland, etc.. Vol. iv. (New Series), p. 48.

No two of these agree, as it happens, in the distribution of sexuality,
and they may be regarded therefore as types of three different classes.
I take them in order, as above.

1. One side male, the other female — sexual structures perfect except
for the distortion resulting from the union of dissimilar halves on the
median line.

Of this Kulczynski's Oedothorax fuscus is an excellent example.
I translate his description, and add some of his figures.

The right half of the cephalothorax is longer and wider than the left (width:
380 /i, left 350/x; length from fore middle eyes to the hind margin: right 970 /x,
left 910;:i), the difierence mainly accounted for by the asymmetry of the hind margin,
of which the shape is shown in Fig. 1.

The eye area is slightly asymmetrical, the right eyes being a little in advance of
the left : what difference there is in the size of the corresponding eyes is hardly per-

1 The peculiar interest of this operation will excuse a note. I have twice witnessed
the exclusion of the seminal fluid. In each instance it was deposited on a leaf — the living
leaf of a tree in the case of Linyphia montana, a dead leaf on the ground in the case of
Lycosa amentata ; in both cases it was immediately taken up into the bulb. Coition was
effected a few minutes afterwards in the former case ; in the latter I could only watch
about half-an-hour, and courtship was still proceeding when I left.

J. E. Hull

173

5 f*"

1

lb

^y>

1. Oedot/iorax/«.«c«« Bl. 6 . After Kulczynski.

2. ,, ,, ,, a— epigyne of normal $ ; & — genital area (after
Kulczynski).

3. Maso simdevallii Vfestr. a — ^ epigyne; b — ^ genital area (after Falconer).

4. Lophomma herbigradum Bl. o .

5. ,, ,, ,, a — 9 Epigyne; b — o genital area.

6. ,, ,, ,, g sternum.

7. ,, „ „ 6 frons and falces.

12—2

174 Gyiiandry in Arachnida

ceptible. Looked at from the front the cephalothorax seems a little distorted,
because the front row of eyes is more strongly curved on the left than on the right,
and the lateral slope of the caput is steeper.

The mandibles are about equal in length (330 /i), but the right is rather broader
(190;i) than the left (170/i), and the setae are somewhat differently distributed. The
fore margin of the fang-groove bears on the right mandible 4, and on the left 5 teeth:
the former are slightly larger and unequally spaced ; the latter are equidistant, the
first and last being obviously smaller than the three between them. The sternum
is asymmetrical ; the left maxilla, seen from below, appears shorter than the right.
The left palp is a completely normal and fully developed male palp ; the right is
female without trace of abnormality.

It is remarkable that in spite of the marked inequality of the two body-halves
the corresponding legs have joint for joint exactly the same length : but there is an
obvious difference between the metatarsi, and more particularly between the tarsi
of the first pair. The latter differ both in shape and pubescence : the left is slightly
clavate, at the base 55 /i, near the apex 70 /x wide ; while the right is of uniform
thickness (55 /x) except for the very apex ; the pubescence of the male (left) tarsus is
much denser and finer than that of the female right....

The asymmetry of the abdomen (1270/x long, 880 /x broad) is pretty obvious, the
right half being broader and more strongly curved than the left : a line from the pedicle
bisecting the fore part of the abdomen leaves all the spinners on its left side. The
right inferior spinner is 138 /x long, the left 130 /x, the former 95 /i, the latter BO^x wide
at the base.

Very striking is the asymmetry of the genital area. The right half is more
strongly chitinized, the hind margin from the middle to the right dark, to the left
light ; the fissure on the left in which lies the entrance to the spermatheca, is abnor-
mally bent inwards, whereas in the normal female it runs straight forward ; on the
outer side of this fissure are two conspicuous spots — that in front oblong blackish,
reaching the fissure ; that behind it round reddish and at some distance from the
fissure. The latter spot is caused by the underlying spermatheca, the former by the
duct leading from it to the external aperture. Of all this the left half of the genital
area shews no trace.

This careful description simply means that (externally)^ the spider
was wholly and completely male on the left side and female on the right,
the structures being strictly normal except for junctional distortion. It
is interesting to note that the careful examination of this spider revealed
the sexual differences in tarsus I which had previously escaped notice.

2. As 1, hut one side imperfectly developed before, the other behind.
This is exemplified (according to my own reading of it !) by Falconer's
Maso sundevallii, — a British example captured by Dr A. Randell Jackson,
time and place unknown. But I have not seen the actual specimen, and
Falconer's opinion is different. I quote his brief description :

1 This qualification is to be understood throughout this paper. No dissections have
been made of any of the gynandromorphs here enumerated.

J. E. Hull 175

The left palpus is of the male form, the palpal organs being well developed ; the
right palpus is somewhat tumid with the appearance of being loosely covered at the
apex, and is thus not quite of the normal female shape. The epigyne is very im-
perfect, the parts on the left side being obsolete, and those on the right being very
much distorted. ...The specimen is thus male on the left side, but not quite female
on the right.

It should be remarked that asymmetry of the body is not to be
expected in this species as the two sexes are of approximately the same
size ; indeed sexual dimorphism is very feeble.

Not having seen the specimen, I can only judge the description of
the genital area — loosely called the 'epigyne' — by the figure given.
To me this figure (in the explanation of the plate — Naturalist, 1910,
p. 229 — described by error as ' epigyne of female ') presents much the
appearance one would expect from an amalgamation of the right half
■perfectly developed) of an epigynium with the left half (probably im-
perfect) of an epiandrium. The spermatheca on the right seems quite
normal — in fact nothing abnormal save the inevitable disturbance of the
median septum. If there be defect it seems to me to be on the left,
which in a normal epiandrium is wholly covered by the dark pubescent
cuticle of the epigaster, but here unmistakably exhibits vestiges of a
female element of about the same strength as the male element in the
right palp. If I interpret it rightly this particular gynandromorph
presents a reciprocal combination — a left side perfectly male in front,
imperfectly male behind ; and a right side imperfectly female in front
and perfectly female behind.

I am the more confirmed in this opinion because it seems to agree
with a Hilaira excisa Cb. taken by me near Haltwhistle, Northumber-
land, in August 1898. The specimen was accidentally destroyed and
never recorded. It was certainly male on the right as regards palp and
carapace, the tarsus of the left palp being crassate and the occipital tuber
(characteristic of the male) nearly bilaterally perfect ; so that the left
anterior region was, at most, predominantly female. I unfortunately
cannot speak with certainty of the genital area. There was asymmetry
both of cephalothorax and abdomen, but I have no accurate note of it.

3. One side perfectly female before and male behind, the other per-
fectly male in front and female behind.

This is the condition of things in my Lophomma herbigradum.
I reproduce my original note.

A fine gynandrous example of this common species was taken at Ninebanks in
the spring of 1910. For the most part this specimen exhibits the usual phenomena

176 Gynandry in Arachnida

of bisexuality ; that is, one side is male and the other female, with no atrophy or
distortion of parts except where mutual accommodation is necessary on the median
line. Naturally this disturbance of structure shows itself chiefly in the sexual region
of the epigaster. In the present instance the female side of the external genitalia
suffers less modification than the male side. In one particular, however, this indi-
vidual differs from all other bisexual spiders I have ever seen or heard of ; for while
the right side of the cephalothorax is male and the left female, in the case of the
abdomen the sexes are reversed — the right side being female and the left male.

In view of the unusual character of this specimen I now supply
further details.

The first two pairs of appendages of the cephalothorax, being uncon-
nected, retain their sexual characters undisturbed. Thus the right falx
(or mandible) is typically male — attenuate and divergent distally, with
four teeth on the forward border of the fang-groove, a proximal pair of
which the first is smallest of all and the second largest, and a distal pair
of intermediate size, the larger (the third) being slightly out of the line
of the rest on the side farthest from the fang-groove. The left falx is
typically female, equal in breadth to the right at the base, but wider at
the apex (i.e. not attenuate or divergent), with five fang-teeth, all in the
same line and equal, except the last, which is smaller than the rest.

Similarly the right palp is typically male. In this particular specimen
the copulatory organs were nearly fully exserted at the time of capture
and are so shown in the figure. They are fully developed in every detail
without the slightest variation from the normal. The same may be said
of the left palp which is completely and perfectly female.

As the cephalothorax differs in form and dimensions in the two sexes,
there is inevitably asymmetry of the carapace. In the first place the
clypeus of the male is rather higher than that of the female; conse-
quently there is in our gynandromorph a certain distortion of the ' face.'
The front row of eyes is straight but tilted upwards on the right (male)
side, on which side also the eyes are slightly nearer together and more
prominent. Another effect of the distortion is that the larger part of
the eye area falls to the left of the median line ; but it must be remem-
bered that part of this excess is normal, the eye area of a female being
slightly wider than that of a male.

In the group of genera to which the present species belongs it is
usual for the occipital area (including the two posterior middle eyes or
not) of the male to be raised into a tuber varying in form according to
the species, limited on either side by a furrow or indentation or pit. In
several species this occipital tuber is so slightly developed as to be
scarcely perceptible, though the lateral grooves (called the post-ocular

J. E. Hull 177

furrows) are generally well marked. This is the case in Lophomma
herbigradum ; and accordingly in our gynandromorph the right post-
ocular furrow is present, quite normal in form and dimensions. But on
the left side there is a corresponding longitudinal impression, very slight
but still obvious, which of course does not exist in a normal female. This
I take to be a merely mechanical effect, and not due to a subordinate
male element on the left side.

The normal female exceeds the male more in breadth of cephalothorax
and abdomen than in length, so that there is less displacement of the axis
of the body than in Kulczyiiski's Oedothorax (where the difference of the
sexes in length is considerable); but the difiference in breadth makes
itself visible in carapace, sternum, and abdomen. In the last, the sexes
being reversed, the lateral gibbosity is of course on the right side. It
will suffice to give illustrative dimensions of the sternum. The greatest
width of the left half is 263 /i, of the right 236 /i.

The legs present no special feature, but I append measurements of
the five distal articles — tarsus, metatarsus, tibia, patella, femur — in the
order named. As it is the proportiatis that matter, I take no definite
unit of measurement but use the length of metatarsus I (= 100) as a
standard of reference :

First pair of legs

j Right :
1 Left :

95
90

100
106

133
132

50
49

153

158

Second pair of legs ...

j Right :
1 Left :

85
85

95
95

112
110

49

48

135
140

Third pair of legs . . .

j Right :
1 Left :

75
73

80
80

92

85

45
44

120
115

Fourth pair of legs ...

J Right ;

87

105

140

45

155

1 Left :

85

105

135

45

150

To these three I must add a fourth of more dubious nature — an
example of Oedothorax retusus Westr. taken by Mr W. Falconer on the
sandhills near Southport in May 1904, and included by him in the paper
on Ahnormality in Spiders already quoted above (see Naturalist, 1910,
p. 199). He describes it thus :

The cephalothorax is raised behind the eyes as in a normal male, but the eleva-
tion is much less lofty, not so massive and totally devoid of the lateral impressions
which are [noty so conspicuous in the latter, while the descent to the ocular area is
also less abrupt. Both palpi are of the male form, but some of their parts, including
the palpal organs, are abnormal in shape, size, and structure. In a normal example
the tibial joint is shorter than the patella, and provided at the extremity with an

1 The brackets are mine. The 'not' is obviously a printer's error.

178

Gynandry in Arachnida

angular prominence directed outwards, and ending in a small pointed apophysis, a
little distance from which is a small, black, sharp-pointed, slightly curved spine,
directed downwards. In the abnormal specimen both palpi are without the angular
prominence, possessing only the curved spine ; the right [tibia] ^ is equal in length
to the patella ; the left [tibia] i is in normal proportion, but towards the extremity
has an irregular false articulation.. ..The epigyne is imperfectly formed but all the
details may be distinctly traced.

It will be observed that the sexual development is everywhere im-
perfect, and that the two halves are sexually dissimilar. Both palps
superficially resemble a normal male palp, but neither is fully developed.
The right palp, for instance, is farther from the normal than the left, as
is indicated by the form of the tibial joint, which makes an approach to
the dimensions of the normal female palpail tibia. If the explanation of
the arrest of male development be the presence of a female element, it
is stronger on the right than on the left.

Similarly the genital area has a general resemblance to that of the
normal female (so much so that Falconer as above quoted calls it an
epigyne), slightly imperfect on the right and still less perfect on the left.
If a male element is the disturbing cause, it is stronger on the left than
on the right.

This state of things may be represented diagrammatically thus :

Anterior

Left

M.F.

M.Fo

F2M2 j FsMi

Right

Posterior

Here M and F represent the male and female element respectively,
the subscript figures varying degrees of influence, of which the figures
may or may not be a measure.

iii. The following British records stand in the names of the Rev. O.
Pickard Cambridge and Dr A. Randell Jackson.

Hilaira excisa Cb. Cambridge : Proc. Dorset, etc., Field Club, 1902,
p. 21.

Adults of both sexes found near Glamorgan 2 and sent to me by Dr A. R. Jackson
in 1901. Among them was a remai-kable bisexual form. One of the palpi was that
of the male spider, the other that of the female ; the form of the caput was that of

1 'Palpus' in the original ; obviously a slip of the pen.

'■^ I quote verbatim. The specimen was actually taken on Maendy Hill, near Ystrad in
the Rhondda valley.

J. E. Hull 179

the male, and the abdomen was of the male form. I have seen a somewhat similar
form in an exotic spider, but never before among the many thousands of British
spiders I have had occasion to examine.

In all probability this specimen will still be preserved in the Cam-
bridge collection, now (or presently to be) in the Hope Museum at
Oxford.

Dr Jackson mentions this spider in a list of the spiders of Glamorgan
{Cardiff Nat. Soc. Trans., Vol. xxxix. 1907). Like Cambridge he says
the caput is the caput of a male, but he does not mention the abdomen.

Poprhomma pallidum Jacks. Jackson: Trans. Nat. Hist. Soc.
Northd. etc.. Vol. I. (new series), part iii, p. 384 {sub P. oblongum Bl.).

A gynandrous specimen taken at Hexham. It has one male palpus, one female
palpus, and a distorted epigyne.

Probably this specimen is still in existence, but Dr Jackson is on
active service and his collection for the time being inaccessible : so it is
impossible to say whether this gynandromorph falls into Class 1 or 3.

Leptyphantes pallidus Cb. Jackson: Trans. Nat Hist. Soc.
Northd. etc., vol. ill (new series), part ii, p. 435.

A fine gynandrous form occurred at Cudham at the end of May (1908). In this
specimen the right palpus was of the male form with well-developed palpal organs.
The left palpus was of the female type. The epigyne was large but asymmetrical.
The central portion was of the normal female type, and so was the left part of the
scapus. The right portion of the scapus was quite short. Thus the specimen was
male on the right side and female on the left.

Agroeca ppoxima Cb. Cambridge: Proc. Dorset, etc.. Field Club,
1913, p. 112.

No figure or description is given — merely the record of the capture
(in Dorset) of a ' hermaphrodite ' of this species. It is, however, im-
portant as being the only definite record of a gynandrous spider which
does not belong to the family of the Linyphiidae.

iv. I believe Dr Thorell somewhere casually records the occurrence
of a ' hermaphrodite ' of the family Lycosidae {Lycosa sp.), but I cannot
trace the reference. The authentic cases here included may be classified
thus:

Linyphiidae :

§ LInyphiina — 2 species.

Leptyphantes pallidus Cb.

Porrhomma pallidum Jacks.
§ Hilairina — 1 species (twice).

Hilaira excisa Cb.

] 80 Gynandry in Arachnida

§ Copyphaeina — 2 species.
Oedothorax fuscus Bl.
Oedothorax retusus Westr.

§ Nerienina — 2 species.

Maso sundevallii Westr.
Lophomma herbigradum Bl.

Drassidae :

§ Clubionina — 1 species.

Agroeca proxima Cb.

Kulczynski's Oedothorax was taken in Galicia in 1880 : the rest are
British (England, 7 : Wales, 1) ; and of these eight species seven belong
to one family, Linyphiidae.

It is perhaps important to have a just conception of some reasons
why the Linyphiidae should take so large a place in this list. The
family is only one of seventeen represented in Britain ; but it is by far
the largest. I have just completed a revision of the British list of
spiders, and make the total of British species 539. Of these the Liny-
phiidae claim no less than 232 : proxime accessit, the Drassid family,
58 — exactly one-fourth of the Linyphiid total. Thus, of every 5 British
species, 2 are Linyphiids. In the north of England, which supplies 5 of
our 8 gynandromorphs, the proportion of Linyphiids is still higher; in
Northumberland quite a half Moreover, the family includes nearly all
the very small and critical species, which the collector finds it necessary
to take for closer examination. I shall not be over the mark if I say
that in an ordinary way 80 per cent, of my own captures (in Northumber-
land) are Linyphiidae.

Still the difference in size of the sexes is so general that the conse-
quent asymmetry of the body, together with the conspicuous difference
of the palps (to say nothing of every collector's keen interest in practi-
cally every adult male !) would make it difficult for a gynandromorph to
be overlooked whatever family it belonged to.

On the whole therefore we may conclude that the preponderance of
Linyphiidae in the records of gynandry fairly represents the actual state
of things. And, briefly, the British figures stand thus — in 232 species
of the Linyphiidae we have 7 cases of gynandry ; in 377 species of other
families 1 only. Taking the figures as they stand, they indicate that
cases of gynandry are a little more than nine times as frequent among
the Ijinyphiidae than in all the rest taken together.

J. E. Hull 181

Whatever the reason may be, it is obviously indisputable, after
making every allowance for the imperfection of the numerical test, that
the liability to gynandry (in Britain) is strikingly greater in the Linyphiid
family than in any other. But I do not think that the tendency
can be ascribed to any particular alliance within that family, for the
two great branches into which it is divided — the Linyphiine and the
Erigonine — are about equally represented among the known gynandro-
morphs.

V. Examples of gynandry in other orders of the Arachnida are yet
to seek. In recent years I have had through my hands many thousands
of specimens, chiefly Acari (particularly Gamasidae, Thrombidiidae,
Oribatidae, and Tyroglyphidae) and have never seen a true gynandro-
morph; nor do I know of any record of one. I may, however, here refer
to an observation of Canestrini (Prospetto delV Acarofauna Italiana, in,
p. 364) on a Tyroglyphid — Rhizoglypkus echinopus F. and R. — which has
some bearing on the subject. In maintaining that Hypopus dujardinii
of Claparede and Rhizoglypkus rohinii of Michael are dimorphic males
of the same species, he says : " My conviction is strengthened still
more by the discovery of a male in which one leg of the third pair is
incrassate as in Rhizoglypkus rohinii, while the other is of ordinary
dimensions " (which is the case with both in the female and in the other
form of male, i.e. Hypopus dujardinii Clap.). It would be interesting
to know something of the genitalia of this abnormal individual ; for the
normal third leg might be either male or female.

But the total lack of records of gynandry in orders other than
Araneae is by no means surprising, for (the Gamasidae and a few other
Acari excepted) sexual dimorphism is so slight that the sexes cannot be
separated without very close examination. Indeed, in the case of the
Oribatidae, it is impossible to distinguish male from female by an
external examination, however close. v

NOTES ON THE GENETICS OF TEUCRIUM
SCORODONIA CRISPUM (STANSFIELD).

By M. C. RAYNER, D.Sc.

(With Plate X.)

An interesting variety of the common " Wood-sage " came into my
possession some years ago and has since been crossbred with typical
plants of the species in order to investigate the genetic behaviour of the
varietal leaf characters.

The results of these experiments are recorded in the present note
which is preliminary to a more detailed account of the structural
peculiarities and behaviour of the plant.

The type species, Teucrium Scorodonia L., the " Wood-sage " or " Ger-
mander," is a common plant, especially of dry open woods, commons and
heaths, locally abundant but often absent from certain areas. The
leaves are variable as to size, details of shape and incision ; two charac-
teristic examples are figured in Plate X, fig. 1. A variety, Teucrium
Scorodonia dentatum Bab., with deeply cut leaves, is recorded but does
not appear to be common.

The variety under discussion, recorded as Teucrium Scorodonia cris-
pum (Stansfield) is of rather more compact habit than the species, the
leaves are broader and blunter and the leaf margins are characteristically
" crisped " or " crested " as shown in the photograph (Plate X, fig. 2).

The inflorescences, flowers and fruits resemble those of T. Scorodonia.

The variety is very distinct and is of some interest for taxonomic
reasons, inasmuch as there appears to be no previous record of such a
form of Teucrium in this country, nor have I noticed any tendency
towards leaf variations of this kind in wild plants in localities where the
plant is abundant. The deviation from type is quite clean and well-
marked in the variety and I have not been able to find a reference to a

184 Notes on Teucrium Scorodonia crispum

" crested " variety of Teucrium Scorodonia as a local form, or to the
species as one in which the abnormality of leaf-structure known as
" cresting " has been observed.

For the original material, I am indebted to Dr F. Stansfield, in
whose garden the plant has been cultivated for many years.

Dr Stansfield's plants are vegetative descendants of a wild plant
found in Devonshire at least 50 years ago, — exact date and locality
unknown. The plants are readily increased by division and by cuttings
and have remained true to type during this long period of vegetative
propagation.

There is no tendency for the leaves to revert to the type normal for
the species, such as is exhibited, for instance, in similar crested leaf-
varieties of Anemone japonica and of Viola sp., in both of which
corresponding forms with crisped leaves are known but are not per-
manent under cultivation. The flowers of these crested Teucrium
plants are normal, viable seed is formed and the plant sows itself freely
under cultivation.

Dr Stansfield has observed many generations of such self-sown seed-
lings in his garden, and they are invariably of the normal type, showing
no trace of the " crested " character.

The possibility of accidental crossing appeared to be excluded, since
the nearest wild plants of 2\ Scorodonia are at least two miles distant.
The correctness of this view is confirmed by the results of the experi-
ments recorded below.

In 1913 flowers of a "crested" plant were pollinated from wild
plants of T. Scorodonia and also from plants growing in the Cambridge
Botanic Gardens.

These crosses gave about 20 seeds, the majority of which germinated,
yielding an Fi generation with normal leaves (PI. X, fig. 3).

These Fi plants have been under cultivation since that time and
have never given the slightest indication of their hybrid origin.

The cross has not been made in the reverse direction.

In 1915 several of the F^ plants were selfed and were also crossed
with the " crested " grandparent, using the latter as pollen parent.

These crosses yielded an F^ generation of approximately 200 seeds
in the first case and of 12 seeds in the second case. From the 200
seeds 89 seedlings were raised, all of which grew to maturity and have
continued to grow without showing any trace of the " crested " habit
(PL X, fig. 4).

None of the seeds obtained from the other cross germinated.

M. C. Rayner 185

The experimental results may be tabulated briefly as follows :

Crested x Type

Fi , 20 seeds, of which the majority germinated giving
plants exactly like the pollen parent.

Fi selfed. -Fi crested.

I i

F-i , 200 seeds, of which 89 reached 12 seeds, none of which germinated,
maturity giving plants exactly
like parents : no trace of ' crest-
ing.'

The experiments are obviously incomplete and must be repeated
and extended before exact hypothesis can be founded upon them.

The mortality among the Fq generation seedlings may be significant
or may be accidental : it requires investigation and analysis and there
is at present no evidence as to its cause.

It is conceivable, for instance, that the seeds carrying the " crested "
character are not viable or die off soon after germination. This would
account for the non-appearance of " crested " plants under experimental
conditions and also for their absence in a wild state.

Such a hypothesis, however, offers no explanation of the apparent
non-inheritance of the " crested " character wheij the variety is selfed.

If, on the other hand, the mortality was due to accidental causes and
the " crested " character is never inherited by seedlings, the possibility
that the original plant was of the nature of a periclinal chimaera
suggests itself as an explanation.

Microscopic examination of the epidermal tissues of the two parents
has not yielded any evidence in support of this view, nor is it easy to
imagine how a plant of this constitution could have arisen in the first
instance. The possibility is being experimentally tested.

The case seems to be worth recording at this stage if for no other
reason than to put on record the appearance of a well-marked and
apparently isolated case of leaf- variation in a wild plant.

A tendency to excessive marginal growth of the leaves resulting in
the " crested " or " crisped " habit is common among ferns and not
uncommon among Angiosperms and in the former it is often inherited
by the sporelings.

There is no record, as far as I have been able to ascertain, of such a
general tendency showing itself in Teucriiim Scorodonia, in which plant
it seems to have manifested itself as a sudden and rare variation.

University College,
Reading,

November 1917.

186 Notes on Teucrium Scorodonia crispum

DESCRIPTION OF PLATE X.

Fig. 1. Leaves from a wild plant of Teucrium Scorodonia L.

Fig. "2. Leaves from a plant of the "crisped" variety of T. Scorodonia L. recorded as

T. Scorodonia crispum (Stansfield).
Fig. 3. Leaves from plants of the Fi generation of the cross T. Scorodonia crispum x

T. Scorodonia.

Fig. 4. Representative leaves from plants of the Fg generation resulting from the selfing
of plants of Fig. 3.

JOURNAL OF GENETICS, VOL. VH. NO. 3

PLATE K

SOME EXPERIMENTS ON THE ROTIFER
HYDATINA.

By EDITH E. HODGKINSON.

Six years ago I began some tests and experiments on the rotifer
Hydatina which were brought to a close in the summer of 1913-
The principal object was to test for strains producing no arrhenoto-
kous females. Many tests were made but arrhenotokous females
occurred in all the families tested ; no pure theljrtokous strain was
found. The Rotifers were isolated in watch glasses during the tests
and the food used was a horse manure solution prepared in the same
way as that used by Whitney^ for his experiments on Hydatina. The
results of the testing of one family are given in the table below (p. 188).
The rotifer laid 42 eggs six of which hatched into arrhenotokous females.
The descendants of the thelytokous females from these eggs were tested
through a varying number of generations. The generations were rarely
completed. The table shows the number of generations tested from
each female and the number of thelytokous and arrhenotokous females
in these generations.

Out of the thelytokous females tested all gave arrhenotokous females
in the first generation except I, XXXIV and XXXIX and these gave
them subsequently.

Attempts were made to alter the relative proportion of arrhenotokous
and thelytokous females by using solutions of horse manure of different
ages and concentrations, but the results were not definite. Other
methods were tried to produce an increase in the number of arrheno-
tokous females but also with no definite results. Protozoa bred in
horse manure solution were removed from the solution and put in
spring water and these served as the liquid and food for the rotifers.
The protozoa were removed from the horse manure solution by centri-
fuging some of the solution in a tube. The protozoa sank to the

1 Science, xxxii, No. 819, 1910.
Journ. of Gen. vii 13

188 Experiments on the Rotifer Hydatina

bottom of the tube and as much liquid as possible was drawn off with
a pipette, then spring water was added. The process was repeated
several times until it could be assumed little or no horse manure
solution remained. Judging from Shull's^ results, it was thought this
method would give a very definite increase in arrhenotokous females,
as the horse manure liquid was practically removed ; but very few were
hatched. The rotifers were kept well supplied with fresh protozoa
during this experiment. Again, the temperature was varied, the horse
manure solution was oxygenated, but no definite change in the pro-
portion of the sexes' could be produced.

No. of gene
rations
tested

T
females

A
females

No. of gene-
rations
tested

T
females

A
females

I

T

5

39

16

XXII

T

2

33

14

II

T

4

60

9

XXIII

T

4

29

8

III

T

5

70

13

XXIV

T

3

20

5

IV

T

4

45

8

XXV

A

—

—

—

V

T

5

48

11

XXVI

T

2

16

9

VI

T

5

42

9

XXVII

T

3

7

5

VII

A

—

—

—

XXVIII

T

3

20

8

VIII

A

—

—

—

XXIX

T

2

27

7

IX

T

3

46

13

XXX

T

15

134

9

X

T

5

56

7

XXXI

T

2

19

5

XI

T

4

32

11

XXXII

T

4

20

3

XII

A

—

—

—

XXXIII

T

4

21

3

XIII

T

5

50

16

XXXIV

T

17

115

5

XIV

T

2

24

7

XXXV

T

14

104

17

XV

T

5

51

9

XXXVI

T

4

18

12

XVI

T

5

34

14

XXXVII

T

4

23

7

XVII

T

5

26

16

XXXVII]

[ T

2

8

2

XVIII

A

—

—

—

XXXIX

T

12

93

11

XIX

T

5

28

14

XL

T

2

19

5

XX

T

4

31

11

XLI

T

3

23

7

XXI

A

—

—

_

XLII

T

2

11

6

During one period of the testing of the family given in detail above
no arrhenotokous females were produced. A very concentrated solution
of horse manure was used but it was not an old solution. This was
suspected of producing the results and on the completion of the test
rotifers were bred in a highly concentrated solution of horse manure
which was swarming with protozoa. The solution was more concen-
trated and the protozoa more numerous than in any test previously
made. No arrhenotokous females were produced by these rotifers.

I then decided to try the effect of a culture of green protozoa used
as food instead of the colourless protozoa in the horse manure solution

1 Journ. Exp. Zool. 1911.

Edith E. HooaKiNSON 189

and subsequently learned that in his paper of 1910 Whitney stated that
an apparently thelytokous strain could be made to produce arrhenotokous
$ $ when fed upon the green flagellate, Chlamydomonas. I found that
rich cultures of green protozoa could be made by the following method.
Bones and meat are covered with water and allowed to stand for a few
days. A little of the solution is put in a glass jar, filled up with
spring water and infected with green protozoa — Euglena, etc. This is
placed in sunlight and in 3 to 5 days a rich culture is formed and is
ready for use. By this means cultures of different ages and in different
conditions can be used, if required.

Experiment I. A thelytokous female, a descendant of XXX in the
test given above, was selected, as few arrhenotokous females had occurred
in this line, and put in the green protozoa culture. The details of the
descendants of XXX when in the horse manure solution are given

Generation

T female

, A female

1

8

1

2

3

0

*3

1

0

4

3

0

5

6

0

6

12

0

7

6

0

•8

15

0

9

5

0

10

10

0

11

17

2

12

10

4

13

26

2

14

8

0

15

5

0

Totals ... 185 9

* A break occurred here in the continuity of the generations as generations 3 to 7 were
bred away from the laboratory.

These are the details of the descendants of the thelytokous female
in the green protozoa culture.

T A

AAAAAAAAAAATTTAAAATTAATTTTATAAAATTAAAAA 12 27
ATTTATATTTAAAAAATAAATATA 10 14

i 1

A A AAA

A A A A (parent died)

0 9

A A A A A

AAATAAAAAAAA

1 16

A T T T T A 4 2

13—2

190 ExpermienU on the Rotifer Hydatina

These are the details of a similar experiment. A thelytokous
female was taken from XXXIV shown in the detailed test, which was
producing few arrhenotokous females. The results in the horse manure
solution were 115 thelytokous females and 5 arrhenotokous females.
The table shows the results of a thelytokous female in the green
protozoa culture.

T T A

-. \

AAAATTTTTAAAAAAAAAAAA 5 16

\

AAT^'TTATTAA 6 5

These experiments indicate that the increase in the number of
arrhenotokous females is due to the change from the horse manure
solution to the green protozoa culture and this is proved by subsequent
experiments. It is to be noted that an increase takes place in the first
generation.

Experiment II. A thelytokous female not taken from a tested
family was put in the green protozoa culture. A thelytokous female
of the second generation was transferred to the horse manure solution
which was very concentrated and the protozoa of which were very
numerous. The family was continued in this solution

T in green protozoa food T A

**TTAATTATATATTAAAA 8 9

_J

** A A A A A*T T A A A AT AT 4 10

* r in horse manure solution T A

\

A A AT T T 33

\

T T T AT T T T T T T A T T T T T 15 2

T T T A T AT T T A T T AA A T A T A T T A A A A A T T T T T T T T T A 22 14

I

T T T T T T T T ATT T T T T T T T T A T T T T T T T T T T T T T T T T 34 2

2^ j^ X X T T T T T T T T T T T T T T T TTTTTTTTTTTTTTTTTTTTT 40 0
T T T T T T T T T T T T T T T T T T T T T T T T T T T T T T T T T T T T T 37 0

T T 2 0

J

T TT T T T T T T T T T T T T T TTT T T T T T T T T T TTTTTTTTTTT 39 0

J

T TTT TT T T T T T T T T T T T T T T T T T T T TT T TT T TT 33 0

** The parents and many eggs in these two generations were killed through being
exposed to the sun's rays on a hot day.

Edith E. Hodgkinson 191

It is to be noted that the arrhenotokous females persisted through
4 generations when the rotifers were transferred to the horse manure
solution, after which none were produced.

Experiment III. A thelytokous female was put in green protozoa
culture ; after laying a certain number of eggs it was transferred to
concentrated horse manure solution in which the remainder of the eggs
of that generation were laid. This was repeated through two more
generations, the parent in each case being a rotifer that had hatched
out in the green protozoa culture. The table gives the results

Green protozoa culture T Horse manure solution

TAAAAAATTT yrrrrrrrrrTrrrrrTTrrrrr

J

AAAAAAATT yrrrrrrrjTrrrTT

\

TTTAAATTTATATT j<xTTTTTTTTTTTTTT

A definite change occurred in the same generation : whilst the
parent was in the green protozoa culture numerous eggs were laid
which hatched out into arrhenotokous females, but after transfer to the
horse manure solution no arrhenotokous females were produced. This
change occurred in each generation.

Expenment IV. A rotifer and its descendants were put in a culture
of green protozoa kept diluted with spring water so that it was com-
paratively poor in protozoa. The results were

Generation

T

A

1

17

11

2

17

9

3

25

19

4

6

10

Further experiments are needed here to show if a decrease in the
food supply and in the assimilation of the green protozoa affects the
output of males.

Experiment V. Four rotifers were isolated and put in green pro-
tozoa culture. They were kept in sunlight during the day and in
electric light duHng the night so that assimilation b}^ the green protozoa
was continuous. The results were

T

A

I

16

13

n

24

5

III

12

26

IV

7

19

192 Experiments on the Rotifer Hydatina

The results are too scanty to prove that increased assimilation by
the green protozoa causes a rise in the proportion of arrhenotokous
females. This is however not unlikely in view of the fact that Shull
and Ladoff ^ have found that increased oxygenation leads to the output
of a higher proportion of arrhenotokous $ $ .

The results recorded in this paper serve to confirm the conclusions
of other observers as to the alteration in the proportion of arrheno-
tokous females brought about by a change of diet from protozoa without
chlorophyll to protozoa which contain this substance. Both Whitney
and Shull have attempted to push the analysis further and to bring
about the alteration in the relative proportions of the two kinds of
females by means of substances of more simple constitution chemically.
Their attempts have shewn that a wide field of inquiry is still open
here, and I do not feel that at present I can add anything to the full
and suggestive discussions which are to be found in their papers.

1 Journ. Exp. Zool. 1916.

THE HISTORY OF PRIMULA MALACOIDES,
FRANCHET, UNDER CULTIVATION.

By ARTHUR W. HILL.

(With Plates XI and XII and one Text-figure.)

The history of Primula malacoides under cultivation is of some
interest, as though of only recent introduction it has been very exten-
sively grown in this country and has already displayed a considerable
range of variation.

The species was discovered by Pere Delavay in March, 1884, in
cultivated fields at Tali, Yunnan, a locality to which it appears to be
restricted, and a description of the plant was drawn up by Franchet in
1886 ^ The introduction of the plant to cultivation is due to Mr G.
Forrest, who collected specimens in the Tali valley in 1905 and 1906.
From the seed sent home to Mr Bully of Ness (of the firm of Messrs
Bees, Ltd, of Liverpool, for whom Mr Forrest was collecting) plants
were raised which flowered in 1908'-'. Mr Forrest records that it is
abundant in the Tali valley near the city at elevations of from 5000
to 7000 feet, and that it was thriving in rather moist sunny situations.
The illustration of a plant raised by Messrs Bees in 1908 is given in
Notes of the Royal Botanic Garden Edinburgh, Vol. iv. PI. XXVII B,
and shews the slender, graceful inflorescences of the original plants^
A figure of a more robust specimen is given in the Gardeners Magazine
for Dec. 5, 1908. The plant figured was shown by Messrs Bees, Ltd,
and given an award of merit at the meeting of the Royal Horticultural
Society's meeting of Nov. 24.

1 Franchet, in Bull. Soc. Bot. Fr. Vol. xxxiii. 1886, p. 64.
"^ Messrs Vilmorin had received seeds in 1895 but they failed to germinate.
3 The same figure is reproduced in the Gardeners' Chronicle of Dec. 5, 1908, p. 397, in
connection with a short note on the plant by Prof. Bayley Balfour.

194 History o/" Primula malacoides, Franchet

According to Franchet's original description the colour of the wild
flowers is rose and the limb of the corolla is said to be 10 — 12 mm. in
diameter.

Prof. Balfour in his note in 1908 says : " To some its tint of pink or
lavender-pink, in contrast with the grey mealy surface of the flower
stems, will be hardly pronounced enough for effect, but time and culti-
vation will alter that, and it may be predicted that this Primula is likely

to become a general favourite " This prophecy was quickly fulfilled.

It may be remembered that in the case of Primula ohconica^ no striking
variations occurred 'until the plant had been in cultivation for some
fifteen years, and double flowers were only developed in the twenty-first
year of its cultivated existence. With P. malacoides however the whole
range of variations was initiated in about four years from the date of its
introduction to this country. A figure in The Garden^ for 1912 shows
the more robust character of the plant as compared with those first
cultivated in this country.

The only coloured illustration of P. malacoides of which we are
aware is that published in the Revue Horticole for 1912, p. 156, which
depicts a plant very like the type as originally introduced. The plant
does not appear to have been cultivated in France until 1911 when it
was exhibited by Messrs Vilmorin at a meeting of the French Horti-
cultural Society in the spring of that year^ M. Mottet, the author of
the article, gives a short historical account of the plant and, what is
of considerable value, has drawn up a detailed description from living
specimens cultivated in Paris in 1911.

According to this description the corolla measured 15 — 20 mm.
across, the segments were deeply notched and the colour was a soft
rose-lilac with a gi-eenish-white centre. The leaves were 6 — 8 cm. long,
oval rounded at the apex and cordiform at the base.

The leaves on our collected specimens are up to 9 cm. long.

From this account it appears that the effect of cultivation has been
to increase the size of the flowers from the maximum of 12 mm. in
diameter given by Franchet to 25 mm., which is the dimension of the
largest flower yet noticed.

In England about this time apparently the plant varied in several
directions. Messrs Sutton and Sons sent to Kew on April 6, 1912, a

1 A. W. Hill, " The history of Primula obconica, Hance, under cultivation, &c. " in
Journ. Genetics, Vol. ii. 1912, p. 1, with plates.

-' The Garden, March 30, 1912, pp. 157, 158, with fig.

'^ Rev. Hort. 1911, p. 218; Journ. Soc. Nat. Hort. Fr. 1911, p. 181.

Arthur W. Hill 195

white sport of P. malacoides (PI. XI, Fig. 3) and informed ns that they
had also a " deep mauve quite distinct from the type."

The flowers of the white sport measured only 1 cm. across, the
corolla segments were narrow, only 4 mm. broad at the apex and almost
oblong with a shallow notch. The eye was deep yellow. The calyces
were very mealy and the inflorescences much more slender than in the
type.

Messrs Bees, Ltd, who also had the white sport, remarked that the
habit of the plant is very distinct fro-ra that of the rose-coloured form,
the inflorescences being much more erect and the distances between
the whorls of flowers much shorter. The flowers of the white form
sent by them to Kew measured 2 cm. across, whilst the largest lilac
flowers measured 2*5 cm. The more robust character attained by the
plant in its fifth year of cultivation is well shewn in the figure in The
Garden for March 30, 1912, p. 1-57.

In this same year P. malacoides also produced double flowers for
the first time, and this sport occurred among a large batch of plants
grown by Mr L. R. Russell at his Cedar Nursery, Ham, Surrey (PI. XI,
Figs. 5—12 ; PI. XII, Figs. 2 and 3). All stages in the doubling of the
flowers were noticed, and on the same inflorescence in some plants every
gradation from single to fully double flowers could be found. The
doubling is like that of the old double white P. sinensis and of P. ob-
conica, being due to out-growths from the apex of the connective of
the anther resulting in a " hose-in-hose " type of flower. In the best
cases one or more whorls of extra segments have been developed, each
segment having the characteristic apical notch.

A good figure of the double form was given in The Garden of
Dec. 13, 1913, p. 624, from a plant grown by Messrs Bees, Ltd, and
the sport appears to have originated in Cheshire independently of its
appearance at Messrs Russell's nursery ^

The double white variety was produced at Kew as a result of crossing
the double lilac with the single white form and flowered in 1915. This
spring (1917) Messrs W. and J. Brown of Stamford, showed double white-,
double lilac and double mauve forms at a meeting of the Royal Horti-
cultural Society.

* Although the double form of P. malacoides originated first with Messrs Russell in 1912
it was Messrs Bees' plant shown at the meeting of the R. H. S. on December 2, 1913, which
received an Award of Merit, see G. C. March 14, 1914, p. 180.

2 A figure of the double white form is given in Gardeners^ Chronicle, March 2, 1918,
p. 91, and the foliage of the plant is distinctly " fern-like."

196 History 0/ Primula malacoides, Franchet

The seed of these three doubles was obtained from a private grower
and sown in the autumn of 1916.

Another break which occurred in 1912 also at Messrs Russell's
nursery consisted in the fimbriation of the corolla segments, a variation
which has commonly developed in primulas under cultivation (PI. XI,
Fig. 13 ; PI. XII, Fig. 4). In the examples seen in 1912 the apical notch
still existed, but the lobes of the segment were more or less deeply cut
or laciniated and there was also a tendency to the formation of tooth-like
projections in the sinuses between the segments.

This latter tendency is especially noticeable in the forms with narrow
segments and wide sinuses (PI. XI, Fig. 15; PI. XII, Fig. 5).

In some cases only a single segment or one side of a segment of a
flower may show fimbriation.

With fimbriation of the corolla is associated also a fimbriation of the
calyx. Normally the calyx bears five simple acute teeth, but where the
corolla segments are laciniate the calyx teeth tend to be conspicuously
and sharply toothed (PI. XI, Figs. 13 and 14). With the double flowers
the calyx is like that of the type.

Fimbriation and doubling then appear to be two quite independent
variations and fundamentally different, though the inception of both is
undoubtedly due to cultivation.

Yet another variation was noticed this same year with regard to the
scent of the plants, the foliage of some being almost scentless, while in
other cases it was distinctly scented, the odour being like that of pine
apples. Attention has also been called to this fact by Dr Mac Watt,
who states that the pine-apple scented form is more robust and sets
seed freely, while the scentless form has smaller flowers and does not
set seed according to his experienced

Variation in the direction of greater intensity of flower colour has
also been marked in P. malacoides ; though the range of variation is
not yet so wide as in P. obconica, both lilac and deep mauve forms are
common, and a deep pink shade was exhibited by Mr James Box in
1914^. This particular form was further distinguished by its vigorous
growth. The plant having " stouter stems, a more robust habit, and
flowers twice the size of those of the type." The flowers measured
2"5 cm. in diameter. This giant form appears to be the result of
beneficial cultivation rather than of any hybridisation, the suggestion
that Primula Forhesii may have been accidentally crossed with P.

1 The Garden, January 6, 1912, p. 2.

'^ Gardener'x Chronicle. March 14, 1914, p. 180, fig. 82.

Arthur W. Hill 197

mcdacoides seems no more tenable than that the deep rose colour of
some forms recently exhibited is due to the influence of the pollen
of P. rosea. According to the editorial note in the Gardeners' Chronicle
numerous experiments have been made to cross P. malacoides and
P. Forhesii, but no fertile seed has resulted.

The most distinct colour form of P. malacoides is that known as
" Rose Queen," the flowers of which are of rich rosy colour. A figure
of this is given in The Garden of Feb. 27, 1915, p. 104. The variety,
was raised by Mrs Denison of Little Gaddesden, Berkhamsted, and
is very vigorous in habit wath fine trusses of flowers. It differs from
other forms chiefly in the breadth of the perianth segments which have
become so much broadened that they touch one another and the flower
has a rounded outline more like that of a polyanthus.

As in the case of P. obconica, the hybridisation of P. malacoides
with other species of Primula has been attempted by many gardeners,
but there is no evidence that success has attended any of the experi-
ments. The species in a remarkably short space of time has responded
to cultural conditions and has sported in the directions of doubling,
fimbriation of the corolla segments, albinism, and various colour
changes as well as in the direction of increased vegetative vigour.
All these changes however must be attributed to the skill of the
gardener rather than to the efibrts of the hybridiser.

P. malacoides has varied under the influence of cultivation in
precisely the same manner as have P. sinensis and P. obconica^, but
the time interval required for the disturbance of its equilibrium has
been very short.

1 It is of interest that P. denticulata which in 1912 was being grown in large quantities
by Messrs Russell at Cedar Nursery, Ham, showed a great range of colour forms, through
every shade of lilac, pale purple and violet to a distinct bluish shade and also white. The
shape and form of the flowers was also very varied from small with narrow sharply pointed
segments to very large flowers with rounded segments very far removed from the type.
Fimbriated flowers were also noticed but no doubling. The leaves were also very diverse
in form in some cases being quite coarse. In this species there is no question of any
hybridisation having been attempted.

198 History o/ Primula malacoides, Francliet

EXPLANATION OF PLATES.

PLATE XI.

Figs. 1 and 2. Primula malacoides grown at Kew in 1912, very similar to the plant as

originally introduced.
Fig. 3. White variety which appeared at Messrs Sutton and Sons in 1912.
Fig. 4. A deep coloured form flowered at Kew in 1914.
Figs. 5, 6, 7. Stages in doubling Fig. 7 being a complete double, see Fig. 11, from Messrs

Bussell in 1912.
Fig. 8. An enlarged drawing of the flower seen in Fig. C, cut open to show the extra

segments developed from the connective of the anther.
Figs. 9 and 10. Details of the development of the perianth-segments from the connective.
Fig. 11. An enlarged drawing of Fig. 8, the perianth cut open showing a complete inner

perianth and further developments commencing from the anthers.

Fig. 12. An extra segment and anther from the flower represented in Fig. 11.

Fig. 13. A large fimbriated flower grown at Kew in 1914 the flower measured 2 35 cm. in

diameter.
Fig. 14. The calyx of the above magnified showing the toothed margins of the segments.
Fig. 15. A flower showing incipient lobes in the sinuses.
The drawings are by Miss M. Smith.

PLATE XII.

Flowers from a group of seedling plants of P. malacoides exhibited by Messrs R. L. Eussell,
Nurserymen, Richmond, at the Royal Horticultural Society's show, April 9th, 1912. The
plants were raised at Cedar Nursery, Ham, and selected from a batch of 15,000 plants.

All the flowers are natural size.

Fig. 1. Flowers showing early stages in flmbriation.

Figs. 2 and 3. Early stages of doubling with two more or less completely double flowers.

Fig. 4. A flower with the segments conspicuously fimbriated.

Fig. 5. Three flowers showing an attempt to produce outgrowths from the sinuses.

Fig. 6. Three large single flowers measuring 2*2 — 2*35 cm. in diameter.

Photographed by C. P. Raffill.

JOURNAL OF GENETICS, VOL VII. NO. 3

PLATE XI

U ^

c.^

'G:

r- ,:n

S'^^-Mi-

't*--^

-MY:

7

-JifC

I'^f'

. J^-

13

^ ^1

15

JOURNAL OF GENETICS, VOL VII. NO. 3

PLATE XII

Fi«. 1.

Fig. 2.

Fig. 3.

Fig. 4.

Fig. o.

Fi.'. 6.

PARTIAL SELF-FERTILIZATION CONTRASTED
WITH BROTHER AND SISTER MATING.

By RAINARD B. ROBBINS,
Instructor in Mathematics, University of Michigan, Ann Arbor, Mich.

In a recent issue of this Journal, A. B. Bruce (1) has published an
article entitled " Inbreeding," the purpose of which was " to express the
genetic consequences if, in each generation, selfing and mating at random
take place in the fixed ratio s : (1 — s)." Bruce gives as one of the reasons
for the importance of such a problem, that the results of brother and
sister mating may be assumed to be given by a proper choice of s. The
excuse for such an assumption is suggestive parallelism in particular
cases. The fact is that the assumption is erroneous, as the results of
this paper will show.

The theory will now be developed for combined selfing and random
mating. The results for the general case of brother and sister mating
have already appeared(2). It is by comparing these results that our
conclusions are drawn. Only one pair of simple Mendelian factors will
be considered.

Suppose we start with a population which can be represented by

r^AA +SoAa+toaa.

This means that the pure dominants, the heterozygotes and the recessives
occur in numbers proportional to 7*0, Sq, to. For convenience we choose
ro. So, to such that

ro + So + to = 1.

Suppose that self-fertilization occurs in a part of each generation
indicated by the proper fraction a- and that mating is at random in the
remainder of each generation. Of course we assume that the individuals
which are self-fertilized are chosen at random in each generation.

200 Brother and Sister Mathig

Let Vn, Sn, tn represent the proportions of the three types of zygotes
in the nth. generation, and let

Tfi + S,i + fc,j = 1 .

The analysis of the formation of the zygotes of the nth generation gives
the following recurrence relations :

Vn = a- (r„_i + s„_i/4) + (1 - o") (2r,,„j + .s-„_i)V4, (1)

S,, = a Sn-J2 + (1 - a) (2r„_i + S„_,) (2tr,-^ + Sn-i)/2, (2)

tn = a- {tn-i + Sn-J4<) + (1 - a) {^n-r + Sn-,fl^ (3)

On the right hand side of each of these equations, the first term is the
contribution of self-fertilization, and the second term is the contribution
of random mating. To solve these equations we notice that

2r,, + s,i = 2r„_i + s„_i = =2ro + So-

Thus we have that 2r„ + s„ is constant. Let

p = 2ro-l-5o.
Then equations (1), (2), (3) simplify giving,

r„ = o-r„_i/2 + [o-p + (l-(7)/D2]/4, (4)

6-„ = crs„_i/2 + /o(l-o-)(2-p)/2, (5)

tn = (Ttn-^l'2. + [a{2-p) + {l-a-){2-pf]l4> (6)

The solutions of these equations are,

■cr^"^ , /30"(1 — p)-\- p-
2(2- a)

^""UJ ^""^ 2(2"^^"^ ^^^

-(1)"".+"^^^^^ («)

.^(i)%..<^--'>^-:-^-^-> («)

The constants Co, d^, e^ are determined by the initial conditions and are

c„ = ro + p(/jo--p-o-)/2(2-o-), (10)

rf„ = So + p(2-p)(o--l)/(2-o-), (11)

eo = 4 + (2-p)(p + cr-po--2)/2(2-o-) (12)

Equations (7), (8), (9) give the results for our problem of combined
selfing and random mating. Incidentally we may specialize these

Rainard B. Robbins 201

equations so as to get the results for random mating and for self-
fertilization. If we set cr = 0, we have, for random mating,

r„=(2ro + 6-„)V4; «„ = (2ro + s„) (2^o + 5o)/2 ; tn = {2t, + s„y/'^.

Setting a = 1 gives for self-fertilization

rn = r„-|-6-o(l-l/2")/2; Sn = So/2^; tn^t, + s,(l - l/2'')/2.

The results for brother and sister mating, starting with a family
VqAA + SoAa + toCia, have been published by the present author (2).
They are

r„ = />/2-Z„/4«+', (13)

s„=2X„/4»+\ (14)

^„ = (2-p)/2-X„/4"+S (15)

in which Ln = ^[^3(1 + V5)«+i -K,{1- V5)«+^] ,

^^ (H-V5) (1-V5), . ,,
and Ks = ^ ^ So + ^ — ^ — - (so - 4ro«o) ,

From these results we readily calculate that as n increases indefinitely,
rn, Sn, tn approach respectively the values 2ro-fSo, 0, 2tQ + s^ In words,
the heterozygous type tends to disappear in brother and sister mating and
the homozygous types approach a proportion equal to that of their respective
gametes in the original population.

But equation (8)) shows that in our combination of self-fertilization and
random mating, the heterozygous type can never disappear, if a is different
from unity. In fact equation (8) shows that

lim*« = p(l-cr)(2-p)/(2-o-).

n=oo

Thus it is clear that no such combination of random mating and self-
fertilization can represent brother and sister mating. In every case of
inbreeding which the present writer has examined, the proportion of
heterozygotes approaches zero as the number of generations increases.
Equation (8) shows that so long as a- =^1, i.e. so long as we have a fixed
proportion of each generation mating at random while the others are
self-fertilized, the proportion of homozygotes cannot vanish. It would

202 Brother and Sister Mating

therefore seem unwise to assume without pretty good evidence that any
form of inbreeding is equivalent to such a problem.

However, as Bruce points out, the combined problem of random
mating and self-fertilization is directly important, since it is just the
sort of mating which occurs in certain problems.

REFERENCES.

1. Bruce, A.'B. Journal of Genetics^ A^'ol. vi. 1917, p. 195,

2. ROBBINS, R. B. Genetics, Vol. n. 1917, p. 489.

November 23, 1917.

GENIJTIC STUDIES IN POULTKY.

I. INHERITANCE OF LEG-FEATHERING.

By R. C. PUNNETT, F.R.S, and the late
Major P. G. BAILEY, R.F.A.

[The results recorded in this paper are the outcome of some
experiments started in 1910 and designed to throw light upon the
inheritance of certain features in poultry, more particularly of weight,
broodiness, and egg-colour. The data presented here are but a by-
product of these other investigations, but in view of the economic im-
portance of the species to which they relate I have thought it worth
while to place them on record. From 1911 until the outbreak of war
the work was carried on jointly by Mr Bailey and myself. Thence-
forward he was unable to take any active part in it though his interest
never flagged until his death in action last year. Nevertheless, deep as
is my regret that it must be so, responsibility for the views expressed in
this paper rests with me alone. R. C. P.]

The data recorded below were derived from two distinct crosses.
In each case the Langshan was the parent with feathered leg, the clean-
legged parent being in one case the Brown Leghorn and in the other
the Gold-pencilled Hamburgh.

Some of the later experiments involved a mating between a bird
derived from the Langshan-Hamburgh crosses and an F^ hen ex
Langshan x Leghorn.

The relation between the various birds used in the whole series of
experiments is set out in the pedigree on p. 204 ; the actual data are
given in the tables on pp. 206-207.

The type of Langshan used was what is known as the Croad Lang-
shan. In our experience the leg-feathering is not very heavy but
extends fairly uniformly down the leg. It was found in all the birds

Journ. of Gen. vii 14

204

Leg -feathering in Poultry

50 O CO ■*
C<1 (N <N e(5

o o o o o

m 00 05 O i-H

^ to «o t- t-

H
<
O

O
t— I

Oh

oo o

2 I — 5r

G

vJ

cS

'cc

bO

0

e6

vJ

ec

^3

tH

CO o
IM 5<I

■** 00

O 3
d °8
2 EH

ft a

a §

T3 r<

o o

00 to
* _

I ^

I iffl

R C. PUNNETT AND P. G. BaILEY 205

which we bred, 38 in number (cf. Table, Pen 6, 1911, and Pen 3, 1912),
and did not shew much variation.

44 Fi birds were bred from the mating Langshan $ x Brown Leg-
horn {/• and its reciprocal. All of them exhibited some degree of leg-
feathering though in all cases it was less than that of the pure Langshan.
The extent of the feathering varied but no attempt was made to grade
the birds into different classes. Of the F^ birds 4 </(/ and 14 $ ?
were used to produce an F2 generation, viz. Pen 10 and Pen 11 in 1911,
Pen 13 in 1912, Pen 4 and Pen 11 in 1913. The number of F^ birds
from these pens was 429. Of these 323 shewed feathering on the leg
while 106 had clean legs — a very close approach to a 3 : 1 ratio. The
extent of the leg-feathering varied considerably though here again no
grading was attempted. But our records make it clear that in some
cases the feathering was rather heavier than in the average Langshan,
while in others it was quite slight. In three cases F^ birds from this
cross were bred together. Pen 16, 1912, consisted of two such birds,
both with feathered leg. The distribution of feathered and clean-
legged birds, 34 to 15, suggest that both parents were here hetero-
zygous. In Pen 11, 1912, an i^2 cT with feathered leg was mated with
an F.2 $ which had clean legs. Of the 23 chickens hatched 11 had
feathered and 12 had clean legs, a near approach to the expected
equality on the assumption that the cock used was heterozygous. In
the third pen, Pen 13, 1912, two feathered-legged birds bred together
gave 7 chickens all with feathered legs. The numbers are too small to
draw any definite conclusion, but it is not unlikely that at any rate
one of the parents here was homozygous for the character. One other
F^ bird, $ 487 ex Pen 11, 1913, was subsequently tested in Pen 5, 1916,
by mating her with cT 35 which, from his breeding, must be supposed
to have been heterozygous. As all the 22 children produced had
feathered legs $ 487 must be regarded as in all probability a homo-
zygous F2 bird. All the data from the Langshan-Leghorn cross are
evidently consistent with the view that leg-feathering behaves as a
simple dominant to clean leg. On this view we must suppose that
the variation in the extent of the feathering, though partly dependent
upon whether the bird is" homozygous or heterozygous for the feathered-
leg factor, must be dependent also upon some other factor or factors
not yet analysed.

The Langshan-Hamburgh results, with one exception, are not dis-
similar. 12 Fi birds were raised in Pen 9, 1912, and Pen 15, 1913.
Of these 11 had feathered legs, while one, which will be referred to

14—2

206 Leg-feathering in Poultry

OS ,^ _, --I '-' >-*

^ r* '2 !:1 05 "3^ oi

S 2 2 55 -^ ^ -^

^ <S o o

«0 ■=* , CO o" ^ "^ '^ I

a s - ^ c c fl '

m ^ r! '^ Oi O) O

(2 &H (2^ ^ (^ (S f^

<B « g X « <U «

CO ii fl — ■ —

'-'-I "> " . T . T . T i-H ?0 ,-1

^^ J

-- -^ i~ *"' O CO

W »<3 ^ OS: <?«^

- J

CO

s

I I I

"-I C5 OJ ^

c CI a ' --H

Ck CM Ph g

!«! H J< (^

<a <Si <a

t^ ft, ^

t; i-H ^ <M th ."i;

CO O 00

p^CLi cli otOiP-iCkaH

^

a

fl

c

C

a

<D

tt

OJ

Oi

O)

OJ

P-i

Oh

Cui

(In

Ph

0-

R. C. PUNNETT AND P. G. BaILEY

207

/ ^

« 1-4 I -"It IM

CO O «5

C94

*" -^iHCOi-l iHOJMTjt i-ti-|i-( >-(

^ «IH «-<' Sh

S X X X

^ s ^- <«• c

«4-.* fl «t^ «<-l

X X X X

<^' ** E2 S

»-l iH rj OS

05 OS 5 "-I

(In Oh

CO
i-H

OS

1— 1

I-H
OS
I-H

CO
I-H

rH

CO

1— 1
OS
I-H

CO
>-H
C5
r-l

CO

I-H
C5
I-H

to

OS
I-H

CO
1-H
OS

CO
I-H
OS
I-H

CD
i-H

«5

I-H

«o

I-H

-Hi"

00
I-H

o

c

1 a

c

0-1

1 c

CM

X

Oh
X

Jii *< f*!

&H 6h

a

(M

SO

CO

s?

iH

I-H

c

50

CO « J2 CO CO r, CO
-H !-H 60 -H rH -t" ^

•2 eS -^ "S '^ ■* -2

S c S

c8 c3 o3

W K W

ffO CO CO
-H -H rH

X? «? I-H

1* Cq Ca TO

PM

OS

I-H

l»t

^

<i>

t>

&C

a

P^

a

Ph ^

a a

CL,

K W

W

.J3 O

OS iffl CO ift

iii PM p^ Ph Ph

208 Leg-feathering in Poultry

below, was recorded as clean-legged. F^ birds with feathered legs
were mated together in Pen 7, 1914 and 1915, Pen 14, 1915, Pen 6
and Pen 7, 1915. Of the 148 birds produced 117 had feathered and
31 had clean legs — expectation being 111 and 37. F^ birds were also
mated with pure Hamburghs in Pen 7, 1915, Pen 14, 1915, and Pen 6,
1916, They gave 93 offspring of which 37 were feathered and 56
were clean in the leg, the latter class being more numerous than was
expected.

One F2 bird, $ 250/13, was mated with a Hamburgh cock and gave
7 chickens, all with feathered legs. Probably she was homozygous for
the character ^ Of her sons, one, (/ 35/15, was mated with $ 487/13,
an F^ bird from the Langshan-Leghorn cross (cf p. 204). Two birds
from this mating were tested in Pens 10 and 20, 1917. Both proved
to be heterozygous. (/ 271/16 was mated with $ 454/13, an F^
Langshan-Hamburgh hen. The result, 18 feathered and 8 clean-legged
chicks, is close to the expected 3 : 1 ratio on the assumption that he
was heterozygous. Further, with $ 210/16, a hen of mixed Hamburgh-
Sebright-Leghorn origin, he gave 12 chicks with feathered and 9 with
clean legs — a proportion near the expected equality of the two classes —
^ 44/16, a brother of ^ 271/16, was mated with a sister of $ 210/16.
The 36 chickens produced are nearly equally distributed between the
two expected classes.

So far the results of the Langshan-Hamburgh crosses may be said
to have borne out those derived from the cross between the Langshan
and the Leghorn. There is however the aberrant case above referred
to where a clean-legged bird, % 64/12, came from the Langshan-
Hamburgh cross. This bird behaved as though she had feathered
legs; with her brother she gave (in Pen 18, 1913) 55 chickens of which
those with and those without feathered legs were almost exactly in the
proportion 3:1. In the following year she was mated with a pure
Hamburgh (Pen 5, 1914). Of the 25 chickens produced 13 had
feathered and 12 had clean legs. One of her daughters, $ 250/13, has
already been referred to as being in all probability homozygous for the
feathered leg.

Occasionally therefore a heterozygous bird may fail to put up any
leg-feathering. Sometimes the leg-feathering may be reduced to one
or two small feathers at the base of the outermost toe. In two cases
chicks on hatching were recorded as clean-legged though subsequent

1 She was also mated with <j 44 in Pen 20, 1917, a bird known to be heterozygous.
Here again, though only five chickens were hatched, all had feathered legs.

R. C. PUNNETT AND P. G. BaILEY 209

examination, when they were nearly full grown, revealed the presence
of one or two very small feathers in this position. Instances of this
extreme reduction of the feathering are comparatively rare. In the
great majority of cases there is no difficulty in appreciating the presence
of the feathering either at hatching or at any later stage in the bird's
life. But apparently in a few cases the feathers fail to break through
in early life, and occasionally, as in $ 64/12, they never break through
at all.

In 1917 a few observations were made on the extent of the feathering
in Pens 10 and 20. Both cocks were birds with legs slightly feathered
though not of the very reduced type. Each was mated with a hen of
mixed Hamburgh-Sebright-Leghorn origin. Presumably these hens
were pure for the clean-leg character for during the past eight years
many hundreds have been bred from birds involving these three breeds
without the appearance of a single case of feathered leg.

There are two points of interest in connection with the feathered-
leg birds from these two pens. In the first place the offspring may
shew a grade of leg-feathering much heavier than that of the male
parent. In Table III these chickens^ have been graded roughly into
three classes: (a) strong, in which the feathering is not dissimilar to that
of the pure Langshan, (b) moderate and (c) slight, including birds with
feathering similar to or less in extent than that of the male parents.

TABLE Iir.

Male oflFsprlng Female offspring

Father Mother Strong Moderate Slight Strong Moderate Slight

<? 44 ? .S88 2 1 1 — — 4

cr271 ?210 1 1 3 — 1 3

<?271 ?454 6 — — 2 3 2

Totals ... 9 2 4 2 4 9

In 6 cases out of 17 from $ $ 388 and 210 the leg- feathering was
markedly heavier than in the father, suggesting that the clean-legged
bird may carry some modifier affecting the extent to which the feather-
ing developes. </ 271 was also mated with $ 454, an F^ bird with
feathering well marked though not as heavy as in the Pure Langshan.
Here again some of the offspring were more strongly feathered than
either of the parents.

1 The numbers are rather less than those given in Table II. Some of the chickens
died without their sex being determined.

210 Leg -feathering in Poultry

The second point of interest appears in all three of the matings.
There is a marked tendency throughout for the (/</ to be more heavily
feathered than the % %. More than half the cTc/* ^-re found in the
highest of the three grades, while the majority of the % % occur in the
lowest one. In his experiments on leg-feathering (" booting ") Davenport
('09) grades his results but he does not appear to have noticed any
connection between the degree of leg-feathering and sex. Since the
sex of the offspring is not recorded in his tables it is not possible to
test this point by his data. It is however one that may well be borne
in mind by any future investigator of this character.

Several observers have published data on the inheritance of feathered
leg in poultry. Of these the results recorded by Davenport are the
most extensive and at the same time the most instructive. Many of
his crosses, especially those in which birds of the F^ and later gene-
rations were used (cf Table 38, pp. 52, 53) are too involved for analysis,
but for the Cochin and Dark Brahma crosses (Table 34, pp. 48, 49)
the data are more favourable, and consideration of them has led me to
the conclusion that a simpler interpretation can be found for them
than that which Davenport has suggested. If we suppose that in these
feathered-legged races two factors A and B may be present, either of
which determines leg-feathering, the different proportions of clean-
legged birds in the various F^ matings recorded by Davenport may
very well occur. In demonstration of this I have constructed Table IV
on p. 211 from the data given on pp. 48 and 49 of Davenport's paper.
The assumption is made that the original parents on the feathered-leg
side all transmitted the factor A to the F^ offspring subsequently tested.
In many instances they also transmitted factor B while in others this
factor was not passed on. It is assumed that the clean-legged breeds
used, whether white Leghorn or Minorca, lacked both of these two
factors. The restatement of Davenport's results set out on p. 211 clearly
supports the view I have just advocated as to the. factors which give
rise to the feathered leg. The actual numbers are in fair accordance
with expectation throughout and in many cases the two sets of figures
are practically identical. It is probable therefore that in Cochins and
Brahmas two factors are concerned in leg feathering, and not one only as
in Langshans. This tallies with the grade of leg-feathering found in
the two former breeds, in which the feathers are more numerous and
more highly developed than in the last named.

Cochins were also used by Hurst who found that the F^ birds pro-

R. C. PUNNETT AND P. Gr. BaILEY

211

TABLE IV.

Pen
650

654
608

659
655

Nature of cross
Black Cochin x Wh. Leghorn

Wh. Leghorn X Dk. Brahma

Ref. nos. of ^i

Ref nos. of F
Father

Suggested
1 constitutions
of Fi Parents

Fi offspring

Expectation

Mothers

f.l.

n.f.l.

f.l.

nj:i.

170, 263 )
278, 364)

265

AaBb X Aabb

133

11

126

18

361

265

Aabb X Aabb

26

9

26

9

602, 828, 640 )
696, 767, 697 )

704

Aabb X Aabb

72

25

73

24

384, 248

409

AaBb X Aabb

78

7

75

10

249, 395, 385

409

Aabb xAabb

107

37

108

36

762, 503, 382 |
250, 737, 387 \

375

AaBb xAabb

146

16

142

20

720, 724, 728 ^
730, 732, 734 .
761, 800, 721 J

666

AaBb X AaBb

53

1

51

3

724, 734, 800 -j
720, 728, 761 '
732, 730, 721 J

254

AaBb X Aahb

96

14

96

14

742, 690, 631 .

618, 700, 703 I

637

AaBb X Aabb

221

17

208

30

743, 599 J

524, 576, 638

637

Aabb X Aabb

57

17

56

18

632 Minorca x Dk. Brahma

duced from crosses with clean-shanked breeds such as the White
Leghorn and Hamburgh all shewed leg-feathering, though in less degree
than the Cochin parent. At the time of Hurst's experiments the idea
of multiple factors had not yet been put forward. It was not at that
time fully realised how important it was to keep entirely separate the
record of each bird used as a parent. Many similarly bred hens were
run with a given cock on the assumption that they were of like con-
stitution though, as we know to-day, this was not necessarily the case.

Nevertheless Hurst's results can to some extent be reconciled with
those of Davenport. In Exp. 6 (p. 135) a number of F^ birds were
raised from (a) 2 Hamburgh ? $ x Buff Cochin <f and (6) from Buff
Cochin $ X Hamburgh (/. The 107 birds reared all had feathered
legs, though the feathering was distinctly less than that of the Cochin.
In Exp. 8 (p. 138), 12 of these i^i ? ? were mated with an F, (^, and
gave 115 chicks with feathered and 4 with clean legs.

Assuming that all of these birds were AaBb in constitution expecta-
tion would be lll'o fl. : 7'5 n.fl. Twelve F^ ^ % (presumably different
ones from the above since the experiment was also made in 1903) were

212 Leg-feathering in Poultry

run with a Hamburgh </ to give 35 f.l. and 21 n.f.l. chicks, expec-
tation being 42 : 14. The number of clean-shanked chicks is rather
higher than expectation, but we do not know for certain that all of
the hens were in constitution AaBb. If one or two were Aabb the
excess in the n.f 1. class would be easily accounted for.

Lastly a Hamburgh-Cochin from Exp. 6 was mated with 12 Leghorn-
Houdan $ % (Exp. 13, p. 141). Here again expectation is fl. and n.f.l.
in the ratio 3:1. The actual numbers were 195 fl. and 112 n.fl. The
considerable excess of n.fl. birds suggests some complication. Possibly
some of the potentially fl. birds do not shew the feathering as in
% 64 (cf p. 209 above). There is also the possibility that some of the
Leghorn-Houdan hens may have carried a factor inhibiting the develop-
ment of the leg feathers. This possibility is indicated by some experi-
ments of Bonhote. From the Silky as the feathered and the Yokohama
as the clean-legged parent he obtained 2Fi birds of which both had
clean legs. Bred together this pair gave 24 chicks of which 8 were fl.
and 16 were n.fl. If the Silky carried a factor A for feathered leg and
the Yokohama an inhibitor factor I, the constitution of the F^ birds
might be regarded as Aali. Such birds in F.2 should give fl. and n.fl.
in the ratio 3:13. The figures obtained by Bonhote, viz. 16 n.f.l. and
8 fl. are not hopelessly out of accord with those expected, viz. 19*5
n.fl. : 4'5 f 1., when the smallness of the numbers is taken into account.

In the only other experiments of this nature with which I am
acquainted, those of Cunningham dealing with a Jungle fowl — Silky
cross, the birds reared were too few to give a result of much value.
The F^ birds had feathered legs, while of 10 F2 birds 9 were fl. and
1 was n.fl.

In so far as our present knowledge of the heredity of leg- feathering
goes we may say that in certain crosses it may behave as a simple
dominant, though dominance is always incomplete. In breeds with
more heavily feathered legs there are grounds for supposing that two
factors for leg-feathering may be present. Further, it is possible that
some strains of clean-legged birds may carry an inhibitor for the
feathered-leg character. The extent of the feathering shews consider-
able variation even in birds similarly constituted in respect of the
factor or factors for feathered leg. Rarely a potentially leg-feathered
bird may fail to shew even traces of the feathering. Probably clean-
legged birds will be found in some cases to carry factors, hitherto not
identified, which affect the extent to which the feathering is developed.
In support of this is the fact that a clean-legged bird, mated with a bird

R. C. PUNNETT AND P. Gr. BaILEY 213

in which the leg-feathering is slight, may give offspring in which the
legs are strongly feathered (cf p. 209).

The experiments of which an account is given above form part of a
series of investigations on heredity in poultry for which the means
have been provided out of the Fund controlled by the Development
Commission,

REFERENCES.

BoNHOTE, J. L. "Preliminary notes on the Heredity of certain characters in a
cross between Silky and Yokohama Fowls." The Cairo Scientific Journal,
No. 9, Vol. vni. April, 1914.

Cunningham, J. T. "Mendelian Experiments on Fowls." Proc. Zool. Soe. 1912.

Davenport, C. B. " Inheritance of Characteristics in Domestic Fowl." Carnegie
Institution of Washington Publication, No. 121, 1909.

Hurst, C. C. "Experiments with Poultry." Rep. Evol. Cowraittee Roy. Soc. ir.
1905.

ON THE OCCURRENCE, BEHAVIOUR AND ORIGIN
OF A SMOOTH-STEMMED FORM OF THE COM-
MON FOXGLOVE {DIGITALIS PURPUREA).

By EDITH R. SAUNDERS,

Lecturer, late Fellow, Newnham College, Cambridge.

In the course of an earlier investigation on the inheritance of the
peculiar modification of the flower known as heptandry, which is occa-
sionally met with in Digitalis purpurea, it was incidentally noted that
some individuals exhibited a distinct variation from the recognised type
as regards surface character^ The typical plant is generally described
as having all the aerial green parts covered with hairs, those on the stem
and petioles being so abundant as to give these structures a downy,
grey appearance. In the particular individuals in question, on the other
hand, the vegetative region of the stem was conspicuously green, smooth
and shining, and the leaves were distinctly less hairy on the upper sur-
face and deeper green in colour, so that the eye was at once attracted to
these plants. The data obtained in regard to this partially glabrous
condition pointed to its relation to the completely hairy state being that
of dominant to recessive, but the records were neither large enough nor
derived from a sufficient number of generations to furnish conclusive
evidence on this point, nor did they serve to show whether or not this
character is constant and breeds true. In order to obtain decisive proof
as to the behaviour of this form the experiments were continued through
several further generations. In passing it is worth mention that a
further object of the investigation was to ascertain whether the two
distinct abnormalities, peloria and heptandry, exhibited any phenomenon
of the nature of linkage, or whether each recessive character was inherited
without relation to the other. The later work showed that these two
modifications are, in fact, inherited quite independently and that in
^ The New Phytologist, Vol. x. 1911, p. 60.

216 Smooth-steninied Form of Foxglove

peloria, as in heptandry, the degree in which the abnormality is mani-
fested, as indicated by the size of the peloric "flower" has some relation
to the order of the axis, whether main or lateral, on which it is borne,
as is the case with numerous other abnormalities which exhibit grada-
tions. A further remarkable peculiarity observed in all the members
(40) of one heptaudrous non-peloric family was the thickening of the
margins of the sepals, on which were borne bodies having the appearance
of rudimentary ovules. Unfortunately no offspring were raised from
these particular plants, and as the character never reappeared in any
later crop there is npthing to add to the bare record.

A serious difficulty all through the work was the almost unavoidable
contamination of the whole available ground with a large quantity of
self-sown seed. In spite of the precautions taken to meet this contin-
gency the appearance of an occasional plant here and there gave ground
for suspicion that it was a " stray" and sufficed to render the evidence not
wholly beyond question. Eventually, however, by breeding exclusively
from white-flowered peloric, heptandrous (i.e. triple recessive)^ parents
whereby the detection of " strays " red-flowered or normal was at least
ensured, unimpeachable records were obtained. These results have now
fully established the fact that the smooth-stemmed condition in Digitalis
purpurea is a well-defined character^ and that pure-bred individuals
exhibiting this character, but otherwise identical with the completely
hairy plant, breed strictly true. We have in fact in the partially glabrous
D. purpurea a form as definite as the partially glabrous (half-hoary)
MattJiiola incana which I have described elsewhere*, to which indeed it
appears to be somewhat analogous. For the half-hoary Stock similarly
has a glabrous stem though the type plant is tomentose throughout.

1 The recessive nature of peloria was established by Keeble, Pellew and Jones (see
The Neiv Phytologist, Vol. ix. 1910). In my earlier paper (loc. cit.) I showed that
heptandry is similarly recessive to the normal, and the same result has been obtained by
Tine Tammes and Shull (see Shull, ZeiUchr. f. induk. Ahstamvnings- u. Vererhungslehre,
Bd. VI. 1912). The white-flowered plants employed were of the pure white form with
greenish spots which behaves as a recessive to the coloured type.

'■^ Although I have consistently )ef erred for convenience only to the stem character, it will
be understood that this feature is always associated with the smoother leaf surface though
the difference in the leaves is less striking to the eye. And here I may add that the extent of
the stem which is smooth is not a definite proportion of the total length but is determined by
the amount of vegetative growth which takes place before the reproductive stage approaches.
A practical outcome resulting from this variability is that in many herbarium specimens
barely sufficient of the stem is preserved to permit of its character being certainly determined.

» See Eeports of the Evolution Committee to the Royal Society, Hep. 1, 1902, p. 33.
For a fuller account later see J. oj Genetics, Vol. v. No. 3, 1916, p. 145.

Edith K. Saunders 217

The leaves which succeed the cotyledons are also glabrous except for one
or two minute marginal hairs near the apex. But as the plant develops
the successive leaves attain a greater degree of hoariness until the lower
surface becomes covered nearly if not quite as densely as in the fully
hoary type plant, though on the upper surface the covering of hairs
remains throughout appreciably thinner. As we ascend from the vege-
tative region of the stem to the bractless racemes hairs usually gradually
extend down on to the axis from the pedicels which become distinctly
hoary on their abaxial surface. In the smooth-stemmed Foxglove, as
stated above, the stem is glabrous, deep green and of a polished appear-
ance from the radical rosette upwards. The leaves, though hairy on both
surfaces, are distinctly less so on the upper surface than in the hairy
form, and are deeper green in colour. The midrib on the under side of
the last-formed leaves of the rosette appears as a smooth green demar-
cation line halving the greyer lamina and petiole, whereas, in the hairy
form, the grey surface appears unbroken. This feature serves as a
useful means of identification in young plants which have not yet de-
veloped the flowering axis. Proceeding up the stem we find that an
overflow as it were of the hairs from the leaf may extend down from the
point of insertion as a hairy streak or band. As we approach the flower
spike these decurrent bands from the uppermost leaves and bracts
coalesce more or less completely so that the axis becomes as completely
downy as in the hairy form. Thus in both Stock and Foxglove we have
the same feature, a transition from the glabrous condition characteristic
of the vegetative region to a more hairy condition in the inflorescence,
the change being much more marked however in the Foxglove.

As appeared probable from the results obtained incidentally in the
course of the earlier experiments on heptandry the smooth-stemmed
form in D. purpurea is found to behave as a simple dominant. The
statistical evidence proving this to be the case is given in the accom-
panying tables where it will be seen

(1) that both forms, when pure, breed true ;

(2) that the F^ generation from a mating between the two is all

smooth-stemmed ;

(3) that F^ plants yield the simple ratio of 3 smooth to 1 hairy

when self-fertilised, and 1 smooth to 1 hairy when crossed
back with the hairy form.

These results are in marked contrast to the complex relations which
are found to hold in the Stock, a full account of which it is hoped to

218

Smooth-stemmed Form of Foxglove

complete shortly. For the moment the general statement that the half-
hoary condition in the Stock is dependent upon several factors, and that
it behaves as a recessive to the fully hoary type will suffice to show that
the underlying factorial relationships are here quite different notwith-
standing the apparent similarity of the two cases.

That the smooth-stemmed plant represents a distinct form of pur-
purea sharply separated from the accepted type and not a fluctuating
condition of the species does not appear, so far as I can discover, to have
been recognised by systematists. I have searched in vain in both British

TABLE I.

Results obtained when the two forms o/" Digitalis purpurea ivere
self- or inter-hred.

Reference to these two forms by name being more convenient than by a lengthy
description I have designated the hairy-stemmed form pabescens, the smooth-stemmed
midicaulis.

Experiment 1 shewing that pubescens
(recessive) breeds true.

Offspring

Experiment 2 shewing that pure bred
nudicaulis (dominant) breeds true.
Offspring

Parentage

No. of
family

pubescens

nudicauKu

Parentage

No. of
family

pubescens

nudicaulis

pubescens x self

1

208

None

nudicaulis x self

1

None

303

or fertilised

2

115

or fertilised

2

,,

263

inter se

3

77

inter se

3

>)

162

4

75

4

,,

49

5

55

6

47

Total

777

7

46

8

37

9

32

10

30

11

28

12

26

13

12

14

87

15

47

16

43

17

41

18

39

19

29

20

21

21

21

22

19

23

19

24

12

Total

1166

None

Edith R. Saunders

219

TABLE II.

Results obtained when crossbreds were self-fertilised or bred
back with the hairy-stemmed form.

Experiments* shewing that nudicaulis
crossbreds x self or fertilised inter se yield
both forms in the simple Mendelian ratio of
3 nudicaulis : 1 pubescens.

Offspring

Experiment 4 * shewing that nudicaulis
crossbreds when crossed hAckwith pubescens,
in accordance with expectation yield the two
forms in the ratio of 1 nudicaulis : 1 pubes-
cens.

Parentage

No. of
family

1
2

imbescens
15
16

nudicaulis
55
37

Parentage
nudicaulis cross

No. of
family

- 1

Offspring

nwdjca/zi/s cross-
breds X self or

piibescens
388

nudicaulis
394

fertilised inter se

3

24

65

breds x pubescens 2

377

343

4

36

104

3

205

228

5

8

42

4

15

18

6

28

75

5

74

62

7

21

69

6

83

88

8

36

94

7

116

146

9

28

78

8

145

173

10

47

128

9

40

38

11

27

73

10

145

158

12

10

64

11

177

156

13

24

53

12

140

144

14

50

191

13

55

57

15

18

75

14

70

48

16

10

25

15

41

33.

17

14

42

16

68

87

18

1

7

17

31

31

19

28

108

18

74

60

20

1

6

19

144

141

21

15

57

20

103

116

22

6

22

21

11

12

23
24
25

6

5

53

23

13

130

22

13

9

Totals

2515

2542

Expectation

2528-5

2528-5

Totals

...

527

1636

Expectation

541

1622

* Beciprocal crosses were made in each case.

and Continental Floras and in numerous works devoted to detailed
descriptions of British Wild Flowers and Garden Plants for any hint as
to the definite nature of the variation above mentioned. In descriptions
containing direct reference to this character of the stem as apart from
the plant as a whole, and such are the exception rather than the rule, it
is usually stated simply to be tomentose or pubescent, or at most with
the amplification "especially above," but without in such case precise
Journ. of Gen. vii 16

2*20 Smooth-stemmed Form of Foxglove

reference to any sharp contrast presented by the lower region. This is
the more surprising as the two forms very frequently occur together in
copses, woods and hedgebanks through the length and breadth of England.
I had hoped to be able to give a more or less full list of the areas in this
country where the smooth-stemmed form occurs mixed with the hairy,
and where it appears not to accompany it, but I have been obliged to
postpone further investigation in this direction owing to the difficulties
which have arisen in consequence of the war. That it is widely dis-
tributed as a wild or naturalised plant there is no doubt. I have myself
observed it as far north and south as Northumberland and the New
Forest ; in the Midlands ; and as far east and west as Kent and Surrey
and Somersetshire and Carnarvonshire. Also wherever I have chanced to
examine plants in gardens I have invariably found both forms growing
mixed together. In herbarium material, whether British or Continental,
I find purpurea represented by either form indifferently and without dis-
tinction. In the Cambridge collections, for example, specimens from
Surrey, Gloucestershire, Wales and Scotland, and from the Black Forest
and the Vosges appear to be smooth S while other specimens from Scot-
land and Germany, from Gibraltar and Castile, Portugal, France and the
English counties Shropshire, Sussex and Worcestershire (Malvern Hills)^
are hairy. Again in the Kew collections, out of some score of specimens
examined, about two-thirds were found to be nudicaulis and one-third
only of the recognised type. There appear, however, to be localities in
England, and no doubt abroad, where pubescens occurs alone — nudicaulis
alone I have never met'*. For the following information on this point
I am indebted to the kindness of friends who were good enough to make
observations in response to my enquiries. Dr J. C. Willis tells me that
of some hundreds of plants examined on Worcestershire and Hereford-
shire Beacons (Malvern Hills) he and his daughter noticed only pubescens.
Statements to the same effect were sent to me by Dr O. Withers in
regard to the neighbourhood of Sidmouth and to his own garden there
where plants had appeared spontaneously from self-sown "wild" seed; by

1 Although in some of these specimens only the upper part of the stem is preserved,
I have no doubt as to the identification being correct in these cases.

2 On the Lickey Hills on the other hand I found both forms.

3 In the only two districts where I attempted an actual count over a small area — a copse
at Holmwood (Surrey) and a cleared piece of land near Stocksfield (Northumberland) —
I found in the former case a slight, and in the latter a considerable predominance of the
smooth form. Information kindly sent to me by a friend in regard to a wood in Somerset-
shire on the other hand was to the effect tbat pubescens was much more abundant than
nudicaulis.

1

Edith R. Saunders 221

Dr D. H. Scott regarding a wood near Oakley (Hampshire), although
both forms were to be found in his garden about a mile away, here
however commercial seed had been introduced^ ; by other correspondents
in regard to the plants which occur abundantly on shingle in the neigh-
bourhood of Dungeness and in woods and coppices around Peasmarsh
(Sussex).

These facts naturally give rise to the question as to when and where
the smooth-stemmed form first appeared. Did it arise spontaneously as
a wild plant or did it originate in cultivation and then become generally
distributed as a " garden escape." Or is the premise here involved not
really established and should we more properly enquire whether indeed
pubescens, as seems to be always assumed, or nudicaulis represents the
original type ? Whichever be the view adopted there remains to be con-
sidered the further question whether the one form arose from the other
by direct mutation, or as the result of hybridisation either with a hairy
species on the second supposition, or with a glabrous or partially glabrous
one on the first. If a direct mutation has taken place, then, according to
the accepted view we should have the case of a dominant mutant arising
from a recessive type. As a comparable case in another member of the
Scrophulareaceae we may cite Linaria alpina where the self-coloured
form concolor, though dominant, is regarded as a variety., the recessive
form with the orange patch on the lower lip as the type^ Here as in
Digitalis the two forms are distinguished by a solitary characteristic, in
every other respect they are identical. Such instances of mutation — if
mutation it be, and in Linaria there seems no ground for regarding it
otherwise — from a recessive to a dominant form, when only one factor
appears to be involved, are exceptional, and it is obvious that on the "pre-
sence and absence " theory of factors they present a certain difficulty. But
the alternative hypothesis which supposes that the smooth-stenimed Fox-
glove or the self-coloured Linaria is derived from a cross with another
species is also not free from difficulty. On this supposition we should
look in these forms for larger or smaller differences from the parent type
in a considerable number of characters such as are enumerated by Neilson
Jones^ as characterising his artificially raised (and hence authentic)

1 I am indebted to Dr Scott for the f urtlier information that this was essentially an oak
and hazel wood, and that the soil was "clay with flints" resting upon qhalk. So far how-
ever the results of investigation do not seem to indicate that the appearance of nudicaulis
is conditioned by the character of the soil.

2 See The Neio Phytologist, Vol. xi. 1912, p. 1G7.

3 " Species Hybrids of Digitalis," J. of Genetics, Vol. ii. No. 2, 1912.

222 Smooth-stemmed Form of Foxglove

'purpurea-grandiflora (= -ambigua Murr.) hybrids which were found to
differ to a greater or less extent from the parent species in every par-
ticular: by Henslow^ in his detailed comparison of a spontaneously
occurring garden hybrid with its supposed parents purpurea and lutea :
and again in the still earlier account by Roth^ of the wild form media
presumed to be a hybrid between ambigua Murr. and lutea. In the case
of Digitalis certainly we should look too for some trace of that sterility
which is so marked a feature of indubitable species hybrids in this
genus^*. But we do not find either the one or the other. For except as
regards surface character the two forms nudicaidis and pubescens are
precisely similar, and both set seed abundantly. Moreover it is to be
noted that such spontaneously occurring forms as are judged to be of
hybrid origin exist for the most part in small numbers, only a single plant
perhaps being found wild or in a garden from time to time^ whereas, as
stated above, purpurea nudicaulis is not only of very general occurrence,
but where it occurs it is abundant. Furthermore, reference to other
species of Digitalis makes it evident that the nature of the differentiating
character in the present case cannot be taken as in itself necessarily

1 Tram. Camb. Phil. 'Soc, Vol. iv. Part 2, 1831, p. 257.

2 Cat. Bot. II. 1800.

* Among mauy references to this fact since the early work of Koelreuter and Gartner
may be mentioned Focke's account of the following hybrids — purpurea-lutea, ambigua-
"purpurea, ambigua-nbscura, ambigua-lutea, ambigaa-lanata, ambigua-laevigata, laevigata-
lanata [Die Pjianzen-Mischliiige, 1881). In each case greateror less difficulty was experienced
■ in obtaining the hybrid artificially, and the hybrids themselves proved in most cases to be
completely sterile. Of these various forms purpuri'a-lutea has been recorded from time to
time by various observers in the wild state. It was noticed by A. Saint-Hilaire and
de Salvert growing mixed with purpurea and lutea in a valley in the Auvergne district in
1808, and for several consecutive years the capsules were examined for seed, but each
season they were found to be shrivelled and to contain only aborted ovules (" Observations
sur la St^rilite des Hybrides," Mem Soc. d'hist. nat. Paris, i. 1823, p. 373). It is stated by
Wilson who raised a number of hybrids between these two species that the stamens were
often wanting, and that when present the pollen was found to be bad and he obtained no
seed (see Report of the third International Conference on Genetics, Roy. Hort. Soc, 1906).

■* Among such cases may be instanced the single plant identified as D. piirpurascens
Roth, found by Le Jolis growing in the environs of Cherbourg amid plants of D. purpurea
(Ann. Sci. Nat., 3rd Ser. T. vii. 1847, p. 220) : the single specimen of D. longiflora noticed
by Lejeune near Verviers and regarded by him as answering closely to Koelreuter's hybrida
(Revue de la Flore des Environs de Spa, 1824, p. 126) : the single purpurea-lutea hybrid
described by Henslow which appeared in his garden among plants of the two parent species
(loc. cit.) : and other similar records to be found scattered through the literature of the
subject which emphasise the rarity of occurrence of these natural hybrids, their variability
of form, and the fact that when they arise they are often little durable (see Lamarck and
De CandoUe, Flore Fran^aise, Vol. vi. 1815, p. 412, and Vaucher, Histoire Physiologique des
Phintes d'Europe, T. iii. 1841, p. 520).

Edith E. Saunders '223

pointing to a hybrid origin. For we find within the genus, besides the
group of types closelyallied to purpurea andlikejju6esce/7.s' having the stem
pubescent or tomentose throughout, as e.g. Thapsi L., tomentosa Link and
Hoffmgg., Mariana Boiss., nevadensis Kze. and duhia Rodr.\ and those on
the other hand which like laevigata W. and K., laciniata Lindl., obscura L.
and ferruginea L. are smooth-stemmed throughout, a considerable
number of forms which exhibit as a recognised normal feature a similar
transition from a glabrous condition in the vegetative to a hairy condition
in the flowering region of the stem. Well-marked examples of this class
are D. lanata Ehrh., a Hungarian species, and the Portuguese form
D. miniana discovered by Sampaio and described by him in 1905, in both
of which the transition in surface character is definitely recognised as a
diagnostic character. Among cases less pronounced or possibly only less
clearly recognised we may probably include ambigua Murr., sibirica Lindl,
leucophaea Sibth., parvijiora Jacq.- and perhaps orientalis Lam. and
lutea L.^; also purpurascens Roth, considered by some as a hybrid form
and fuscencens W. and K., the last-mentioned a rare Hungarian plant

1 All these types appear to be closely related to purpurea. Beautiful illustrations of
tomentosa and Thapsi are given in Flore Portugaise by Link and Hoffmansegg, where
tomentosa and purpurea are compared in detail. According to Brotero [Phytographie, T. ii.
1827, tab. 149, p. 159) tomentosa is not to be regarded as distinct from puipurea but merely
as a southern race of that species to which it is linked by intermediate forms. Thapsi is
described by Lindley {loc. cit.) as also very similar to purpurea, and by Koelreuter (Acta
Acad. Petropol., 1777) is regarded in the same light as tomentosa. G. W. F. Meyer takes
the same view {Ghloris Hanoverana, 1836, p. 323). Mariana is another form, according
to Boissier {Voy. Bot. T. n) closely akin to purpurea, while dubia is considered by
Rodriguez {Plantas vasculares de Menorca, 1874) to be intermediate between Thapsi and
minor, the flowers of the latter species being described by Linnaeus as again very similar .
to those oi purpurea.

^ From systematic descriptions of these species where little importance attaches to the
nature of the stem surface as a diagnostic character it is less easy to gather the precise
degree of difference between the upper and lower region. Though no specific statement
may be met with in the text reference to the illustration, where such is given, will some-
times clearly show that the plant exhibited this transition character, as is seen, for example,
in the case of D. ochroleuca described and figured by Jacquin in Fl. Aust. i. p. 36, pi. 58,
1773.

3 Though lutea is generally spoken of as glabrous Lindley describes the stem as very
glabrous or pubescent, and Henslow also draws attention to the possibility that absence of
hairs in this species is not a fixed character. Specimens from near Siena in the Bunbury
Collection (Cambridge) show numerous hairs on the margins and on the veins of the under
surface of the leaves, and on bracts and sepals, but the stem appears to be hairless. Whether
the appearance of hairs on the stem in this case is a fluctuating character and varies with
conditions or whether there are two distinct and constant forms I have not sufficient data
to determine. It is possible that the same question may arise in regard to orientalis and
perhaps also with laevigata as well.

15—3

224 Smooth-stemmed Form of Foxglove

universally regarded as a hybrid, but whether on other grounds than its
intermediate character does not appear. It has not yet apparently been
artificially bred since statements as to its parentage commonly indicate
uncertainty as to whether it should be regarded as derived from a cross
between amhigua and lanata or ambigua and laevigata.

D. lanata was originally described by Ehrhart' as having the stem
glabrous, the raceme woolly. The figure and description of this plant
in Curtis's Botanical Magazine" emphasises the same character, the stem
being described as " smooth at the lower part, woolly above." So also
does the account given by Waldstein and Kitaibel"' of the plant referred
by them to this species. They describe the stem as naked but woolly
with white hairs among the flowers, this feature being also clearly seen
in the coloured plate which shows the stem destitute of hairs and
purplish-red in the lower portion, becoming green and covered with hairs
in the region of the inflorescence. Similarly Sampaio's description of
D. miniana as quoted by Coutinho states that the stem is everywhere
glabrous except the axis of the inflorescence which is more or less
tomentose*. Comparison of the descriptions by numerous writers of the
other forms enumerated above and examination of herbarium material
indicate that this same feature is characteristic of all of them in varying
degree. Now to several of these types, if not to all, systematists are
agreed in according specific rank, hence in the particular case of purpurea
the stem character cannot be held to afford ground for regarding nudi-
caulis as a crossbred. But there is yet another alternative supposition
which is not excluded and which appears to me worthy of consideration.
May it not be that nudicaulis is a precursor rather than a derivative of
pubescens ? As between such forms our conception of which is type and
which variety, in the absence of any historical record is largely deter-
mined by our knowledge of the relative numerical abundance of the two
forms and possibly may not invariably reflect the true genealogical order.
A mutation in this reverse direction, i.e. from nudicaulis to pubescens
would on the " presence and absence " view be simply the usual case of
variation by the loss or dropping out of a factor. Once in existence the
new recessive (on this supposition pubescens), in any area to which it
chanced to spread unaccompanied by nudicaulis, might well continue to

' Beitrdge vii. p. 152.

2 Vol. XXIX. PI. 1159.

3 Descript. et Icon. Plant, rar. Hungariae, i. p. 76, pi. 74.

* I have been unable to see or obtain Sampaio's original paper in A Revista, Porto, iii.
1905, but his account is quoted by Coutinho in the Bol. Soc. Broter., xxxii. p. 199, See
also Coutinho, A Flora de Portugal, 1913, p. 561.

Edith R. Saunders 225

remain pure for a considerable period, and be taken in such areas to
represent the type form of the species. It is possible that the Malvern
Hills afford a case in point, for according to Anne Pratt the Foxglove
suddenly appeared in abundance on one of these hills after, for some
reason, the soil had been turned up\ The same argument may be
advanced in the analogous case of Linaria alpina mentioned above, where
indeed a hybrid origin seems even more improbable.

How far a fuller knowledge of the geographical distribution of the
two D. purpurea forms will serve to throw light on their genealogical
relation remains to be seen. The home of the genus appears to be in
Central and Southern Europe, numerous species occurring in Germany,
France, Spain, Italy, Austria ; it is also represented by several forms in
Russia, Belgium, and Portugal, purpurea itself being found in all these
cuuntries, as well as in England, Denmark, Norway and Sweden where
it is the only recorded species. I have already drawn attention to the
wide distribution over Great Britain of purpurea nudicaulis in company
with purpurea pubescens. The further fact that herbarium material
examined at Kew was found to include specimens of nudicaulis as well
as pubescens both from Norway and Sweden, taken together with the
descriptions of the species in Neuman's Swedish Flora (stem cylindrical,
hairy above) and in Blytt's Norwegian Flora ([plant] covered with short
downy hairs especially above) renders it highly probable that nudicaulis
will very generally be found with pubescens in these countries, as in
England. If this should prove to be the case it strengthens still further
the argument against hybridisation as an explanation of the appearance
of nudicaulis. It may be urged no doubt that the absence of all mention
of glabrousness as a feature of the stem in the early descriptions of
D. purpurea, such as those of the older herbalists, renders the view here
suggested improbable. But seeing that neither in modern works do we
find any clear reference to the existence of two forms despite the abun-
dance and wide distribution of nudicaulis this objection may be put
aside. A similar omission is very general in the case of Linaria alpina ;
many botanical works dealing with the flora of regions where both forms
are to be found make no mention of the concolor form. And here
perhaps the matter must rest until further evidence throwing light on
the question is available.

^ The Flowering Plants and Ferns of Great Britain, Vol. iv.

226 Smooth-stemmed Form of Foxglove

Summary of Conclusions.

1. Digitalis purpurea occurs ^ under two distinct forms, the one
commonly accepted as the type with the stem grey and densely pubes-
cent throughout, and the leaves very hairy, here designated pubescens ;
the other, not apparently hitherto distinguished, with the stem green,
polished and smooth from the base to the flowering region where it
becomes pubescent, and leaves less hairy, designated nudicaulis.

2. The two forms are similar in every respect except as regards
surface character.

3. No difference in fertility was observed between the two forms
which set seed abundantly.

4. The distinguishing feature of purpurea nudicaulis is a character
common to several other species within the genus.

5. Both forms, when of pure parentage, breed true.

6. Nudicaulis behaves as a Mendelian dominant when crossed
with pubescens and the erossbreds yield the monohybrid ratio 3 nudi-
caulis : 1 pubescens when self-fertilised and 1 nudicaulis : 1 pubescent
when crossed back with the recessive pubescens.

7. The facts stated under 2, 3, 4, 5 and 6 do not support the idea
that nudicaulis is of hybrid origin.

8. The alternative hypothesis of the origin of the one from the
other by direct mutation raises the question as to which represents the
original type, unless we postulate parallel development from a common
ancestor.

9. In the absence of conclusive historical or geographical evidence
to the contrary, and in view of the fact that there is a certain difficulty
in supposing the origin of a dominant mutant form a recessive type, the
view that nudicaulis may be the earlier form and pubescens the derivative
seems at least worthy of consideration.

10. The same argument may be advanced in the case of Linaria
alpina where also, according to the accepted view, we have in the recessive
spotted form the type and in the dominant concolor form the variety.

Edith R. Saunders 227

And further in regard to certain abnormal features : —

11. That the two modifications of the corolla, peloria and heptandry,
both recessive to the normal, are inherited quite independently.

12. That as a rare occurrence the margins of the sepals may be
thickened and bear structures having the appearance of rudimentary
ovules.

The expenses incun'ed in the course of this investigation have been
defrayed in part by grants from the Royal Society and from the British
Association for the Advancement of Science.

After the foregoing account had been written I received from Professor
Coutinho, to whom I here wish to tender my most cordial thanks, in
response to my enquiry, a specimen of D. miniana obtained from the same
region as Sampaio's original plant (Duriminia, Serro de Castro, Laboreiro,
near Alcebaga, about lat. 42° N.), and also further interesting information
regarding the occurrence of this form. Professor Coutinho writes that
D. iiiiniana is a rare plant, occurring within a very restricted area mixed
ivith purpurea ; that it was found again last year by one of his assistants
in the original locality, the specimen sent being one of those obtained
on this occasion. The appearance of the specimen, taken together with
Sampaio's account, and the further details supplied by Professor Coutinho,
make clear the close affinity of this form with purpurea. In fact whether
it is really distinct from that species seems to me open to question seeing
that the smooth stem emphasised as the distinctive character of miniana
is a constant character of purpurea nudicaulis. The dimensions given
for the stem (2 cm. in diameter at the base, 17 dcm. in height), the deeper
colour, smaller size (10—18 mm. long) and narrower form of the corolla
which is only slightly ventricose may possibly be fluctuating characters,
and in that case may be discovered associated with the pubescens habit,
although having hitherto escaped detection. The description of the
leaves of miniana would be quite applicable to a plant of purpurea.
Although the length of the pedicels in purpurea is generally not more
than equal to that of the bracts whereas in this particular specimen of
miniana they far exceed the bract-length, this appears to be a decidedly
variable character since in his description of miniana Sampaio states the
precise contrary, and in purpurea I have occasionally found a pedicel
exceed the bract by perhaps 2 mm. or more. Beyond the corolla characters
therefore there seems to be no very sharp distinction between purpurea
nudicaulis and miniana, and it remains to be seen whether the latter

228 Smooth-stemmed Form, of Foxglove

form will remain constant under different conditions, a point which I
hope now to be able to test on the material received from Professor
Coutinho. Hybridisation as an explanation seems improbable since
D. Amandiana Samp., the only other native form with a glabrous habit,
is not mentioned by Sampaio as an inhabitant of this region but as
occurring in arid and stony places along a certain part of the course of
the Douro, and the Tua and elsewhere.

Volume VII AUGUST, 1918 No. 4

THE INHERITANCE OF TIGHT AND LOOSE
PALEAE IN AVENA NUDA CROSSES\

By A. St CLAIR CAPORN.

(With six text-figures.)

Experiments were conducted with three varieties of ordinary oats
with tight paleae, viz. : —

Thousand Dollar.

Ligowo.

Nubischer Schwarzer (Nubian Black).
As the F.2 generation of the cross with Nubischer Schwarzer has not
yet been harvested, being part of this year's crop, it may be necessary
to issue the results obtained from it in a supplementary account, though
reference to its earlier stages will be made in this paper.

A short description of each of the parents will indicate the nature
of the characters involved.

Thousand Dollar.

This is a typical tight grained oat of the spreading, or open, panicle
kind. The grains are of good size, fairly long, and firmly ensheathed
in the paleae. They do not, except when it is exceptionally prolonged
and violent, become naked on threshing. The inner paleae are thin
but stiff, the outer thick and curled round the edges of the inner. So
well developed is the intervascular sclerotic tissue that the fine longi-
tudinal ribs of the outer paleae appear only as faint markings, the
whole surface, save at the extreme tip, being perfectly smooth.

The spikelets are usually two-grained, though a panicle will often
contain a few which are 3-grained. Very rarely a plant will be found
which has two or three 4-grained spikelets. In this case the upper-
most grain is generally very small and all the peduncles are short and

1 The work herein described was taken up at the F-, stage. The data concerning the
parents and the Fi generation, however, were rather vague and scanty, and had to be
reinvestigated by me concurrently with the analysis of the F3 crop.

Joum. of Gen. vii ♦ 16

230 Ifiherltance of Tight and Loose Paleae in Oats

straight. It is important to note this condensation of the occasional
large spikelet in the tight oat as compared with the extreme laxity of
the Avena nuda type.

Ligowo.

In all its essentials this variety resembles Thousand Dollar. It
also has white, tight, awned outer paleae and spreading panicles. The
grain, however, is somewhat plumper and there seems a greater tendency
towards 3-grained spikelets, though no 4-grained have been noticed,
possibly because no large field crop has been available for inspection as
in the case of Thousand Dollar.

Fig. 1. Characteristic spikelet from' a "pure loose"
containing 8 flowers.

Nuhischer Schwarzer

is a shorter strawed oat than the two just described. It also differs

in being a black tartarian. Save for the size and shape of the grain all

other points are similar,

Avena nuda.

Two distinct features characterise the spreading panicles of this
species: —

{a) The long, dangling spikelets in which there may be as many as
nine flowers (cf. Fig. 1), and

(h) The readiness with which the grain drops out from between
the paleae when mature.

A. St Clair Caporn 231

The parental stock used showed impurity if the "nuda" type be
taken as one in which the paleae are all absolutely membranous, that
is to say, with no trace whatever of intervascular sclerosis. For some
plants had at least one obviously tight grain, often more, the pro-
portion of tights even rising to 40 "/^ in extreme cases, while others
would have a grain or two with paleae moderately stiffened in a
broadened midrib region.

One important character, however, is common to all these forms :
they never throvj offspring with all the grains tight. In this respect,
therefore, even if not in others also possibly concerned in the partial
tightening of the paleae, they may be regarded as constituting a pure
line.

In colour the parent plants were mixed greys and whites. Greys
are very easily overlooked in Avena nuda because colour is developed
only on the small inner paleae.

The ^1 Generation.

Spikelets.

The panicles of the F^ plants contained a varying proportion of
many-flowered spikelets of the " nuda " type. In the basal regions the
two-flowered spikelet predominated.

Paleae.

Every conceivable gradation from the "pure tight" of the tight
grained parents to the wholly membranous palea occurred.

On rubbing out, it was noticed that sometimes there was an excess
of tight grains, and sometimes not. Six separate panicles, which all
threw the same range and similar proportions of F^ types, and must
therefore have been Fi heterozygotes of the same factorial composition,
were constituted as follows

TABLE I.

Avena nuda 9 x Thousand Dollar ^ .

Panicle 1. Excess of loose grains. Tights about ^ of total

j> 2. ,, ,, ,, ), •§• ,1

„ 3. ,, tight ,, ,, i ,,

Ligowo 9 X Avena nuda $ .
Panicle 4. Excess of tight grains. Tights about f of total

)j &• )i »> >> >) 2 ' >>

,, 6. ,, loose ,, ,, 1 ,,

It was at first thought that the reason for this fluctuation might lie in
the fact that single panicles were examined instead of whole plants ; but

16—2

232 Inheritance of Tight and Loose Paleae m Oats

a more detailed analysis of the entire plants in this generation of the
Nubischer cross proved that, although there are cases in which the
exact nature can only be ascertained when the basal tillers are also
taken into consideration, the main panicle is generally indicative of
the whole plant. The necessity for inspecting the other panicles arises
chiefly when the excess of either tight or loose grains is very great.
In such cases all, or most, of the paleae of the preponderating type
may be concentrated into the largest panicle, and the opposite extremes
have to be searched for on the subsidiary j)anicles.

In the detailed analysis just referred to the panicles on each jPj
plant were taken separately, in the manner shown in Table II, and the
paleae borne on the different branches classified node by node, com-
mencing at the base. A " pure tight " palea has no paperiness, a "pure
loose " is entirely membranous, while an " intermediate " is a very
variable mixture of these two extremes (cf Fig. 2).

ABC

Fig. 2. Types of paleae found on F^ plants. A = pure tight; a grain
is shown in both dorsal and ventral view, the latter showing the
margins of the outer pale clasping the inner one. B = intermediate
forms. C = a pure loose palea. The sclerotised parts are repre-
sented black.

The percentage of " pure tights," it will be observed, ranges from
20-4% to 68-7 7 J that of the "pure looses" from 7-8% to 46-8%.
The rise of the one appears to be correlated by the fall of the other,
not only when different plants are compared, but also in different
zones in the same plant. In plant A, for example, the first nodes
of the four panicles bear 77 pure tight and 7 pure loose paleae out of a
total of 115, i.e. the relative proportions are as 67 7o '■ 6"! Vo-

In the case of the second nodes they are as 45'8 °/^
For the third nodes they are as ... 268 %

For the fourth they are as ... ... 5 °/^

And for the last they are as *^'^°lo

20-8 7,
39 7o-

45 7o.

44-473

A. St Clair Caporn

233

TABLE 11.

JVubischer Schwarzer x Avena nuda.
Analysis of paleae on 5 Fi plants.
[P. t. = pure tight. P.l. = pure loose. Int. = ± membranous.]
Distribution of paleae at

1st node

2nd node

3rd node

4th node

5th node

Plant Panicle P.t. Int. P.L P.t. Int. P.l. P.t. Int. P.l. P.t. Int. P.I. P.t. Int. P.l.

A 1 27 8 2 22 15 15 2 4 5 0 1 3 0 0 0

2 16 18 49 10 9037025024

3 23 51 15 30532131157

4 11 009 12 144. 2040161

Percentage
P.t. P.l.

D

Totals

77

31

7

55

40

25

11

14

16

1

10

9

2

13

12

45-2°/,

21-4°/,

1

4

17

12

5

9

20

2

5

7

0

5

8

0

3

8

2

27

4

2

13

11

0

5

10

11

1

1

12

2

7

19

3

4

7

6

5

17

10

5

10

11

5

5

6

0

5

18

4

4

2

1

0

8

10

1

3

7

0

2

8

0

2

14

Totals

39

30

21

23

45

40

13

28

36

6

13

34

2

17

59

20-4%

46-8%

1

7

10

7

2

13

12

1

8

10

0

5

7

0

1

10

2

26

9

4

13

9

7

7

8

5

0

6

7

0

2

4

3

11

14

9

7

15

6

3

8

5

0

3

2

0

3

5

4

6

8

0

10

6

2

9

4

1

5

3

0

2

3

7

5

10

6

2

3

8

5

10

1

5

2

3

6

0

4

6

6

13

2

2

5

12

10

1

5

3

6

5

15

0

0

8

7

7

2

0

11

11

5

8

4

4

3

3

8

3

2

7

8

24

1

0

11

7

2

8

6

8

4

1

2

4

1

2

Totals

104

47

24

62

81

49

47

44

36

20

29

47

9

16

49

36-4%

30-9°/o

> 1

30

0

0

18

0

0

5

2

2

1

1

0

1

1

0

2

10

8

2

12

12

2

3

4

3

0

2

6

0

2

0

3

12

0

0

6

0

0

10

1

1

1

1

2

0

2

4

Totals

52

8

2

36

12

2

18

7

6

2

4

8

1

5

10

63%

16-2%

; 1

11

8

0

16

5

0

10

3

0

2

0

0

2

2

2

2

3

0

0

4

1

0

4

1

0

1

1

0

0

2

0

3

19

1

5

5

1

1

1

1

0

1

1

1

tip broken off

Totals 38 9 5 25

1 15

12 4 2 68-7% 7-8°/o

234 Inheritance of Tight and Loose Paleae in Oats

%

Fig. 3. Diagrammatic representation of an Fi plant as a single panicle. For
purposes of the diagram the different types of paleae are indicated each in
a separate spikelet, though naturally all three kinds frequently recur in the
same spikelet.

^k =Pure tight type of paleae.

B^ - Intermediate type of paleae.

y^ - Pure loose type of paleae.

A. St Clair Caporn 235

This gradual diminution in the number of tight paleae, and increase
in the number of loose, towards the tip is a tendency noticeable in
every heterozygous panicle.

Colour.

Owing to the impure condition of the Nuda parent as regards
colour, the plants raised from the crosses with the white husked
varieties, Thousand Dollar and Ligowo, were a mixture of greys and
whites. The presence of greys, however, demonstrated the dominance
of the grey colour whenever the cross grey x white actually occurred.

In the case of Nubischer Schwarzer x Avena nuda the F^ colour
was a bright brown sometimes overlaid with a faint greyish flush.
Here the variability of the i\ may also be due to the fact — established
in another experiment — that the black parent is really made up of at
least three different kinds of blacks, represented zygotically by the
formulae

BBB'B' GG

BBB'B'gg

BB b'b'gg

where B and B' are factors for blackness, and G for grey colour.

The F^ and F^ Generations.

Paleae.

No meticulous examination of the pales in the F^ generation was
made. The mixture of types looked so bewildering that there seemed
no alternative but to rub out the panicles and note the actual propor-
tion of tight and loose grains. This w^as done.

TABLE III.

1 Thousand Dollar <? x Avena nuda ? 213 panicles gave 64 with all grains tight

2 Avena nuda c? x Ligowo ? 172 ,, ,, 49 ,, ,,

3 Reciprocal Cross 133 ,, „ 31 ,, ,,

Grand Total ... 518 „ „ 144 ,, ,,

Here appeared to be some evidence of a 3 : 1 segregation, though
the approximation is rough, expectation on the ratio mentioned being
1 29 "5 instead of 144. When, however, it is remembered that only
single panicles were examined, that these were occasionally damaged
at the tips, or had lost a few spikelets elsewhere, and that although,
as was subsequently proved in the case of pure tight-grained forms,
at any rate, the single panicle is indicative of the whole plant about

236 Inheritance of Tight mid Loose Paleae in Oats

85 times in every 100, the possibility of the number 144 being an
abnormally augmented figure is not difficult to conceive.

The likelihood of a heterozygous panicle being taken to indicate a
homozygous plant is well illustrated in Table II, plant D, panicle 1.
If such a panicle were only slightly damaged, the chances are all in
favour of the two loose grains with membranous paleae being the first
to break off, while it is also possible that the four paleae classified as
" intermediate " might be such near approaches to the pure tight form
as to be separable from the grains only by extra hard rubbing.

A random selection of 26 taken out of the 144 apparently " pure
tights" yielded an F^ generation fairly representative of the whole
group. The plants of 22 rows were uniformly tight grained on every
panicle. The remaining 4 rows showed distinct splitting. Twenty-
two out of twenty-six is equivalent to 122 "pure tight" out of 144
tentatively considered pure. This is a much better agreement with
the 1295 expected on a 1 to 3 basis, the corrected result for the F.^
generation reading : —

Pure loose and
lieterozygotes Pure tight

Observed 396 122

Expected 388-5 129-5

More substantial evidence of this agreement will now be offered in
connection with the detailed classification of other F.^ results.

A large number of sowings was made from F^, heads having a
mixed character. Double rows were planted, so that the tight grains
in each sowing might be placed in one section, and the loose in the
other. This was done in case the nature of the husk should in any
way be related to the zygotic constitution of the embryo, and hence
have some bearing on the segregation, though it was hardly expected
that maternal characters would so affect it. As a matter of fact they
did not. The double rows, however, if for no other purpose, were very
useful in that they enabled one to obtain a few more F^ plants.

In order to determine just which among these ^3 rows were showing
splitting after the manner of offspring of F^ plants, and could therefore
be regarded as an exact repetition ^ of an F^ generation derived from
actual ^1 individuals, the seed from half a dozen F^ panicles which had
been saved over was sown at the same time. These F^ panicles had
shown on rubbing out a proportion of "pure tight" grains varying
from one-fifth to three-fifths of the total. Allowing for a reasonable

' In regard to the throwing of one-quarter pure tights, that is to say.

A. St Clair Caporn 237

amount of variation in the relative numbers of the different con-
stituents, the six F^_ rows were essentially alike. All included "pure
tights," tight-containing types similar to the F^ parents, and one or
more of the forms such as have been described as occurring in the
original samples of Avena nuda.

It was found that practically all^ the plants of mixed type picked out
of the F^ generation gave rise to F^ rows of the kind just described, no
matter what the relative proportions of tight and loose grains. Some
even among the panicles sown as " pure tight " and " pure loose" behaved
in the same way, these, as has been explained before, being obviously
either damaged panicles or cases in which the preponderance of one
type of grain was so great as to render classification uncertain unless
the whole plant were examined. The number of these imagined
" pure " types which split in this way was small, and it has already been
shown (p. 236) that in the case of the pure tights, in which the observed
figures were too high, the excess over expectation was due simply to
these mistakenly placed heterozygotes.

According to the nature of the paleae the progeny thrown by each
of the similarly segregating ^3 rows were placed in the following
groups.

I. Pure tights. — All the grains on the plant enclosed in tough,
wholly sclerotised paleae (cf. the respective tight parents).

II. Tight-containers'^. — One or more " pure tight " paleae, the rest
varied.

III. Hardbacks. — No " pure tight " pales. Pales partly mem-
branous, partly hardened, the hard portion varying from a slightly
thickened midrib to the stiffening of nearly the whole palea.

IV. Penulti-looses. — As in III, but the hardening is never found
above the lowest palea in any of the spikelets. These verge very closely
on the

V. Pure looses. — All paleae absolutely membranous.

119 sowings gave rows containing 2445 plants and exhibiting the
splitting just described. The following table shows the distribution of
the various forms.

1 Exceptions are the "tight-containers" dealt with subsequently. See Table IV et seq.

■'' To avoid possible misapprehension it may be stated that this term is devoid of any
implication as regards the genetic properties of the plants to which it applies. Most
" tight-containers " throw " pure tight " plants among their offspring, but there are some
which do not throw any (cf. p. 239).

238 Inheritance of Tight and Loose Paleae in Oats

1 Thoiisand Dollar i x Avena nuda ?

2 Avena nuda <? x Ligowo ?

3 Ligowo i X Avena nuda 9

Grand Total

TABLE IV

Total
F:i plants

Pure
tights

Tight-
containers

Hard-
backs

Penulti-
looses

Pure
looses

? 1133

283

610

161

61

18

657

163

341

113

24

16

655

164

359

95

25

12

2445

610

1310

369

110

46

Fig. 4. Two characteristic spikelets from "tight-containers." Sclerotised parts

represented black.

I
(I

Fig. 5. Characteristic spikelet from a " hardback.
Sclerotised parts represented black.

Fig. 6. Characteristic spikelet from a " penulti-
loose." Sclerotised parts represented black.

A. St Clair Caporn 239

The first thing to notice is that the pure tights make up almost
exactly one-quarter of the total. The agreement with expectation is
remarkably good.

The tight-containers call for closer study, as their group is really a
mixture of two. These groups can only be defined and separated by
further breeding. For whereas the one type always throws pure tights,
the other, which is outwardly indistinguishable from it, never does so.
In some cases as many as 35 — 65 plants and more have been raised
from tight-containers of this second kind without any pure tights
appearing among them. It was also found that out of a fair random
sample of 118 tight-containers 7 behaved in this way. 7 out of 118 is
equal to 78 out of 1310; so that the results in Table IV may now
be summarised as follows :

78 Tight-containers not
throwing pure tights.
610 Pure tights : 1232 Tight-containers : i 369 Hardbacks.

throwing pure tights ) HO Penulti-looses.
46 Pure looses.

'603

This definitely establishes the 1:2:1 ratio, or, in other words, the
determination of complete tightness by a single independent factor.

But what is it that causes the heterogeneity of the last terra in
the ratio, the 603 plants which, while so obviously of different sorts,
have the one common property of being unable to give rise to pure
tights ? The 46 pure looses compose roughly one-sixteenth of the
group. Added to the penulti-looses they form one-quarter. These
proportions are significant : plainly other factors must be coming into
play.

At the present stage of the investigation, mainly through an in-
sufficiency of available experimental evidence, it is not possible to give
the exact number and mechanism of these factors (whether they are
complementary or cumulative, or even inhibitory, for example), but it
seems tolerably certain they are there, and that only through them
can the very variable constitution of the ^i individual^ (see Table II)
be accounted for.

Let us suppose, in order to illustrate this view, that there are
3 factors :

X = a factor capable of rendering all the paleae on the plant pure
tight.

240 Inheritance of Tight and Loose Paleae in Oats

Y = a fjictor capable of rendering some of the paleae on the plant
pure tight.

Z = a factor capable of rendering some of the paleae on the plant
more or less sclerotised but never wholly tight.

It follows that all pure tight forms must be homozygous for X, no
matter whether they contain Y or Z, separately or together, or not.
Similarly all plants which never throw pure tights, i.e. all " nuda "
forms, must be recessive for X. They, too, may or may not contain
X and Y. Now let any two individuals of these two classes be crossed.
It is evident that quite a number of F^ combinations may result :

XYZ • xYZ\

Xyz • xyz
XYz • xyz
XYZ xyz }

are a few.

One feature will alone be constant : the heterozygosity of X. All,
therefore, will throw one-quarter pure tights in the F^ generation.
This is consonant with our proven results. Now if in the F^ combi-
nation the presence of one dose of X determine that, say, x °/^ of the
paleae shall be pure tight, then it is not unnatural to infer that the
remainder will be variably tightened and hardbacked according as
Y and Z are present both homozygously, both heterozygously, one
homozygously and one heterozygously, and so on, or both absent.

It should be added that the functions assigned to the factors Y and Z
are, of course, merely hypothetical. Just as possibly either or both
may have some negative, inhibiting action, thus lowering the basic x °/^
of pure tight grains in the F^ plant (cf Plant B in Table II), while
quite possibly they may govern the number of grains per spikelet and
thus have an indirect influence on the hardening of the paleae. For
it has been noted that progressive tightening of the paleae always
reduces the number of flowers in a spikelet, until, in the pure tight
forms, the repulsion between the many-flowered habit and tightness of
paleae is complete.

Some further light is cast on the complex nature of the various
" nuda " types in Table V ; but admittedly it is only a glimmer. A few
interesting points, however, stand out. The results of two seasons'
plantings are contained in the table, as, owing to the tediousness of
the classification, with its careful scrutiny of every palea, only a limited
number could be undertaken at one time. In 1916 the parents were
simply single panicles rubbed out and called " various nuda forms " ;

A. St Clair Caporn

241

D

TABLE V.

Profreny

Parental

Tight-

Hard-

Penulti-

Pure

Nature of

Number

containers

backs

looses

looses

Parent

5/8

2

6

11

1

-—

5/25

7

13

1

1

Tight-container

9/3

2

17

1

1

—

5 A 13

4

1

5

2

—

5A/7

4

5

2

2

—

5Ain

5

5

2

1

Tight-container

5/32/1

12

12

7

6

Hardback

9/5/2

1

27

11

7

Hardback

9/5/3

11

35

28

15

Hardback

9/5/4

39

23

17

5

Tight-container

Totals

87

144

85

41

5J/1

12

6

.

1

9/3/2

61

6

—

5

Tight-container

Totals

73

12

—

6

5/1

6

8

2

5/2

6

9

1

—

—

6/3

3

16

6

—

9/5

6

12

2

—

—

9/7

6

9

6

—

—

9/5/1

6

21

2

—

Penulti-loose

9/5/5

22

32

8

—

Tight-container

Totals

55

107

27

—

5/4

16

6

.

5/23

32

3

—

—

Tight-container

5/26

42

4

—

—

Tight-container

9/4

7

9

—

—

—

9/10

3

12

—

—

—

9/43

14

4

—

—

Tight-container

5.-1/19

16

1

—

—

Tight-container

9/3/1

33

28

—

—

Tight-container

9/4/1

54

6

—

—

Tight-container

9/4/2

34

5

—

—

Tight-container

Totals

251

78

—

—

5/5

—

8

10

14

5/6

— "

9

5

24

—

9/6

—

3

1

8

—

5AI2

—

5

11

7

5AI8

—

4

3

6

—

Totals

—

29

30

59

—

5/8/1

—

—

9

30

Pure loose

242 Inheritance of Tight and Loose Paleae in Oats

but in 1917, when whole plants were used, the exact type was always
noted. Hence the indication in the table of only some of the parents.

Groupings have been made ; but they are not meant to show the
identity of the members in more than one broad, fundamental character.
Thus in group A, where some parents Avere tight-containers and others
hardbacks, the different progenies have one common feature in that
they cover the whole range of the four " nuda " forms. The total for
the group is 87 tight-containers, 144 hardbacks, 85 penulti-looses, and 41
pure looses. The fact that the tight-containers make up almost exactly
one-quarter (expectation is 89) is of interest. On the other hand it
must be recognised that there is some lack of uniformity in this first
section of the table. This is even more marked in the next section, in
which two families each lacking penulti-looses have been placed. It is
very possible that No. 5^4/1 may really belong to group A, and that
No. ^I^j'2,, with its huge preponderance of tight-containers, should go to
group D, especially as the five pure looses were in this case wretchedly
small plants all under two feet high, and with only a few spikelets each
on most.

In the following set of seven, however, a moderate degree of uni-
formity exists, though the two known parents were of distinctly unlike
kinds. Here pure looses are not found, but on a rough approximation
the tight-containers again form a quarter of the total (expectation
47).

The plants in Group D threw neither penulti-looses nor pure looses.
The proportion of hardbacks in the individual families fluctuated con-
siderably ; yet it is noteworthy that all the classified parents were of
the same sort. It was also remarked that among the offspring tight-
containers the tendency to produce a large relative number of pure
tight grains was much greater than in any other group. Numerous
plants had from 35 — 45 "/^ of pure tight grains. Taken as a whole the
group exhibits a 3 : 1 ratio.

Tight-containers were entirely absent from the last two parts of
the table. In E the hardbacks form a very uniform quarter of the
progeny, while in F, which has even less variety, the same is true of
the penulti-looses. Group E consists of 1916 sowings only and un-
fortunately nothing definite can be said about the parents.

Five distinct breeding systems, therefore, have been recognised
among the " nuda " types :

Group A. Forms throwing tight-containers, hardbacks, penulti-
looses, and pure looses.

A. St Clair Caporn 243

Group G. Forms throwing tight-containers, hardbacks, and peniilti-
looses.

Group D. Forms throwing tight-containers and hardbacks.

Group E. Forms throwing hardbacks, penulti-looses, and pure
looses.

Group F. Forms throwing penulti-looses and pure looses.

Very likely more exist. A pure line of pure looses could probably
be isolated from No. 5/8/1, and Nos. 5/23 and 5^/19 of group D may
conceivably serve as a basis for obtaining tight-containers which breed
pure. Very likely, too, the lack of uniformity in the large groups may
signify a mixture of different lesser groups. But so far as the experi-
ment has at present been carried, and with a limited amount of data
available, nothing beyond this one important empirical result has come
into prominence ; namely, that irrespective of the actual proportions of
the offspring-constituents in individual members, certain groups broadly
similar in the nature of their progeny can be established.

Spikelets.

One of the main objects for which the crosses were made was to
ascertain whether it would be possible to transfer the many-flowered
habit of the Avena nuda spikelet to the tight-grained forms. A type
with a combination of these two characters would plainly be of pre-
eminent agricultural value; for not only would it yield a very heavy
crop, but the grain would also not be liable to shake out. Occasionally
seedsmen have put on the market new varieties which have been
extracted from Avena nuda crosses, and for which it has been claimed
that the slightly higher yield they gave was due to this combination ;
but the observations set forth in this paper, dealing with some thousands
of plants of all sorts, do not support these contentions. Far from doing
so : they tend rather to foster the idea that the presence of these two
features fully developed in one and the same plant is physiologically
impossible.

The many-flowered spikelet appears to be a function of the mem-
branous palea. In "nuda" forms the spikelet with 6 — 10 flowers is
found right through the panicle. As soon as tight grains occur in it,
however, as in the F^ generation, the multiflority is partly suppressed,
the reduction always taking place in those spikelets bearing the tight
grains. When one comes to the pure tight type extracted from the
F2 generation, the maximum number of grains per spikelet ever found
is four, and that very uncommonly. Only a few of the spikelets on the

244 Inheritance of Tight and Loose Paleae in Oats

whole plant are 4-grained, and when bred on for another year the
character often fails to appear in any of the offspring.

Now none of the varieties sold ever go beyond four grains to the
spikelet ; they are never 4-grained throughout ; and they do not
remain constant throwers of 4-grained forms. Moreover, it has been
noted at the beginning of this paper that four grains may occur
sporadically in varieties normally with two. (Specimen XI.) It can
hardly be held, therefore, that extracted pure tights with some 4-grained
spikelets really show a transfer of the high "nuda" number. It is
more likely that they are just exhibiting the effect of superabundant
nutrition.

These facts incline one to conclude that when membranous paleae
are replaced by thick, stiff husks, the extra growth which would have
produced the additional grains is used up instead in the process of
strengthening the paleae.

Colour.

In this investigation the inheritance of colour has been followed
merely for the subsidiary purpose of discovering whether there might
be any connection between it and the different degrees of sclerosis of
the paleae. The figures from the crosses involving the white tight-
grained parents seemed in some respects to point to a preferential
linkage between white colour and tight husk ; but the results lacked
consistency, and adverse conditions so weathered the crop in 1916 that
the correct classification of light greys was rendered extremely difficult.
The F2 generation (raised this season) of the cross Nubian Black x.
Avena nuda, moreover, did not show any signs of the supposed coupling
being perfectly normal. In Table VI are 1225 plants raised from 60
coloured F^ plants showing splitting. The second cross suffered the
least from weathering and mice.

I

TABLE VI.

Total

Fi plants

Pure

tights

Tight-
containers

G W O W

Thousand Dollar <j x Avena nuda ? 439 170 93 48

Avena nuda 3 x Ligowo ? 259 87 55 22

Ligowo c? X Avena nuda ? 182 88 43 26

G

253
145
107

W
92
45

47

Hard-
backs

G W

08 18

44 17

21 15

Pure

Penulti-
looses

G W G W

27 9 8 3

10 1 5 2

7 — 4 —

Grand Total 880 345 191 96 505 184 123 50
Expected 919 306 229| 76^

44 10 17 5

A. St Clair Caporn 245

The whole group showed a rough approximation to a ratio of 3 grey :
1 white, but among the pure tights the ratio was practically 2:1. If,
however, it were a case of partial coupling of tight paleae with white-
ness, a much greater bias in favour of greys than actually exists should
occur at the opposite end of the scale.

Nothing similar was found among the pure tights in the Nubian
Black crosses carried on to the jPg generation this year. Segregations
in four families are shown in Table VII. All these are quite normal

TABLE VIT.
Nuhischer schwarzer x A. nuda.
Pure tights only from F^ generations.

F^ Colours

Parent Gametes

<

f 1 plant

Black -brown

Grey

White

Nubian Nuda

A

31

131

—

Bb'G X bb'G

B

40

22

1

BB'G X bb'gs

G

87

71

—

BB'G X bbG

D

IH
1 All dark.

1»

—

BB'G X bbG

2 One dark.

one very

light.

3 Or BB'g X

bb'G. •

segregations. The first conforms to the ratio 3:1, the second 60 : 3 : 1,
and the last two 15 : 1, the parent gametes producing the Fi plants
being as shown, where B and B' are factors for black, and G for grey
colour. This cumulative black in the Nubian parent will be discussed
in another paper. The point to note here is that the colour inherit-
ance is on ordinary lines and that no linkage of a particular colour
with tight paleae is manifested. The apparent indications, inconsistent
and irregular, of coupling seen in parts of Table VI must therefore be
ascribed to the weathering, damage by mice, and consequent experi-
mental error. For evea on the whole count the error is large.

Summary.

1. In "pure tight" varieties complete sclerosis of all the outer
paleae is found. The spikelets are from 2 to 4-flowered.

2. In Avena nuda all, or most, of the paleae are membranous ; a
few tight grains may occur up to a proportion of over 40 °/^ ; but " pure
tight " forms are never thrown. 6 to 10-flowered spikelets are found.

Journ. of Gen. vn 17

246 Inheritance of Tight and Loose Paleae in Oats

3. The Fi plants have grains enclosed in pure tight, pure loose,
and several intermediate kinds of paleae. The range of variation is
very great.

4. The method of segregation in the Fg generation shows that the
inheritance of what may be called the " pure tight " character depends
on a single factor. At first sight the number of F^ types appearing in
this generation seems to exceed expectation. This is because they
include tight-containing " nuda " forms which can be separated out
only by further breeding. Such forms never throw pure tights.

5. Much diversity exists in the quarter of the F^ generation in-
capable of yielding any pure tights. These " nuda " forms appear to
result from secondary factors; for, according to their type and the
manner of their breeding, empirical groupings can be made among
them. They require more extensive study.

6. In extracted pure tights the number of grains per spikelet
never exceeds four, no cross-over of the large " nuda " number taking
place.

7. There is no repulsion between grey or brown colour and pure
tight paleae.

AN ACCOUNT OF AN EXPERIMENT TO DETER-
MINE THE HEREDITY OF EARLY AND LATE
RIPENING IN AN OAT CROSS.

By A. St CLAIR CAPORN.

(With one text-figure.)

Systematic experiments on the maturation of cereal crops have
not commended themselves to Mendelian workers, partly because the
matter involved the study of characters obviously influenced to a
certain extent by climatic and edaphic conditions, and partly because
the amount of labour necessary seemed hardly commensurate with the
results apparently to be expected. Observations therefore have been
of a rough and ready kind, generally incidental notes made during the
course of experiments with some other principal object in view. It
was known, for instance, that when a spring wheat like T. polonicum
was crossed with a winter Rivett, the F^ generation was more or less
an intermediate in regard to time of ripening. It was also known that
in the F^ generation a very extensive scatter took place, though the
bulk of the crop tended to be earlier than the late parent. But there
was always uncertainty as to precisely how much stress to lay on the
fact that one parent was a winter wheat and the other a spring, and it
was largely on this account that no fixed standard of ripeness was
determined on, and complete statistical analyses made. While the
results of the investigation described in this paper cannot be said to
be other than puzzling and empirical, they can at any rate claim to
present in tabular form the exact maturation values of a whole F^
generation as established by the F^ progeny.

The two parents chosen were Mesdag and Hopetown. Their ripening

times in 1913 were

Mesdag. July 26th.

Hopetown, August 13th.

17—2

248 Early and Late Mipeiiing in an Oat Cross

These dates, while giving the time when the whole mass of each
variety was mature, take no account of the extremes of the ripening
periods. It will be seen, however, that there is a reasonably large
gap between them, and it may be stated here as well that all the
individuals of a sowing of the early parent are always completely ripe
some time before the first Hopetown panicles can be harvested. There
is no overlapping whatever.

Fig. 1. For explanation see p. 249.
In some respects Mesdag has distinct affinities with the wild oat,
A.fatua, which may possibly figure in the pedigree. The panicles are
extremely long and with a great drooping spread. They are borne on
a tall weak straw, as in the wild oat, but the considerably larger and
heavier grain makes it appear even weaker. The lower grain in each
spikelet is strongly awned and of a dark brown colour which resembles
that of the wild oat in that it masks a grey factor. By way of contrast,
however, hairiness is absent and the grain is of exceptionally fine size
and plumpness.

A. St Clair Caporn 249

Hopetown grains are a good deal smaller and narrower. They have
white husks and stiff, twisted awns. The straw is of medium length.

The two parents present a marked difference in habit in their earlier
stages. In Fig. 1 the taller plant on the left is Mesdag, that on the
right Hopetown.

Both are of the same age and were grown in adjacent rows. The
Mesdag plant is just about to send up the floriferous shoots, here limited
to one main head. The late ripening Hopetown, on the other hand,
tillers profusely; so that at the time when the Mesdag panicles are
completely out of their sheaths, its own more numerous panicles have
yet to burst through.

The F^ and F.2 Generations.

Little information as to the rate of ripening in these two gene-
rations was available ^ As far as can be gathered, the F^ plants, which
were strongly awned, with brown grains somewhat lighter in shade
than those of Mesdag, and also with more compact panicles, were ripe
before the late parent, yet not quite so soon as the early parent.

The F^ generation was harvested when the majority of the plants
were fully mature ; that is to say, when a certain proportion must have
been over ripe. Even then, however, there was a number of very
green plants to be seen. A sample of 24 plants contained 4 without
any yellowness at all in the glumes. This sample probably gave a
rough idea of the general composition of the whole F^, ; but the possi-
bility of picking out entirely green panicles from an assortment made
up of so many different stages in ripeness not occurring until most of
the crop had been rubbed out for the next sowings, no real confirmation
of this assumption can be brought forward. Nevertheless it will be
worth noting here that the F^ rows derived from the 4 green panicles
all tended to ripen late. (Table I, Nos. 103 — 106 inclusive.)

The Fs Generation.

A far more detailed study of the F^ generation was riiade. 106 F^
heads were rubbed out and planted, all on the same day. A row of
each of the parents was also sown in proximity to the rest of the crop.

^ The experiment, which came into my hands after the Fy crop had been harvested,
was unfortunately the victim of several interruptions, and suffered accordingly from
want of attention. All that could be done was simply to grow on the first two generations
without making any careful observations. Needless to say, the lack of a concise con-
ception of the Fi ripening period is keenly felt in any attempt to explain later results.

^

250 Early and Late Ripening in an Oat Cross

TABLE I.

Fs Rows, Numbers

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

Aug. 10

n

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

1

2

3

4

5

6

7

8

9

10
11
12
13
14
15
16
17
18
19
20
21
Later

1 —

1 — — 1 —

1 1 — —

2 2 4 2
— 11 —

1 — — 1

2 1 _ — _

1 — — 1 —

— 3 — 1 —

1 1 1 — —

1 — — 2

3 — —

1 —

1 —

2 — —

m

1—1

1 _ _ _
1 — — 1 1 —

2 l__ 2 — — — 2 2
— 11112211 —

1 1 1 — _ l__ 2— 1 — —

1 !____ 1 i__ 2 1 —
11112—21—2—1-
1____ 1_ 2 — — — 2 1

2 2 2— 1 2 1— 1— 3 — —
___13__221121

1__. !_______ 1 1 5

!l

m

Sept.

2 — — 3 1 2 2 — — —

1— 112 — — 21 —

__— 4 — — 1— 1 —

_ 2 — — l__ 6 1 —

4— 31— 1 — — 32

1— — — 11— 3 44

2— 2 1— 1 2— 1 —

1

_ _ _ 1 _ _ 2

2 — 1 2 2 — —

2 1 — — 1 _ 2

2 4 3 — 2 — —

_ — _ 3 _ _ _

_ 2 — — — — 1

— — 1

It

— — 1 1 3— 2 — — — 1
2 — — — — — — 3 1 1 —

- 1 — — 1 !_____
__ !______ 2 —

— — 1

* One taken a little prematurely.

t Stem nearly severed before ripe. Date therefore unreliable.

A. St Clair Caporn

251

"

TABLE I

(continued).

^3

Rows, Numbers

A

-~s

24

25

26

27

28

29

30

31

32

33

34

35 36

37

38

39

40

41

42

43

44

45

46

Aug. 10

—

—

—

—

—

—

—

—

—

—

—

—

—

■

11

—

—

—

—

—

—

—

—

—

—

—

— —

—

—

—

—

—

—

—

—

—

—

12

—

—

1

—

—

—

—

—

—

—

—

— 1

1

—

—

1

1

1

—

—

2

2

13

1

—

—

1

—

—

—

—

—

—

—

— —

—

2

—

1

—

—

—

1

1

14

1

—

1

—

1

—

1

2

—

1

—

— —

—

2

—

—

—

4

1

2

—

3

15

—

1

—

—

—

—

1

2

—

3

—

— —

—

—

1

—

—

1

—

1

2

—

16

—

1

1

1

1

—

—

—

1

1

—

— 1

1

1

—

1

2

—

—

1

—

17

2

1

3

—

2

2

2

1

1

1

1

1 1

1

1

1

—

1

—

—

1

1

—

18

3

—

2

1

2

1

—

2

1

—

—

— 1

1

—

—

—

—

—

1

—

1

—

19

1

2

—

—

2

—

1

3

—

—

1

— 1

4

1

—

1

1

3

1

1

—

2

20

1

—

—

—

2

—

2

—

2

1

1

2 1

1

1

1

3

—

2

1

—

—

1

21

1

2

1

3

—

—

1

1

—

—

—

— 2

—

1

—

—

—

—

2

—

2

—

22

—

—

1

1

2

—

1

2

—

—

1

2 —

2

1

1

1

1

—

1

—

3

—

23

1

—

1

1

—

1

—

—

—

1

—

— —

—

1

—

1

—

2

1

—

—

—

24

—

—

2

1

1

2

1

2

2

—

1

3 —

3

—

—

2

1

1

1

1

1

1

25

—

2

1

2

1

—

—

—

1

—

4

— 1

—

2

1

1

—

1

2

1

1

2

26

1

—

—

—

1

2

—

—

3

—

—

2 —

3

2

1

—

4

3

1

2

1

1

27

—

2

—

—

—

—

—

1

1

1

2

1 1

1

—

2

—

1

— .

1

3

—

3

28

2

3

—

—

1

—

3

1

2

4

2

1 1

1

1

1

2

1

1

—

—

—

1

29

—

—

—

1

1

1

2

—

—

1

1

1 1

—

—

1

—

—

1

1

—

—

—

• 30
31

—

1

2

—

—

1

2

1
1

—

1

2
2

1 1
— 1

1

—

1

1

1

—

—

1

—

1
1

—

Sept. 1
2

—

—

3

1

1

3

—

2
1

1

—

1

2 1
2 2

2

—

1
2

1
1

—

—

—

1
1

—

—

3

—

2

—

—

—

—

1

—

1

—

—

— 1

—

—

—

—

3

—

1

—

—

1

4

—

1

— •

—

—

2

—

—

—

—

—

— —

—

—

1

—

—

—

2

—

1

1

5

—

—

—

2

—

—

—

1

1

—

—

1 1

—

1

1

—

1

—

1

—

—

—

6

1

—

—

—

—

—

—

—

—

—

—

— —

—

—

—

—

—

2

—

—

—

—

7

—

—

—

2

—

—

—

—

—

—

—

— —

—

—

—

■ —

—

—

—

—

—

—

8

—

—

—

—

—

—

—

—

—

—

—

— —

1

—

—

—

—

—

—

^

—

—

9

—

—

—

—

—

—

—

—

—

—

—

1 —

—

—

—

—

—

—

—

—

—

—

10

—

—

—

—

—

—

—

—

—

—

—

— —

—

—

—

1

—

—

—

—

—

—

11

—

—

—

—

—

—

—

—

—

—

—

— —

—

—

—

—

—

—

—

—

—

^

12

—

—

—

—

—

—

—

—

—

1

—

— —

—

—

—

—

—

—

—

—

—

—

13

—

—

—

—

—

—

—

—

—

—

—

— —

—

—

—

—

—

—

—

-^

—

— .

14

—

—

—

—

—

—

—

—

—

—

—

— —

—

—

—

—

—

—

—

—

—

—

15

—

—

—

—

—

—

—

—

—

—

—

— —

—

—

—

—

—

—

—

—

—

—

16

—

—

—

—

—

—

—

—

—

—

—

— —

—

—

—

—

1

—

—

—

—

—

17

—

—

—

—

—

—

—

—

—

—

—

— —

—

—

—

—

—

—

—

—

—

—

-18

—

—

—

—

—

—

—

—

—

—

—

— —

—

—

—

—

—

—

—

—

—

—

19

—

—

—

—

—

—

—

—

—

—

—

— —

—

—

—

—

—

—

—

_

—

—

20

—

—

—

—

—

—

—

—

—

—

—

— —

—

—

—

—

—

—

—

—

~

—

21

—

—

—

—

—

—

—

—

—

—

—

— —

—

—

—

—

—

—

—

—

—

—

Later

—

—

—

—

—

—

—

—

—

—

—

— —

—

—

—

—

—

—

—

—

—

—

252 Early mid Late RiiJening in an Oat Cross

TABLE I (continued).

Fs Rows, Numbers

47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69
Aug. 10 — — — — — — — — — — — — — — — — — — — — — — —

11 ___:__ !_________________

12 — — — — — 1— !_____ l_ i__ i_.___

13 1 2— 1 1 1— 2 — — — — — 2 1 — _ i__ i_ 1

14 — — — 1— 4 — — 2 — — — 1 1 — _— 2 — — 1 — —

15 _ i__ 1 4 1___ i__ 2— 2 1__ i__ 1
16_1_323-1— 322 122 2 1 — — 13— 11

17 111 12— 111— 212 — — 2— 211— 2 —

18 1 2 — — 2 1 2 1 1 4 1 — — 1— 3— 1 1____-

19 3_1__3 i_i 1 i_i 1444 1__1 l_

20 1 2 3 — — — — 3 l_— 1— 1 2 — — 1— 2— — —

21— 1 — — — — 2 — — 1— 1 1 1 1 2— -— 1 2— 2

22— 1____ 1 1_ 2 3— 2 — — 2 3— 2 2 1 — —

23 2 1— 1 1 — __ 2 — — 2 — — 2 — — — 2— 2 1 —

24 — — 2— l__ i__ i___ 3 i_ 1 3 i_ i_

25 111111111 — — 22 — — — — 311111

26 1 2— 1 6— 2 — — l____ 4 2— l__ i l_

27 1— 1 1__ 4 32— 1— 4 — — — 1 3 — — 2 1 —

28 l__. _____ l_ 1 1 1 i__ i__ 2 3 1 —

29 — — 2 — — — — 1— 1____ i____ i___

30 — — l_____ 1 l___ i»__ 2 — — — 2 — —

31 1— 1— l__ l_ 3— 2— 2 1— 1— 2 — — 1 3
Sept. 1 1— 3 — — l___ l_ 2— l__ i__ 2— 1 1

2— 3 1________ 1__ 1— 1 1 1 2 — — —

3 — — 1__ 1______ !________ i_

4— !____.____ i_ i____ 2 1— 3— 3

5 1 — _________ 2 — _— 1— 3 2 — — 1 —

6 — — — 2 — — 1— 1 1— 1 iti__ i_ i__ 2 1

7 1- 1

10-' 1_III_I III

11 . 1

12 —

13 i:

14 _

15 —

1

2

1« 1

17 — — — ________________

18 _ — _________________

19 — — — ________________

20 — — ______. ___ ________

21 J

Later — — — —

* Probably earlier,

t Sheath not fully split open. Therefore possibly earlier.

t The first panicle aborted.

A. St Clair Caporn

253

TABLE

[ {cordinued).

^3 Rows, Numbers

/^

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

Aug. 10

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

11

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

12

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

1

—

1

2

—

13

1

—

—

1

—

—

—

—

—

—

—

3

—

1

—

—

—

—

—

1

—

—

—

14
15
16

1

—

2

1

—

—

—

—

1
3

—

2

1
2

1
1
1

—

1

—

—

2

2

1

—

—

1

_

1

1

1

_

1

2

1

1

1

17

—

2

1

1

—

—

1

—

—

2

3

1

2

—

—

1

—

—

1

—

—

—

18
19
20
21
22

1

1
1

1
1

1
2
1
2

1

—

—

—

—

1

2

—

1

1
1

—

2

—

3

1

1

1

1

1
3

2

1

1

—

1

—

—

—

2

2
3

1
2

1

1
2

2

—

2

2
2

—

1

2

1

1

1

1

1

23
24
25

—

1

1
1

1

—

1

1

—

—

1

—

—

1
2

1

1

1

1
2

3

3

1

—

1
2

1
1

—

4

_

__

1

2

2

2

___

._

26

—

1

1

—

—

1

2

—

2

—

2

—

1

3

—

—

—

1

—

—

—

—

—

27

1

2

1

—

—

—

1

—

1

1

—

1

—

2

1

4

1

4

4

1

3

1

1

28
29

—

3

1

1

3

2

—

2

2

—

—

1

1

1

—

—

1

1

2

1

1

1

4

30
31

—

1

—

2
3

—

1

1

2
3

—

—

2

2

2

—

3

1*

1

1
2

—

—

1

1

4

Sept. 1

—

—

1

2

2

1

3

1

—

1

—

1

—

—

4

—

2

—

—

—

1

—

—

2
3
4

4

3
1

1

1

1
1
1

3

I

—

3

—

1

—

—

1
3

1

1
1

2

1

1

1

1

—

—

—

—

—

3

1

1

_

_

1

2

_

1

_

5
6

7

—

—

—

—

1

1

1

3

—

1

1

1
2

1

—

1

—

—

—

—

1

—

2

1

—

1

—

1

1

2

1

1

1

2

_

_

_

2

I

I

8

9

10

—

—

—

2

—

—

1

—

1

—

1

1

—

1

—

—

—

—

1

—

—

—

—

_

_

1

_

1

_

2

_

1

_

11

—

—

—

—

—

—

1

—

—

—

—

—

—

1

—

—

—

—

—

—

12

1

—

—

—

—

—

—

—

—

2

—

—

—

—

—

—

—

—

—

1

13

2

—

—

—

—

1

—

—

—

—

—

—

—

—

—

—

1

—

—

1

14

—

—

—

—

—

—

—

—

—

—

1

—

—

—

—

15

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

16

—

—

—

—

—

—

—

1

—

—

—

—

—

—

—

—

—

—

—

—

—

—

1

17

—

—

—

—

—

—

—

—

—

—

1

—

—

—

—

—

—

—

—

—

—

—

18

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

19

—

—

—

—

—

—

1

—

—

—

—

—

—

—

—

—

20

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

—

21

—

—

—

—

—

1

—

—

—

—

—

—

—

—

—

—

—

—

Later

—

—

—

—

—

—

—

—

Ist panicle damaged. Therefore probably earlier.

254 Early arid Late Ripening in an Oat Cross

TABLE I {continued).

F3 Rows, Numbers Parents

93 94 95 96 97 98 99 100 101 102 103 104 105 106 ^esdag ?o°wn"

Aug. 10 — — — — — — — — — — — — — — 1 —

11 ______________ 3 _

12 __ — _ 2 — — — — — — — — — 1 —

13 — — — — l_________ 2 —

14— 1— 1 1— 1 — ______ 3 —

16 — — l_-___ l__ i___ _ _

17 ______ 3_______ _ _

18— 2 2— 2 — — 1 — _____ l_

19 — — 1 3 l___ 1_ l_ l_ _ _

20— 2 2 — — 1 l__ 1_ i__ _._

21— 1 — ___ i__ i_ 1 i_ _

22 — — — — — — — 1— 2 — — — 2 — —

23 113 — — 2 — — — — 11— 1 2 —

24 — — 1 — — — — 21 — — 1 — — 1 —
25— 1- 1 2— 1 1— 2— l__ _ _

26 — — — 11— 31— 2 — — 32 — —

27 2-p- — — 2 — — 13 — — — — 1 — —

28 — — 2 — — 3 2— 1 — — — — — — —

29 — — 1 — — — 1— 12— 2 1— — —

30 1_— 1— 2— 1__ 1 ._ _ _ _ _

31 — — — 1— 1— 2— 4 3 — — 1 — —
Sept. 1— 2 — — — — — — — 2 — — — — — —

2__— i____ i___ 2— — —

3 l_____ i___ l___ _ _

4 1— 1 — __ 12 — — — — — 1 — 1

5 2 — — 1— 1 — — — — — — — 3 — —

6 — — 22— 122111— 2— — 2

7 — — — l____ l_ i*i_ 2 — —

9__ i__ i__ 2 — — — — 1 — —

10 1— 1 — — _ — — _—-— 1 — — — 1

11 2 1_________ i__ _ _

12 ________ i__ i__ _ _

13 ______________ _ 2

14 — — — — — — — — — — — — — 1 — 1

15 — — — — — — — — — — — — — — — 1

16 — — — — ___ — — — __ — — — —

17 1 — — — __________ _ _

18 — — — — — — — — — — — — — — — 1

19 — — — — — — — — — — — — — 1 — —

20 — — — — — — — — — — — — — — — —

21 — — — — — — — — — — — — — — — —

Later — _ 2 — — — — — — — — — — — — 2

Parents

* Ist panicle damaged. Therefore probably earlier .

A. St Clair Caporn 255

Between the plots 1 ft. 3 in. paths were left to facilitate examination
of the growing rows from both ends. The same object was also attained
by shortening the rows a trifle, so as to make them contain 24 grains
instead of the usual 28.

A panicle was held to be fully mature when the last trace of green-
ness had vanished from the tips of the paleae. Finding the ripe heads
by this method is not so laborious as might at first be imagined. For
the colour of the glumes is more or less an indication of the ripeness of
the grains they enclose. To pick out the ripe ones, therefore, it is
only necessary to peep in between the glumes at the base of the
yellowest panicles. This system of gauging ripeness was preferred to
any based on the times at which the panicles emerged from their
sheaths, because the rate of ripening of any two panicles leafing at
the same time appears to vary a good deal.

At every daily inspection the heads cut out were given separate
labels bearing the number of the row and the date. After the whole
crop had been thus gradually harvested, the ripening period for each
row was obtained and recorded. The results are collected together in
Table I.

The Mesdag row started ripening on Aug. 10th. It had finished
on Aug. 24th. Over 70 °/„ of the plants came to maturity within the
first five days. Ten days elapsed before the first panicle in the Hope-
town row could be cut. Here again the period was 15 days, but
the sequence of the ripenings was much more regular. (See end of
Table.)

On examining the table it will be noticed that no row is as late as
the late parent, and none capable even of reasonable approximation
to it. Nor is there any exactly as early as Mesdag, although two,
Nos. 17 and 97, extend only two and three days respectively beyond
the Mesdag period. If, however, as is likely, a certain amount of
latitude must be given, then these two rows are to be reckoned
equivalent to parental Mesdag in earliness. Two out of 106 may possibly
indicate a 1 : 63 ratio.

Another survey of the table reveals the fact that there are 24 rows
whose ripening period, starting early, extends over the gap between
the parental times and finishes on the first day of the Hopetown harvest
(Sept. 4th). The reverse condition, in which ripening starts immediately
after the last Mesdag, is not found at all. The number 24, however, is
possibly significant ; for it offers a very close agreement with 265, which
is the expectation on a 1 : 3 basis.

256 Early and Late Ripening in an Oat Cross

The general conclusion arrived at as the result of this experiment
therefore is that there is ground for supposing that early and late
ripening are mendelian characters, but that they are dependent on
more than one factor, possibly three. ^2 plants homozygous for one of
these, it would seem, are definitely early in the sense that their F^
rows all reach complete ripeness before any individuals of the late
parent are mature ; yet, in all except 2 cases out of 24, they cannot be
called exactly as early as Mesdag. Apparently it required homozy-
gosis in all three factors to produce these two perfectly early rows.
The diversity of the remaining 22 early ripeners, where in some
cases {e.g. Nos. 20, 52) the accumulation of the ripening occurs
during the first part, in others {e.g. Nos. 28, 44, 54, 61, 72) in the
middle, in others again {e.g. Nos. 29, 49, 71) at the end of the period
from Aug. 10th to Sept. 4th, certainly supports the inference that
although one factor is here constant and homozygous, thus imparting a
common measure of earliness, the other two, in various heterozygous
and homozygous combinations, are responsible for the different in-
cidences of the greatest intensity of this earliness.

There can be little doubt, too, that the extent of the tillering has
much to do with the rate of ripening. The tillering power of late
forms is always good; that of early, very poor. Owing to the con-
centration of growth among early plants into one or two panicles only,
these are generally bigger and bear better grain than those of late
plants ; but this advantage does not compensate for the diminished
yield due to the small number of heads. There is thus an inevitable
sacrifice of crop when it is attempted to render a late kind early. The
only hope, apparently, lies in increasing the output, per individual panicle,
of an already early variety. This can best be done by extracting it again
from a fair-sized F^ generation of a cross with a type which, quite apart
from any ability to tiller profusely, has above all larger panicles and
larger grain of better quality.

SUMMAKY.

1. Earliness and lateness are not instantaneous, sharply defined
characters : they are spread over a period.

2. The ripening times of the parents used did not overlap.

3. Poor tillering power was characteristic of the early parent.

4. Fi plants more or less intermediate.

I

A. St Clair Caporn 257

5. The ^2 crop was harvested all at once. Hence no definite data
were available. Evidence of early, late, and many intermediate forms,
however, was visible.

6. Systematic, gradual harvesting of the F^ generation was under-
taken, together with parent rows grown under exactly the same
conditions, to determine the constitution of the individual F^ plants.

7. Analysis of results. Earliness is possibly a function of 3 factors.
A type which is comparatively early, in that its -^3 period never extends
into the period of the late parent, is segregated on a 1 : 3 basis.

8. Suggestions are made with regard to the improvement of early
varieties.

Since completing this paper the writer has been referred to the
work of Hoshino (1915) on peas and rice. In the case of the latter
this investigator made observations on the " shooting times " in a cross
between an early and a late variety. By " shooting time " is meant
the time at which the first spikelet of the inflorescence emerges above
the ligule of the sheathing leaf Apparently in rice this is also an
exact measure of the relative maturing time of the grain later on.
Hoshino's F3 results were remarkably similar to those given in this
paper for a different cereal examined on a different system. Two only
out of forty-six randomly taken F3 rows were as early as the early
parent, while twenty-one stretched across the "gap-period," but not
into that of the late parent. As only ten plants per row were raised,
the number twenty-one is probably too high, and would, had there
been more plants, have been nearer 11 — 12, which is the expectation
on the 1 : 3 ratio. Although he overlooks this important fact, however,
Hoshino nevertheless strongly favours the multiple-factor hypothesis
as an explanation, deducing from the very small number of pure eaflies
like the original early parent that there are three factors.

The value of this present paper, therefore, does not lie in the
originality of its inferences, but rather in that similar conclusions to
those of the Japanese worker have been reached perfectly independently
by different methods, and also in a cereal.

LITERATURE.

Hoshino. 1915. Imperial Unioersity^ Sopporo, Japan. Vol. vi. November.

ON A CASE OF PERMANENT VARIATION IN THE
GLUME LENGTHS OF EXTRACTED PARENTAL
TYPES AND THE INHERITANCE OF PURPLE
COLOUR IN THE CROSS TRITICUM POLONI-
CUM X T. ELOBONL

By A. St CLAIR CAPORN.

(With Plates XIII and XIV, three charts, and one text-figure.)

T^E work set forth in this paper was initiated by Professor Biffen,
who made the cross in 1913 and carried it on to the F2 generation.
Thereafter, from the analysis of the F^ figures, the experiment has
been in the hands of the writer. Rust, together with weakness of
straw, was often responsible for considerable damage, lowering and
shrivelling the grain output per plant, and thus increasing the difficulty
of discovering faint colour traces usually borne on only a few of the
riper grains.

The Polonicum Parent.

Triticuni polonicum is a tallish wheat about 4 — 4^ feet high. The
tillering is poor, most of the plants having only two stems which difi*er
from ordinary straws in being solid. The species is moderately sus-
ceptible to rust.

The ear is very long, lax, and bearded. Owing to their great length
the outer glumes are a characteristic feature. Practically smooth, flat,
and deeply keeled, they spread out stiffly at flowering time, giving to
the ear the appearance of having been rubbed from the tip downwards.
(Plate XIII, figs. 1 and 2.)

Average glume length (82 ears) = 29'23 mms.

The frequency distribution is shown in Chart I.

260 Inheritance of Glume Length and of Colour in Oats

For purposes of measurement the glume of the fourth grain-con-
taining spikelet from the base of the ear was always used. Empty

CHART I.

o

PLI

9

\ /i

8

)

7

-

1
1

1

\ / '. /

\

,6

-

\ j' 'i /

<

4

-

V ', ;

\

3

_

;

'„'

« , — -

2

_

(

'

'■' \

A

1

1 1

1

1 1 1 1

1 1 1 1

1

III II Tnj

I

17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41

Glume Lengths in mms.

basal spikelets were never counted. As the number of measurements
available in this experiment was not over large, and the separate
plotting of the lengths involving fractions of the unit would have jolted
the curve considerably, it was decided to record them as nearest whole
numbers. The method of dealing with these non-integral numbers is
shown in Table I which contains the data for Chart I.

TABLE I.

Glume Lengths in

Millimetres

20 21 22

23 24

25

26

27

28 29

30 31

32

33

34

35

36

37

38

39

40

Whole
Numbers

0 10

4 4

2

1

6

5 3

7 0

7

4

1

2

2

0

2

1

1

\ below

040

\ i

li

li

li

li 1

i 2

2

1

0

1

0

1

0

0

0

\ above ..

\ 0 \

\ li

14

li

14

1 i

2 i

1

0

1

0

1

0

0

0

0

Totals for ) j^ ^1 , K «
each length) 222

4 9 74 41 9i 2i 8i 5 2 3 3 1 2 1 1

The only fraction considered was the half unit, or '5 mm. An
example taken from it will explain how the table has been compiled.
In addition to the six ears with a glume length of 27 mms. and the
five of 28 mms. there were three of 27*5 mms. Of these three, half
{i.e. \\ ears) go to the whole number which 27*5 mms. is immediately
below {i.e. to the 28 mm. length), and half to that which it is "5 mm.
above {i.e. to the 27 mm. length). This method of tabulation has been
employed in every case where a series of glume measurements has
been taken.

A. St Clair Caporn 261

As is to be generally expected when the glumes are so elongated,
the grain is also longer than that of ordinary wheat. It is hard and
pointed, has a very deep furrow, a whitish colour, and a translucent
appearance.

Plate XIV, fig. 4 (a, a), shows typical Polonicuni glumes — maximum
and minimum.

The Eloboni Parent.

The parent stock was raised from a sample of grain brought from
Abyssinia ; but, owing to the ravages of rust, to which this variety is
extremely susceptible, the strain was unfortunately lost in the second
and third years of the experiment.

Here again the tillering is poor and the plants small. The ears
are of small size, narrow, fairly lax, and bearded. (Plate XIII, fig. 3.)
In contrast to T. polonicmn the chaff is very short, the glume lengths
ranging from 8 mms. to 13 mms. A series of 51 short-glumed Elobonis
in the ^3 generation had an average length of 10*58 mms. ; but, as will
be shown later, there is reason to believe that the extracted parental
types have a markedly different average from the original, for which
the figure, in all probability, would be fully 1 mm. lower. No extensive
measurements of the glumes of the Eloboni parent, however, were
made, and as the stock had entirely died out when the experiment was
handed over to me, the conjectured average glume length is based
mainly on an analogy with the behaviour of the " pure long " glume.

The grain is short, stumpy, hard. It has a tendency to shrink and
become slightly wrinkled.

It is in the purple colour, however, that the chief feature of this
wheat is to be found. Unlike maize, where the aleurone layer con-
tains it, the pigment is located in the pericarp and therefore belongs
to the same generation as the plant which bears the grain. When
the grain has been well soaked in water, the pericarp of the caryopsis
may readily be peeled off. It consists of

The Epicarp. A tissue of narrow cells elongated >y The first care-

in the direction of the long axis of the fruit and with
pitted lateral walls. At the apex of the grain it bears
a few hairs.

The Hypoderm and Mesocarp. Similar to the epi-
carp but with fewer pits and firmly united to it.

ful peeling
generally re-
moves these
three as one
skin.

Joarn. of Gen. vii * 18

262 Inheritance of Glume Length and of Colour' in Oats

The second
skin.

The Girdle Cells. Very regular, delicate cells with '
pitted lateral walls and long axes at right angles to
those of the components of the first skin.

The Tube Cells, or Endocarp. Scattered vermiform
cells, often coalescing, running lengthwise in the grain.
They resemble later tubes and the longer the grain
the more conspicuous they are. They are always best
developed at the embryo end and on the side remote
from the groove. To them often adhere

The Spermoderm and the Perisperm, two delicate
cell laminae which are not strictly pericarp, but the
outermost layers of the seed.

The purple colour is found in the sap of the girdle cells, staining all
or part of it according to the distribution and intensity. The position
of this pigmentiferous layer is shown in Plate XIV, figs. 9 and 10, the
first of which shows it in a transverse section of the grain, and the
other in a superficial view of the girdle cells, more particularly where
they are not covered over by the fragment of epicarp. After immersion
in dilute sulphuric acid (1 in 20) for a short time, the grain turns
a deep crimson. This reaction, in conjunction with microscopical exami-
nation, has proved very useful in the detection of faint purple tinges
in grains which showed a slight streak to which the overlying portions
of the pericarp gave a doubtful, brownish appearance. For whereas
the acid produces a distinctive coloration, water causes the antho-
cyanin first to fade and then to become dull green, — a colour which
is much commoner than crimson, far less vivid, and by
distinctive.

In both figures the acid reaction colour is illustrated.

no means

fl

The F^ Generation.

A type of ^1 ear is shown on Plate XIV, fig. 5. All the ears approxi-
mate to this type, of which the glumes are intermediate in shape and
size between those of the parents, though the fluctuations range from
glumes slightly longer than those of Eloboni to others a little shorter
than the minimum Polonicum glume. Plate XIV, fig. 6, illustrates these
two types (c and a) as well as the more general average (6). The grain
is also of intermediate dimensions, as will be seen fi-om Plate XIV, fig. 7,
m which it is placed between the parent grains for comparison.

The ^1 grains were all purple.

A. St Clair Caporn ' 263

The F^ and Fr^ Generations.

Glumes.

183 F2 plants were examined. As is usual in T. polonicum crosses,
it was found that, while certain ears were clearly long- or short-
gluraed, there was a very large proportion which verged on these
extremes, as well as those which were obviously intermediates. The
glume lengths were therefore measured and plotted in the form of a
curve (Table II and Chart II).

The curve (dotted), it will be seen, falls into three periods, indi-
cating that segregation, although indistinct, has nevertheless taken
place. Professor Biffen^ has already explained the real significance of
such a curve in other crosses of this kind. It results from the over-
lapping of three separate curves representing the lengths of homo-
zygous short, homozygous long, and heterozygous short-long glumes.
Where the curves overlap there is a piling up effect equal to the sum
of the values of the constituent curves ?it these points. Thus : —

Curve t

Curve X.

The triple nature of this type of compound curve is more plainly
recognisable when a greater number of glume lengths is available for
measurement. The dot-and-dash curve on Chart II represents the
glume lengths of 151 F^ individuals derived from 10 heterozygous F.,
ears, and when added to the first curve, produces the large continuous-
line curve in which the three periods are far more distinct. It must
be admitted that the third region is far from smooth. This is due to its
diffuse character. By adopting the smoothing line (sm), however, one
can indicate the real position. The point P marks the summit of the
overlapping curves — always very pronounced where a high curve overlies
a very low one.

1 Biffen, R. H., Jo\xrn. Agr. Sc„ 1905.

18—2

264 Inheritance of Glume Length and of Colour in Oats

TABLE II.

Glume Lengths in Millimetres

7 8

9

10

11

12

13

14

15

16

17

18

19 20 21 22 23

24 25

26

27 28

29

30

31

Numbers]

1

13

3

4

6

5

11

13

9

7

5 7 6 2 2

1 0

2

0 0

0

0

0

1 below 0 i

0

3

6

4

24

2

3

4

3

3

n h n n h

1 4

0

1 4

1

0

0

1 above ^ 0

3

6

4

2i

2

3

4

3

3

2i

\ n ^ i 1

I 0

1

4 1

0

0

4

elcitngth} i 24 4 22 13 IO4 IO4 10 18 20 15 I24 8

CHART XL

4 34 24 4 3 14 14 1 0 4

39

'

38

_

37

_

11

36

_

1

A

35

1

1 \

34

_

1

1 \

33

_

1

1 1

32

_

1

1 \

31

_

1

1 \

30

1

1 \

29

_

1

I \

28

-

1

\

27

_

\ 1

\

26

_

\ I

\

. 25

\ j

\

W) 24

a

<o 23
a, 22

1 2^
3 ?n

-

\

1-

_

\ \ /

/\ \

1 -

« 17

^ 16
0
« ^5

-

H ;

5z; 14

_

1 \ \

13

_

■I

\ ''

i \''' \

12
11

-

'i
i

\\ 1 • ^

M

P

10

,/

v- -./.■ \

\ \ ^

\

9
8

7
6
5

- ■

i

V

\ \ -
\ V
\

\

\
\

'%A

4
3
2
1

.1111

j i)

1 1 1 1

\
\
\
\

1 1 1 1

1 1 1 1

1 1 V i"-'^

C

12 3 4

3 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 3

0 31 32 33

Glume 1

jengths in mms

i.

i

A. St Clair Caporn 265

The 183 F2 ears, each representing a separate plant, were all rubbed
out and planted ; but although all had been measured for the purpose
of compiling Table II, they were not specifically identified in the
sowings. They yielded 170 ^3 rows with a sufficient number of plants
per row to permit of accurate classification. The segregation — an un-
doubted 1:2:1 ratio — is shown in Table II a.

TABLE 11 A.

Number of F3 Kows

Character of F^ Segregation

Observed

Expected

Short only

41

42-5

Short, medium, and long ...

87

85

Long only

42

42-5

A series of sowings in which the glumes of the parents had been
carefully measured proved that the maximum for pure shorts was
12*5 mms. and the minimum for pure longs 18 mms., while the upper
and lower extremes of heterozygosity were 20o and 12 mms. respec-
tively. The extracted pure long types, it was fi-equently noticed,
seemed to be somewhat more compact than T. polonicum, and a more
careful study of them was therefore made. The glumes of the progeny
of certain pure longs of known size were measured and compared with
those of an equal number of Polonicum plants grown in the same
season and under identical conditions. 82 plants of each kind were
examined. The influence of the parental glume length on the progeny
in the case of these pure long forms was nil ; for a plant with a 19 mm.
value gave a row of offspring with the same average length as did one
of 275 mms.

In Chart III the taller curve has been plotted from the extracted
pure longs (Table III), while the dotted curve represents the frequency
distribution of an equal number of Polonicum ears. The average
length of the former was 24-15 mms. ; that of the latter 29"23 mms.
The difference is very considerable. Furthermore, there is a tendency
towards greater uniformity among the extracted longs (cf the large
proportion of glumes 24 and 25 mms. long). In all probability the
extracted pure shorts exhibit much the same sort of alteration in
average value of the glume length when compared with the Eloboni
parent, with the difference that instead of being lowered, the average
length is increased. For although no proper comparison could be made,
owing to the destruction of the Eloboni stock, nevertheless, as far as
could be judged by instinctive eye tests, most of the extracted Elobonis
appeared to be appreciably longer in the glume.

266 Inheritance of Glume Length and of Colour in Oats

TABLE III.

Glume Lengths in Millimetres

Whole Numbers
\ below
I above

18 19 20 21 22 23 24 25 26 27 28 29 30

1 15 7 41 8 12 33 5 02

0 ii0 2|i2 2J22 1i01

I i 0 2i \ 2 2i 2 2 11 0 10

Totals for each length IJ 2 5^ 9^ 7 3^ 12J 16^ 7 6^ 6^ 1

CHART III.

17

'

16

_

15

rC 14

1 ^^
O 10

-

!-l

S, 9

1 3

-

/ 1 ''

\ ' 1 /
* 1 , ;

\ ' ' '

\

o 6

"\ ' / ' '

\

►S 5

\ '' '' I /

\

52;

\ // *^ '

\ tf ', ,'

'j

4

-

V "•'

\

»

3

_

.'

\ A '"'

> .•- — •,

2
1

/l

Ir'-r'' 1 ^^

/
(

1 1 1 1

1 1 1 \

V' \

1 1 t 1

1 1 1 'ni

17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41

Glume Length in mms.
Average Length of Extracted Longs = 24 -15 mms.

Parent

= 29-23 mms.

Briefly, the result of crossing a long-glumed wheat with a short-
glumed is this : —

Along with the ordinary segregation there is established in the F^
generation a kind of telescopic effect, whereby the means of the two
homozygote curves are brought nearer to that of the heterozygotes
than the means of the parents would actually be. This condensation
persists right through into the F^ generation in which, owing to the
possibility of isolating the ' pure long ' and ' pure short ' curves, it can

A. St Clair Caporn 267

be the more readily observed. There is thus every indication that
this slight change in regard to the average glume length of extracted
pure types as compared with the parents is a permanent one.

Grain Colour.

Before detailing the numerical results of this aspect of the experi-
ment it will be convenient to describe here the different colour types
found in the F2, generation. Full coloured purples resembling the
Eloboni parent, but of light and dark shades, were thrown. Besides
these, however, a number of streaked individuals appeared. The colour
of the streak is a dull brownish purple. The purple element is always
intensified by treatment with dilute sulphuric acid, which turns it
bright crimson. In Plate XIV, fig. 8, the various colour types are shown.
Grain a is the ordinary Polish type, of a light yellow translucent appear-
ance, but with no purple colour in the girdle cells, and hence, as far as
this cross is concerned, classifiable as ' Non-coloured.' The grain 6 is a
full purple, or ' Flushed ' form, while c and d illustrate the ' Streaked.'
h', c', d' show the colour changes in acid. The streak shown in d is
by no means the minutest recognisable, which may be so extremely
faint, possibly because the place for the strongest streak on the plant
may be on an unripe secondary tiller grain, as to be nearly invisible to
the naked eye, even when treated with acid and examined in the
brightest light. In such a case microscopic inspection reveals crimson
stained nuclei in a small cluster of girdle cells, but very little pigmen-
tation of the sap. (Plate XIV, fig. 10, x.) Care has to be taken, in
circumstances like these, to ensure that the colour is really crimson in
acid, not reddish or pinkish-brown ; for nuclei and their adjacent proto-
plasm often take on these tints in the grains of non-coloureds which have
been much weathered or somewhat rusted. In Plate XIV, fig. 10, some
of these potentially misleading colours have been introduced for the sake
of comparison with the acid reaction of the real purple pigment.

A noticeable feature in ears containing full coloured purple grains
is the way in which the same kind of flush stains the glumes. The
glume purple is most intense just before the grain hardens. As com-
plete maturity is slowly attained it fades away altogether. In pale or
particoloured forms its presence is less vividly marked, a faint purplish
margin to the glume being the sole indication.

Development of colour in the grain seems to be dependent to a
variable extent on direct exposure to sunlight. This is especially the
case in streaked forms where the width of the gape of the paleae.

268 Inheritance of Ghime Length and of Colour in Oats

occurring when the grain is at its plumpest, appears largely to deter-
mine the size and intensity of the streak. No doubt this also accounts
for the fact that only some of the grains in the ears of a streaked plant
actually bear the streak,* as but few get exposed in this way, in the
long-glumed plants particularly. That direct sunlight exposure does
not alone determine colour production, however, is proved -by the
existence of the flushed forms, wherein the greatest colour area is de-
veloped in well covered up portions of the grain. The action of direct
light may start in these grains the chemical change which renders the
colour visible, but that it is essential for its continuance is a matter of
doubt.

On the basis of the colour classification already indicated the F^
generation consisted of

28 Flushed, 8 Streaked, 136 Non-coloured.

Or, eliminating all which gave less than F^ plants next season,

27 Flushed, 8 Streaked, 123 Non-coloured.

The nature of these F2 plants was determined firstly by examination
of their grain in the ordinary way, and subsequently, after the more
accurate acid test had been adopted, by re-examination of any seed
left over from sowings for the F^ rows, or, when none remained, by
considering the F^ segregations analogically.

Thus, in the case of the 8 streakeds, only one (which had a very
strong streak) was detected on rubbing out by the rough, hurried,
unaided eye test then used. Later, after the sowings for the F.^
generation had been made, the grain left over from three of these was
tested in acid and streaks were discovered microscopically. Now as no
grain remained over from the other four F^ sowings, that the streak was
present in their F^ parents was deduced from the fact that the progeny
showed exactly the same type of segregation as that of the undoubted
streaked forms.

In the same way the 123 non-coloureds were definitely identified as
such. After sowing, surplus seed was left over from 37 F^ plants. It
averaged about 8 grains per plant, and, coming from short-, inter-
mediate- and long-glumed forms indiscriminately, constituted a good
random sample. All these grains gave a negative colour reaction in
sulphuric acid, and the analogous segregation in the F3 rows of the
remaining 86, — roughly eye-tested even though no grains were left
over for subsequent microtesting, — settled the correct classification of
the whole group.

A. St Clair Caporn" 269

It should be emphasized that whenever an insufficient number of
plants forjudging was present in an ^3 row of which the parental colour
was doubtful, or roguing was suspected, these cases were not included
in the count; so that the 27 flusheds, 8 streakeds, and 123 non-
coloureds contain no individuals about which any uncertainty exists.
Furthermore, although no particular note was made regarding the
matter, very probably the grains from which these 180 or so F^ plants
were derived all came off a single F^ plant. For, in spite of poor
tillering power, plants of later generations with a fair amount of room,
such as those at the end of rows, were observed with five or six well
filled ears; whereas among the ^1 individuals the encouragement to
tiller is considerably greater, since they are always purposely well
spaced.

Apparently, then, one, and only one, heterozygous combination gave
rise to the -^2 generation. This conclusion is supported by the sim-
plicity of the ratio : it is too even to be a fortuitous mixture of different
F2 segregation systems. For 27 : 8 : 123 is, considering the not very
large number of plants involved, a reasonably close approximation to
3 : 1 : 12, the actual expectation being 30 : 10 : 118.

In view of the full purple colour of the F^ grains, the minor pro-
portion of flushed forms in the F^ generation is very striking. No less
so is the behaviour of the 123 non-coloureds, of which

111 threw non-coloureds only, and
12 „ „ „ -f- streakeds.
[In the 12 non-coloureds absence of colour was apparently a dominant,
as the streakeds made up the smaller part of the offspring.]

Now if the ratio 3 : 1 : 12 be expressed in the form 48 : 16 : 192,
the third term can be conveniently split up on a 1.5 : 1 basis so that
the whole may be restated thus,

48 : 16 : (180 -f 12).

On applying this to the F^ results the comparison with expectation
reads as follows :

Observed. 27 flusheds : 8 streakeds : (111 -|- 12) non-coloureds.
Expected. 30 „ : 10 „ : (111 -f- 7)

While, therefore, the probability cannot be pressed to the verge of
certainty, there is still ground for believing that the two kinds of non-
coloureds in the F^ generation were to each other as 1-5 to 1.

To speculate on the Mendelism of these figures at this stage were

270 Inheritance of Glume Length and of Colour in Oats

eminently unsound. Additional data are needed. Further breeding
on of all the colour-yielding forms in the cross furnishes us with them ;
but, as will be seen from Table IV, into which they have been gathered,
although of precise and distinctive kinds, by their very multiplicity
they tend rather to complicate than to clarify the problem.

For the sake of convenience letters have been used to denote the
various glume and colour types in the table. Whenever the parental
colour is enclosed in brackets, the identity of the ^3 segregation with
others of definitely known parentage has been the means of its deter-
mination.

In the first part of Table IV are the flushed forms which bred true.
They are followed by one which gave rise to 18 flusheds and 6 streakeds.
The possibility that this may really belong to the next group, which
possesses non-coloured elements besides and conforms to the ratio
12F : SS : IN, is admitted; but other examples of the same sort of
segregation (picked out of Fi results and therefore not included here)
show that in any case a type throwing SF : IS does exist. Some
members of the 12 : 3 : 1 section, perhaps, by reason of their incom-
pleteness will call for a few words of explanation. No. 4 has 17 flusheds
and 4 streakeds but no non-coloured. Although it is quite conceivable
that it may belong to the 3^ : IS class, the low proportion of streakeds
gives cause for hesitation in so placing it. Unfortunately only one of
the 17 flusheds was carried on for another generation; but as it yielded
15 flusheds and 3 streakeds — again a low proportion of the latter —
with one of the streakeds very faint indeed and very likely hetero-
zygous, it was decided to retain No. 4 where originally placed. In
No. 3 there were no streakeds. Out of four of the flusheds which were
grown on for another year, however, one gave rise to 14 flusheds and
6 streakeds. The other three bred true. If it were a case of two
cumulative factors making for the ' flush ' character, and no ' streak '
factor in the zygote at all, this type of segregation could not have been
produced. -415 flushed : 1 non-coloured ratio must therefore be rejected.
Moreover, as the single non-coloured forms less than one quarter of the
total. No. 3 goes to the 12^ : SS : IN group instead of to the following
set, wherein the non-coloureds are produced in considerably greater
relative numbers and the streakeds constitute the smallest term in the
ratio 27 : 9 : 28. By analogy No. 7 does likewise.

Two more groups bred from flushed parents remain. In neither do
any streakeds appear. The first has flusheds to non-coloureds in the
ratio 3:1, the second 9:7.

A. St Clair Caporn

271

TABLE IV.

[L = Long. il/= Intermediate. -St = Short. F= Flush. S = Streak. ^= Non-coloured.]

>■„

irlUIllC

No. of J'^s
plants
m row

Colour Segregation

Group Totals
Parental Fa - ^

>I08.

character

F

.S

N

grain colour F

S N

Ratio

13

St

8

8

—

—

F

14

St

23

23

—

—

F

72

FureF

20

M

21

21

—

—

F

102

St

20

20

—

—

F

8

St

24

18

6

—

F

18

6 —

3: 1

2

L

16

12

2

2

F \

3

M

10

9

—

1

F

4

M

21

17

4

—

F

6

L

25

22

2

1

F

7

L

13

12

1

F
F

119

22 13

12 : 3 : 1

9

M

21

15

3

3

(115-4)

(28-9) (9-6)

10

L

12

8

3

1

F

12

St

15

8

4

3

F

15

M

21

16

4

1

F ,

17

St

25

9

6

10

F

18

M

22

8

2

12

F

44

15 38

27 : 9 : 28

21

M

26-

12

5

9

F
F

" (40-9) (13-6) (42-4)

106

M

24

15

2

7

11

M

25

19

—

6

F

16

M

26

20

6

F

70

— 21

3 : — : 1

19

M

18

14

—

4

F

(68-25)

(22-75)

103

L

22

17

—

5

F

1

L

20

11

—

9

F \

101

M

25

16

—

9

F

43

— 37

9:— :7

105

L

12

5

—

7

F

' (45)

(35)

J

M

23

11

—

12

F

72

St

8

—

8

--

[S]

86

M

6

—

6

—

[S]

63 —

Pure S

93

M

27

—

27

—

S

100

M

22

—

22

—

s

60

St

23

—

17

6

SI

129

M

21

—

15

6

s

71 21

— :3 :1

9

M

21

—

17

4

IS]

(69) (23)

h

M

27

—

22

5

IS]

45

M

7

—

2

5

im'

49

St

22

—

6

16

[N]

21 66

— : 1 :3

84

L

11

—

1

10

[N]

-

(21-75) (65-25)

and

110

M

23

—

6

17

N

or 3 : 13

112

St

24

—

6

18

[M>

74

L

16

—

8

8

N

85

M

25

—

8

17

N

96

M

16

8

8

[N]

109

M

25

—

11

14

N

-

58 78
(59-5) (76-5)

— :7:9

136

M

17

—

7

10

[N]

145

M

18

—

8

10

m

159

31

19

—

8

11

N

104

M

4

2

—

2

F

—

— —

—

35

?

1

1

—

—

?

—

— _

—

a

St

5

2

—

3

F

—

— —

—

160

M

8

2

—

6

N

—

— —

—

146

M

23

1

5

17

N

—

— —

—

34

M

5

1

1

4

Very

?
faint.

'~~

—

272 Inlieritance of Glume Lengtli and of Colour m Oats

The 8 F2 plants with streaked grains were divisible into two classes
of 4 each : —

(1) Pure streakeds.

(2) Heterozygous streakeds.

The latter provided a mixed progeny exhibiting a 3 : 1 segregation.

The next two groups of F3 generations, both derived from non-
coloured parents, have already been mentioned. Their chief character-
istic is the dominance, or pseudo-dominance, of absence of colour. The
first is composed of one quarter streakeds ; in the second the proportion
is seven-sixteenths.

At the base of' the table all the anomalous colour-throwing Fq plants
have been collected. That the cause of the anomaly is traceable in
several would sufficiently justify their being cast out altogether, but in
the interests of exactitude and to avoid criticism on the score of de-
liberate omission, they have all been inserted. No. 104 contains only
4 plants — too small a number for an accurate determination of the
proper ratio. No. 35 has a similar fault. In addition the three or
four parental grains were so badly shrivelled and rusted that the colour
question had to be left unanswered. The mere germination of the one
was in itself a surprise. In three out of the remaining four the facts
all point to certain roguing. T. polonicwn and its derivatives are
clearly susceptible to more roguing than most wheats on account of
the wide gaping apart of the flowering glumes when the anthers are
dangling out. It is quite likely that sometimes cleistogamy has not
been effected before this occurs, and that very occasionally a stigma
receives foreign pollen. In certain cases this roguing and its origin
are readily discoverable. For instance, a purple-grained F^ ear with
long glumes derived from a long-glumed F2 ear gave an F4, row con-
taining 14 plants, all of which were long-glumed and bearded. The
exception had slight scurs, intermediate glumes, and purple grain —
obviously the result of one of the stigmas on the parent ear receiving
pollen from a beardless, short chaffed non-purple belonging to another
culture altogether. Something similar happened to No. a in which it
would seem that a purple F^ grain was contaminated with pollen from a
beardless non-purple. For the F.^ ear was a beardless purple which
threw an F.^ row containing 4 beardless plants, of which one was plainly
of Squarehead origin. No. a may be a pure purple, and the unravelling
of the F^ figures would doubtless result in the settling of this point ;
but they are too few. More troublesome to explain is No. 160. Beard-
less in the F^, thus indicating roguing of an Fi grain, and of a very

4

A. St Clair Caporn 273

dark, translucent colour, which, however, was not purple, it nevertheless
gave rise to an F^ in which two flushed purples appeared (both beardless).
It is of course not at all unlikely that here the foreign pollen came
from another kind of non-coloured altogether, namely, one possessing
an inhibitory factor ; but in a case like this, with only 8 F^ plants to
study and the roguing as far back as the jPj generation, it is not safe to
attempt deductions. In No. 146, on the other hand, the roguing took
place later, — in a single F^ grain to judge by the bearded nature of
the F2 ear and the F^ of 22 bearded + 1 beardless flushed. Hence if
the purple plant be not a stray from some other row, and the rogue
pollen from a non-coloured form, the F^ segregation may possibly be
correct, though I doubt it.

Bearded like the parent the last F^ row, No. 34, came from one of
about half a dozen F^ ears with such shrivelled, sprouted grains that it
was not thought they would grow, and they were consequently planted
speculatively with the colour undetermined. In any case, however,
five plants are scarcely enough on which to decide the dominance or
otherwise of purple in this particular row. But although extraneous
pollination does not appear to have taken place, it is not unlikely that
the purple individual came out of No. 35 which possessed but a single
plant : mistakes are easily made in harvesting two adjacent rows such
as these with extremely few and scattered plants in each.

To sura up this gi"oup of doubtfuls : of the six, Nos. 104, 35,
a, and 160 can be totally rejected on account either of insufficient
numbers or of hopeless roguing at an early stage, while only the
remaining two, in which the degree of doubt is slightly less, though
still very strong, may qualify for a few further words in the following
discussion of the results.

At the outset let it be understood that no attempt is being made to
formulate a theory to tit the facts. Tentative suggestions and inter-
pretations may have to be adopted to explain some of them and also to
aid in the readier appreciation of certain aspects of the problem ; but
for all that, the contradictory nature of the figures makes them so
thoroughly baffling that at present it must be honestly confessed no
scheme which will embrace them as a ivhole can be advanced.

In some general features the colour inheritance has points in common

with that of aleurone colour in maize. East and Hayes', in dealing

with a case involving purple, red, and non-coloured forms, write that

" the only difficulty in alining the results obtained with the ordinary

1 East and Hayes, 1911, p. 83.

274 Inheritance of Glume Length and of Colour in Oats

behaviour of the known factors, is the fact that almost none of the
^3 generation show the same ratios as the F^ generation." They then
instance a ratio of 1843 purples: 188 reds: 545 non-purples obtained
from ^1 seeds which is obviously 12 : 1 : 3. Nothing like it appeared
in the following generation, but, while admitting later that the de-
ficiency of reds is too great to justify this step fully, they seize upon
the slender chance of their figures representing a ratio of 9 : 3 : 4 and
thence unfold a carefully formulated theory with which most of the
other results accord fairly well.

Now a 12 : 1 : 3 ratio appeared in the F^ generation of this wheat
cross, but the opportunity for harmonising it, — or, at any rate, attempting
to do so, — with any even remotely similar F^ segregation was absolutely
lacking. For in all the subsequent 12 : 1 : 3 ratios the colours of the
terms were transposed. In F^, 12:1:3 represented non-coloureds,
streakeds, and flusheds respectively; in F^, flusheds, non-coloureds, and
streakeds. The nearest approach to the F^ ratio, if one ignore the far
too high proportion of streakeds, was in the rogued row, No. 146 ; but
because of the adulteration no confidence can be put in it. Further-
more the F^, seeds which gave rise to the row differed from the F^
grains in one vitally important point. They were not purple. In any
case, however, even were as many as 4 factors involved in the pro-
duction of colour, out of 256 plants 16 heterozygotes identical with the
Fi type would be expected, — in other words at least 9 out of 158.

It may be pretty definitely asserted, therefore, that as far as this
particular purple x non-purple cross is concerned no F^ ratio arising
from purple seeds like those in Fi and resembling the ratio of the F^
generation exists.

If the full flushed purples yielding offspring free from streaked
forms be lifted out of Table IV and separately studied, it will be noted
how very much akin in their behaviour they are to certain of the dark
maize purples investigated by East and Hayes. The production of
flusheds to non-coloureds in the ratios 3 : 1 and 9 : 7 indicates F^ mono-
and di-hybrids which are functions of two complementary colour-making
factors. When also these data are taken in with those of the remaining
full coloured plants in the F^. generation, 4 pure purples out of 27 give
further support to this idea. Nevertheless, as East and Hayes point
out in a very similar situation \ as soon as these individual ratios are
taken in conjunction with the main mass of data, it has to be recognised
that " tri-hybrids and tetra-hybrids are possible which give such results."

^ East and Hayes, p. 85.

I

i

A. St Clair Caporn 275

This is particularly conceivable when some of the relations of the
streaked forms are considered.

In maize certain 'splashed' purples are found; but "they are zygotic
variations which are not inherited, for their progeny are exactly like
the progeny of the dark purple seeds. Further, the patches are not in
a regular pattern nor does the selection of seeds of this nature have the
slightest tendency to fix the phenomenon as a separate character^"
Evidently, then, in wheat the streaked plants, many of which are
capable of being bred true, are of an entirely different kind.

But there is another colour type in maize with which the appear-
ances of structural affinity are in some respects quite well developed,
though only in some. This type is known as ' particoloured,' the colour
being a very faint purple. If we imagine such diluted pigmentation
localised in one or two areas, and may be occasionally a little intensified
owing to this concentration, the ' streaked ' form results. In view of
certain similarities in the behaviour of their segregates which will now
be described, the visible differences between particoloureds and streakeds
are thus best likened to those between chemical allotropes.

East and Hayes describe the occurrence of particoloureds in two
crosses ^ In the second they aye designated ' light purples,' not because
they are unlike particoloureds in outward appearance, but presumably
to emphasize an important difference in the gametic formulae.

Where two factors, P and a colour factor C, are concerned in the
production of colour, it is held in the first cross the absence of C from
the combinations ccPP and ccPp does not prevent the formation
of a very small amount of colour. Hence the particoloureds. The
ratio of 9 purple : 3 particoloured : 4 non-coloured in the F2 genera-
tion lends support to the theory. Moreover, if allowance be made for
the fact that "particoloureds especially when non-starchy are not
always distinguishable fi-om whites," the agreement is fairly good. It
is better in ^3 if pure starchy groups be chosen. Again, ratios of
3 purple : 1 particoloured, 3 purple : 1 non-purple, and 3 particoloureds :
1 non-purple, as well as pure purples and particoloureds, occurred in
the ^3 generation. All these facts uphold the theory. But there is
clearly a weakness when it is stated" later that pure white races,
apparently devoid of any inhibitor of colour, may also have the gametic
formula cP. Moreover in yet another family, wherein it is alleged

1 East and Hayes, p. 68.

2 „ „ p. 67, Family (8 x 54) and p. 81, Family (60 x 54).
=* „ „ p. 102 (foot)— 103.

276 Inheritance of Glume Length and of Colour m Oats

particoloureds of exactly the same nature are thrown, the selfed whites
gave one particoloured ear out of every four in the next generation ^
Now this last is a feature which marked certain non-purples in the
second cross. One is inclined to ask whether there is really any difference
between the particoloureds of the two crosses. Surely in the circum-
stances the cP hypothesis cannot alone account for the particolouring
in the first maize cross any more than, as will shortly be questioned, the
action of an inhibitor can alone be responsible in the second.

Here, however, a useful interpolation may be made to note that
nearly all the above segregations of the first cross, with the exception
of the 9:3:4 ratio, are also to be found among the results set out in
this paper. It is even possible that the ratio 9 flushed : 3 streaked :
4 non-coloured is present too; for if the two groups in Table IV
throwing all three colour types be added together, then comparison
with expectation is as follows : —

Observed. 163i^ : Z1S : blN.
Expected. 141i^ : 47>Sf : mN.

Bearing in mind the small number of plants, one can hardly say the
discrepancy is very much worse than in East and Hayes'

Observed. 638 purple : 210 particoloured : 306 non-coloured.
Expected. 649 „ : 216 „ : 289 „ ^

To return to the second maize cross. Here East and Hayes hold
that the cP seeds are whites and that the particoloureds develop when
the colour inhibitor I undoubtedly operating in this cross is present
heterozygously along with P and C. The fact that in this case no
particoloureds breed true favours the conception. But if this be so,
why do quite as many of these light purples (particoloureds) yield ratios
of 3 particoloured : 1 non-purple and 1 particoloured : 3 non-purple as
give 1 : 1 (possibly 9 : 7 or 7 : 9) ? [Note that the first of these is also
thrown by streakeds in the wheat cross, but not the others, which come
from non-coloureds.] Also, unless the non-purples lacked P and C,
they must by the theory be homozygous for I and therefore incapable
of throwing particoloureds. But this is not so. For in two cases the
following proportions were found ^:

' East and Hayes, p. 78.

^ East and Hayes, p. 72. Table 21 A. Starcliy seeds only, i e. the most favourable
Eelection.

» East and Hayes, p. 100. Table 25 E.

i

A. St Clair Oaporn 277

Non-coloureds throwing

80 particoloured : 294 non-coloured
80 „ :216

160 „ : 510 „ =1:3 (^""^ 3 : IsV

\ or /

While another had

105 particoloured : 204 non-coloured,

which is probably equal to 7 : 9 as the classification would err on the
side of the non-coloureds \

Again this is a point of similarity to the wheat : the non-coloureds
shown in Table IV behaved in this manner.

In brief, then, the ' streaked ' and ' non-coloured ' aspect of the
wheat cross in its relation to the same in the two maize families can
be synopsised thus : —

Resemblances to 1st cross. Streakeds (particoloureds) throwing

All particoloureds.

3 particoloureds : 1 non-coloured.
Resemblances to 2nd cross. Streakeds (particoloureds) throwing

3 particoloureds : 1 non-coloured.
Non-coloureds throwing

1 particoloured : 3 non-coloureds ( 3 : 13)

i „ + 1 ,, (approx.).

The outstanding point is that in one and the same cross (wheat)
the inheritance of streaking gives support to two separate, radically
opposed theories on the transmission of particolouring (maize). The
conclusion is either that one of these two schemes is incorrect and the
other as yet imperfectly elaborated, or that both are incomplete ways
of expressing what may be in reality two kinds of particolouring de-
pendent on the same system. The evidence of this paper seems to
support the latter idea. For as at present expressed by East and
Hayes, although each does so in part, neither hypothesis by itself will
fit the wheat cross completely. On the cP hypothesis the origin of
streakeds from non-coloureds is left unexplained, while the full flushed
purple of the ^i grains and the failure of F2 non-coloureds to throw
flusheds are against the presence of the usual type of inhibitor. To the

1 East aud Hayes, p. 96. Table 25 A.
Journ. of Gen. vii 19

278 Inheritance of Glume Length and of Colour in Oats

writer it seems not at all impossible that streaking is dependent on an
entirely different factor (or factors) from flushing, and that apparent
reversal of dominance in the non-coloureds throwing streakeds may be
due to an inhibitor affecting 'streak' alone. For I cannot inhibit
' flush ' as in maize, because, as has just been remarked, no non-coloured
forms throw flusheds, nor do any streakeds \ But whether or not this
same factor I affects flushing to any extent in an indirect manner is
not clear. It plainly has no visible action in the Ft^ generation, where
the purple flush is fully dominant; but the remarkable deficiency of
full coloured forms in the F^ generation is a puzzle to explain otherwise.
That, and the lack of any F^ segregation with the F^ ratio, is the snag
on which every theory so far launched seems fated to founder.

Stabilisation.

Several pure lines have been extracted during the experiment.
Most are full coloured purples with long glumes. That no pure long-
glumed flusheds appeared in the F2 generation is to be attributed to
the very small number of pure flusheds produced rather than to any
coupling of ' short glume ' with ' flush.' The glume distribution in the
intermediate No. 20 disproved the latter suggestion and no difficulty
was experienced in obtaining them in numerous F^ cultures. Outwardly
similar, genetically they are very varied. Three pure flushed lines
which were bred from the \'2F : SS : IN group, for instance, may or
may not be alike. They may give ^1 mono- or di-hybrids on crossing,
they may mask a streak even as a black oat will contain, but not show,
hypostatic grey, or they may be pure for the flushed character alone.
The same remarks apply to the pure purples of the 27F : 9S : 28iV
section in Table IV, although no ^''4 cultures of them were actually
isolated in quantity.

No streaking occurred in the SF : IN and 9^ : *7N groups, and,
assuming that in both cases the colour is due to the interaction of
the factors C and P, then the extracted pure purple races must all
have the formula CCPP.

As for the various pure cultures obtained from No, 20, all that can
be said of them is simply that they are pure flusheds, but whether alike
or different and of exactly what nature can only be decided from their
behaviour after being crossed with whites of known constitution.

Pure streakeds, both long- and short- glumed, have also been stabilised,
and lastly non-coloured Elobonis.

^ An extremely doubtful exception which may or may not contain a stray, and of
unknown parentage, is No, 34. But see remarks on p. 273.

JOURNAL OF GENETICS, VOL VII. NO. 4

PLATE XIII

n

JOURNAL OF GENETICS, VOL VII. NO. 4

PLATE XIV

a ^

Cambridge nnrv«r»irty Press

A. St Clair Caporn 279

SUMMAKY.

In the first part of the paper the parents are fully described ; an
accurate method of measuring and tabulating glume lengths explained ;
and it is shown how in the F^ generation a marked change in the
average glume length of homozygous ' longs ' as compared with the
average of the parent, T. polonicum, under equal conditions persists
right through into the ^3 generation.

The second half is concerned with the purple pericarp colour, — the
manner and extent of distribution in the tissues, its development and
detection in minute quantities, and the inheritance. The latter is dis-
tinguished by one cardinal and unaccounted for anomaly : segregations
analogous to the F^ segregation have not been found in the F3 gene-
ration. Streaking, a character which suddenly appeared in the F^
generation, has resemblances to particolouring in maize. These are
discussed.

Various cultures have been stabilised in respect of colour and glume
length.

EXPLANATION OF PLATES.

PLATE XIII.

Fig. 1. Ear of T. polonicum. Spikelets in frontal view. Natural size.
Fig. 2. „ „ „ lateral ,, „

Fig. 3. Ear of T. eloboni. „ ,, ,, „

PLATE XIV.

„ . ^ . , rr, , . T ( a. 32 X 4i mms.

Fig. 4. a, a'. Extreme forms of T. polonicum glumes ^ , „, .
° (a. 21x4 ,,

b. Typical T. elohoni glume. 10 x 4 mms.
Fig. 5. An Fi ear. Spikelets in lateral view. Natural size.
Fig. 6. a. 'Long' type of intermediate glume. 17|x4|mms.

b. 'Average' ,, ,, „ 16x4^ „

c. 'Short' ,, „ ,, 14^x4 ,,
Fig. 7. a. T. polonicum grain, llj x 3^ mms.

b. Fi grain. 10 x 4 mms.

c. T. eloboni grain. 9x3^ mms.
Fig. 8. Colour types.

a. T. polonicum grain \

b. ' Flushed ' grain I vt i. 1 1

, , . > Natural colours.

c. 'Streaked gram I

d. Faintly 'streaked' grain J
b', c', d'. Acid colours.

19—2

280 Inheritance of Glume Length and of Colour in Oats

Fig. 9. Portion of transverse section of a purple grain showing pericarp and adjacent
tissues. X 190.
^Z = aleurone. Ep—e^\CBX^. (J = girdle cells. iH=me8odenn and hypoderm.
P = perisperm. <S = spermoderm. r= tube cells.
Fig. 10. Superficial view of girdle cells in part overlaid by epicarp and mesocarp after
removal of the underlying tissues, x 190.
^= False sap tints, — tannin yellow, dark reddish brown, and pink.
Z= Colour content of an extremely minute streak.

LITERATURE.

BiFFEN, R. H., J905. Mendel's Laws of Inheritance and Wheat Breeding.

Journal of Agricultural Science^ Vol. i. pp. 36, 37.
East, E. M. and Hayes, H. K., 1911. Inheritance in Maize. Connecticut

Agricultural Experiment Station Bulletin, No. 167, pp. 57 — 104.

REPORT ON TESTS OF SELF-STERILITY IN PLUMS,
CHERRIES, AND APPLES AT THE JOHN INNES
HORTICULTURAL INSTITUTION.

Prepared by IDA SUTTON,

Student in the Institution.
(With Plate XV.)

The work to which this report relates was begun in 1911 by
Mr W. O. Backhouse and Mr M. B. Crane. It has been continued in
each succeeding year, and among others who took part in it must be
mentioned Lt. Sherrard, Prof. Kusano, Prof. Chibber, and Miss Garlick.
My responsibility in connexion with the observations began in 1915.

Preliminary reports on the earlier stages of the work were made by
Backhouse (1, 2, 3). It is greatly to be regretted that this report had
to be prepared without consultation with those who preceded me in
the work, most of whom are absent on foreign service.

The objects were in part economic, for since the work of Waite and
others it has been recognized that failure of fruit crops is not unfre-
quently due to self-sterility ; and in addition it was expected that the
work might elucidate the physiological nature of this phenomenon. At
the present stage no comprehensive discussion of the genetic problems
can be offered. In view of the recent experiments of others two main
questions arise, (1) whether self-sterility is a simple Mendelian recessive
character; (2) whether the older observers were right in considering
that in such cases self-steriles are fertile with the pollen of any other
variety, or whether there are not, rather, several classes of individuals,
between which there is what East has called " Cross-incompatibility."
As regards the first question there is nothing in our results which
negatives the view that the property of self-sterility may be a recessive,
but until a later generation can be tested, the only evidence bearing on
this aspect of the matter is the fact that the results with plums and
cherries are consistent with the supposition that the plants consist of
two larger classes, self-fertiles and self-steriles, with a smaller number
of plants of intermediate properties. These and presumably some of
the self-fertiles may be supposed to be heterozygous. The self-sterile
class forms a fairly homogeneous group, and the occasional indications

282 Self-Ste7'iUty in Plums, Cherries, and Ap2}les

of partial self-fertility are probably attributable for the most part to
errors. When, for instance, in the case of Coe's Golden Drop fertilized
by Jefferson, the record shows a few fruits formed, out of a large
number of flowers pollinated, the fact may mean that compatibility
exists in a very slight degree, and if this could be confidently asserted,
it would be tempting to suppose that the tree may be mosaic in that
respect; but conceivably some error may have occurred. Between
these interpretations we have at present no means of deciding, and the
facts can only be given as they stand. The characters of the seedlings
will in some cases hereafter provide a criterion.

As regards the question of cross-incompatibility the work was begun
in the expectation that this was the only acceptable account of the facts.
The older belief implied the paradox that there are as many categories
as individuals; that, as a friend put it, "each lock would open with every
key but its own." Nevertheless the results obtained have not favoured
the theory of cross-incompatibility. Of clear incompatibles we have
found four only, Coe's Golden Drop, Crimson Drop, Coe's Violet and
Jefferson. But Crimson Drop and Coe's Violet are known to have
arisen as bud-sports from Golden Drop^ and it is not altogether impossible
that Jefferson had a similar origin I In apples no such case was seen.
Cherries have not been so fully investigated and some records of failure
with certain pollens may possibly be an indication of incompatibility,
but pending repetition and confirmation their significance is doubtful.
We are thus driven to suppose that at least in plums and apples the
former belief is at least approximately true, improbable as this may
appear.

The evidence from apples is not strictly applicable in this connexion,
for as our records relate primarily to the formation of/r'mfe the power
of the pollen in producing seeds is very imperfectly recorded. In plums
and cherries well-formed fruits with rarest exceptions contain developed
seeds, but in apples seedless fruits are common, and indeed in Duchess
of Oldenburg and Antonowka are almost the rule, nor does seedless-
ness necessarily involve any diminution in the size of the fruit (e.g.

1 As to the origin of Coe's Violet, see 0. Thomas, Rev. Hort., 1870, xui. p. 515.
Mention is also made of the independent occurrence of a sport, presumably identical, in
Trans. Hort. Soc. Vol. vi. 1826, p. 393. Another violet sport was observed by Mr Vert
of Ruabon {in litt.). For the origin of Crimson Drop see R. H.S.J. 1901, xxvi. p. ccxxvi.

2 Hedrick, U. P., Pluvis of Netv York, 1911, p. 256, writes : " The parentage of the
variety is unknown ; Floy thought it was a seedling of Washington ; Elliott suggested
that it was 'from a seed of Coe's Golden Drop, which in growth and wood it closely
resembles.'"

Ida Sutton 283

Antonowka) though in Lord Derby, for example, such diminution was
observed.

In work of this kind where the measure of sterility or fertility must
be the proportion of fruits actually set, it is obvious that many circum-
stances must inevitably blur the accuracy of the final result, for in
almost all varieties only a proportion, generally small, of flowers can in
any case succeed in forming fruits. The allowances to be made for
these irregularities nmst depend largely on personal judgment, expe-
rience of each variety, and details Avhich could only be recorded at great
length. The classified lists appended are therefore given as representing
approximate interpretations of the results recorded in the Tables.

The stage at which fertilisation fails probably differs in various forms.
Mr Collins has made preliminary investigations on this subject. In
Goes Golden Drop, typically self-sterile, he finds that whether the
flowers be selfed, crossed, or left unpollinated, the style is definitely cut
ofi" by an abscission -layer at its base on about the fifteenth day after
the flower opens. The plant's own pollen germinates on the stigmatic
surface and tubes penetrate the nutrient stylar tissue, a phenomenon
seen by previous observers in other self-sterile plants.

At the beginning of the work the flowers were bagged on trees in
the garden, but from 1912 all experiments were done on trees grown in
pots in an orchard-house. This house is provided with mosquito-bars,
and various other elaborate precautions were taken to ensure the exclu-
sion of insects. It is believed that errors from that disturbance have
been almost w^holly avoided.

The numbers of flowers given are in general accurate, but some of
these numbers were obtained by estimation based on the number of
cor}'mbs used. The word " em " means that the flowers were emascu-
lated. In the absence of this expression, the anthers were not removed.
All flowers were hand-pollinated. The number of flowers tested is
given when a record was made. At first this was not always done.

The trees, with very few exceptions, were in the finest condition,
being for the most part in exhibition state. The horticultural treatment
was under the continual supervision of Mr Allard, the Superintendent,
and great credit is due to him for the fine health in which the plants
were kept, without which such observations would be altogether un-
reliable.

From the horticultural point of view the conclusions to be drawn
from our experiments are (1) that many important commercial varieties
set little or nothing unless cross-pollinated ; (2) that for the pollination

284 Self- Sterility in Plums, Cherries, and A^yples

of these self-sterile kinds pollinisers must be planted ; (3) that provided
that a variety produces plenty of pollen and flowers simultaneously with
the variety which it is intended to pollinate, any variety, at least of
plums and apples, will probably serve for this purpose, apart from the
special case of the Coe varieties of plums and their presumable co-
derivative; Jefferson.

The paper by Stout published in this volume gives (p. 102) references
to the more important papers on the physiology of self-sterility. To
these may be added the following, relating for the most part to horti-
cultural observations made in this country.

REFERENCES.

(1) Backhouse, W. O. Rep. Brit. Ass. 1911, p. 599.

(2) . Gard. Chron. Oct. 28, 1911, p. 299.

(3) . Gard. Chron. Nov. 23, 1912, p. 381.

(4) Berry, G. P. ""Investigations in Cherry Orchards." Journ. Bd. Agr. Vol.

XXIV. p. 288.

(5) Chittenden, F. J. "Pollination in Orchards I. II. III." Journ. R. H. &'.

Vols. XXXVII. p. 350, XXXIX. p. 366, p. 615.

(6) MiDDLEBROOKE, W. J. " Pollination of fruit-trees." Journ. Bd. Agr. Vol. xxii.

p. 418.

(7) Waite, M. B. " Pollination of Pear Flowers." U.S.A. Dept. Agr. Div. Veg.

Path. Bull. 5, 1894.

Self-Sterile
Coe's Golden Drop ^
Coe's Violet
Crimson Drop
Jeff6rson
Bryaaston Gage
McLaughlin's Gage
Early Green Gage
Old Green Gage
Reine Claude d'Althan
Pond's Seedling
Wyedale

Frogmore Orleans )
Late Orleans )

Prune d'Agen

Primate (small numbers of
Flowers selfed, but no
other cross on tree)

Frogmore Damson

PLUMS.
Partly Self -Fertile
Rivers' Early Prolific
Early Favourite
Cox's Emperor
Early Orleans
Farleigh Damson

Self-Fertile
Denniston Superb
Early Transparent
Reine Claude Violette
Golden Transparent
Reine Claude Bavay
OuUin's Golden Gage
Belle de Louvain
Monarch
Prince Eugelbert
Prune G^ante
Gisborne
Victoria
Czar
Pershore

Yellow Magnum Bonum
Early Mirabelle
Myrobalan Red
Belgian Purple

Ida Sutton

285

CHERRIES.

Self-Sterile
Amber Heart
Black Eagle
Black Heart
Black Tartarian-
Bigarreaa de Schreken

„ Noir de Guben

,, Napoleon

,, Jaboulay

,, Frogmore Early
Early Rivers
Elton

Governor Wood
Guigne d'Annonay
Kentish Red
Toussaint ?
Waterloo
White Heart

Partly Self-Fertile
May Duke
Archduke

Self-Fertile
Flemish Red
Late Duke
Morello

APPLES.

Self-Sterile
Beauty of Bath
Cox's Orange Pippin
Gascoyne's Scarlet
Lane's Prince Albert
Norfolk Beauty
Northern Greening

n ■ . r. ij /Very small
Grimes's Golden ) number of

Yellow Newtown 1 flowers

I selfed

Partly Self-Fertile
Worcester Pearmain
Bramley's Seedling
Lord Hindlip
Lady Sudeley
Old English Broadleaf
Doucin

Improved Doucin
P. prunifolia
P. baccata
Golden Russet

Self-Fertile
Antonowka
Annie Elizabeth
Baldwin
Cellini Pippin
Coronation

Duchess of Oldenburg
Golden Spire
King of the Pippins
Lord Derby
Red Winter Reinette
Ribston Pippin
Stirling Castle
Sturmer Pippin
Washington
French Paradise
Crimson Bramley

286 Self-Sterility in Plums, Cherries, and Apples

s

3 i 1 I I I lllll I I \ \> \ ISSI lo|ooooooj,-| I

fi

>>

00 § o o

^ K;

X iS O tH
S w (jq

o> «o

: a

43 •£* - «

p. I I I I I I 43 _?-

cS'l I I I I I

O fl

Q -S « 3

» 2 1^ 2 "S -n I ^ ,S -S

|>0t-00 l«DOOcgt^O§000:^

>.

p

a>

■a

O

_M

■H

TS

02

.03-2 5^ • ^

02

I 111! ^ ^JiS-^-l i lili ^%l\^l -2 I ^ ^ "11 -'1.2^1

"rtQ^pq gp^OWOSpHP-i^pH p^H:]aHO o >^ p^S pqpqAnO

•iH

S o o . o o

I g s ^ ""

o . .^

ca ' cQ

o o

Ida Sutton 287

a

r-IOlOiSool I loW^O*^*!'^ ' ' It-I lOOlOl I I I I l«SOO(N(M

(N(MlrH^ (M(Ml-l'-'t-l >Cl-lr-l ^lOr-llHi-H

o
a

:::© .•::::o '-S

Q Q t3

C > <= O . «' „

fl o fl I I c o .2 .S

= I

ai ..

U> &C

M s

Q

d'Agen
Empero
gh Dam

c fl "

in c3 >4

Golden

g 00 a>

CD

t*s >. -TJ ;

5 '^ "3

<D

M^O

f^ O P^

o

Pi

IB

(3

s .-. "

"^ ft

>> ^ -^ a. scflflo „

Jooooo |o;^M«|oo I I |mg|||=?^|>g=^ocoo^oooooo I

s

— >: >: t> > f> >

iH „ i-HjH rHlMWi-ICO

• a

t^= «^ -SO ^-^ ^ '^fl®|fl

•S « S I j>>=3^§§ S-S ^.2 0.^ § 2 g 2 §

f \

I ^ o o

: ^

• O

r'S

"aj

a

«

, !>H

- fl

fl

ji

-S

C3

d

- .a

-J

.^

- o

* o

~ 00

O

S

s

o ^ cl

> 2

« o

S

S -I «

O O Ha

288 Self-Sterility in Plums, Cherries, and Apjfles

O O O ^ i-H

>-.

t>^

>^

a

a

a

n1

as

(T!

CO

S

o

CO

"*
•^

S S

S<) O t- U5

^ §

■* »o I CO t-

i-t CO l-l (N

I I I I I I I

>« S 5S "-^

us 00 JO t~
(M 00 TJH

•^

PI ® is

El O H S

0) !> TJ

08 ^

P4

o

Q

a

fl

<

-o

ri

o

o

a

iti

a

o

MO >^ d^ o

" « a g

•2 ;2 ^

O 03 o

OHO

!«

O

C5

O P^

Ph

O

o

o o

<s S a I 2

o OJ

O <M 05 O

CO t^ <>> O O

1

■^

o

05

T-l

•^

(T)

00

o

o

1-1

o

CO

00

-*

-*

ICI

CO

<M

IM

CO

Fm

>,

•

■

*

•

P^

«3

•cs

:

'o

ri

fe

1

Ph

03
60

o

T!

^

>5

C3

a

P5

fcH

01

s

a

60

<

0)

a

ft

a

f

>

=

2

Ci5

a

o

_6p

'3

O

:;

o

S

>

T3

*!

a

Si

c3

,>^

O

Ph

O

1-5

Ph

^

N

I 9

pa

IIIIIIM^SSIIII

Ph

pq

P5

P5

60 6p c

CD O ^3
a g S

Ph

H

c5 S;^

S '^ o

a TJ >. S

•S -S

Ph O pq Ph t> Ph

I 5

0*0(8 0

^s

CO ^

5 00 00 ^ ^

fe !>>

S 2 §

-Bo

k2 00

a a

S S S_

3 --I o

S S

03

ta

OS

o

Ph

08

I— I
o

'a

P3

e3

Ph

2
C5

C(3

C5

S3
c3
•J

03

Ida Sutton

289

1112

s s

1^

> I I I I I I I I

[i< S l>

(N

I I «> rH

I I Oi 00

I I

i I I "^ I I I I I I I I I M I I

I I

c3 «

JS o

ec

a CD

« i

a

a> !3

^.

toil

1 1 o ^

1 °°

^ §

■» J^

TS

y >>

"v H

0

«a »?

O «3

o

H pq

o fe

&^

^5 -a

.2 o

> O

w w

o o

I I I I I I I I i

13

a

O

rt

2

OQ

O

«

T)

O

O

^

o

6

«

o

>

•iH

o^

: r^ s

m

w

a

leans
yanst
ine C

o

U U V

Plh

o m rt

lO "* tH ^<U i-l -*

. . t^ >i >, . . .

l§lg'o'H§§§1'B'B

[i, g CO o

^ ^ iH (M
® .^ C- C-

PM 2 rjt rH

I r.2 I I i I s I I I I I

I I P^ pq

I I I I I I I

I I §

9 o

o a>

« «

8 3-

H « 2

-» -g I

o -H »-i

O M fe

a O

» s

O H?

: O

O

O f> ^ fM d^ o »^

.2 "o
> C5

O Q

S

-_ c

I I I I I I I i ^

0) M ." 0) c3

^ S

6C ^

S pLj fii m pq fi4 O

O i-t o o

ft o

?o O §

«o CC O O OJ o

S S

^ H

::) o o

iH «

eg
C5

CS

O

o

n
5

I

o

>»

290 Self -Sterility in Plums, Cherries, and Apples

a a

^

^SS §^

(M ^

S!>>So>5 5^f^

d

o

D • .

1

^ =
O

0

fl .

a

1

2

C5

>^ 5n

rt ^0

0)

o

"^

W ►^

tj

^■>

s ^

o

^

•3

(4

«5 tH

05

(M

(N r-l

O O (M -*
t> t- CO <M
S<) lO (M «

a c I' 3

SD -fll '^

I § § »

O 3 3 53

>H »H ea

(^ (1| Pm

I I I I

a
a g

a

a §

g? 2 >•■ g > ;^ >■* S >:

PtH

I I

^ 'i '^ a S Ts

. O O ^ eS . O

> 1^ s ^ IS p> a

I I

o ^

ti

a

S P-.

0)

=3 "S

a

'3

S

a;

O
>>

c3

M

O

c>-,

"-t

^

o

k4

Pm Ph

H

^J

PL,

o

be

00

a

a

.2 W

c3

a
o

a

0)

a

2

a

S

o

o

O

!=- "rr

c3 2.

-2

2

F

O

>^.

on

." s>

^3 rn

a

>

s

O

c8

o
O

O Ph

^ o

Q fiH S'

o H cc
C5 %. «

o .«
O P5

■^ i-l 05 U5 CO

>> >5 >J >^ >»

a a a a a

05 oS CO cd CO

l^gjo I I Ig^g^^o

J^

f> f> > t> p>

S Oh-

^ : (^

(U

«

a 03

a

c3 Q

^ 2

60

<u ^- ^ a ^
io >- M a >H

►^ H O Plh Pm

Us

^

2

« ^

g

a

eS

s*-

a>

N

•U'

u

«

o

§^

t

T)

o

o

rt

A.

I I

o

a ^

oj 60

^ a
>»;a

2 ta S ^

CJ

a o H oQ
^ !tt oj > a

H Hs U rt PlH

::^

56 §1

II

I ^ I I I I I I I I I I I

m

O O o <N iM

i-H tC d O

<X> 01

D3
Pi

o

.15 >^

§

P4

Pm

»

>s

t/2

Si

X

a

o

M

o

d

eg

>

too

53
O

S

tJ

«

a>

ca

T3

CO

v

ns

^

a

n

^

W)

a

•s

H

o

^

o

g

Ida Sutton

291

00 (N t- « O

^ >S <-" t>

C<l S f-H i-l

CO <M U5 1-1 WS « O

r- iH t> «o »o >o

■^ 05 CD O "-I t-
r-l -* CO IN -* "H

O 00 -* OJ

0«D»OiMOt^t~00>0 I

Oi«0»0>-Ht~C-CO-^JH-c^O I
i-H i-H (M e<I r-H CO i-H

n

5C

o o

u u

ft P^

o H
d3 -

: o 2 o

■ >-> c3 &:

Q g ft

. <I> O 05

: 2 2

c 'o 2 'o

o O o j^

(u ID " Cab -

O Ph ►^ O Pq O

f^ (30

S _o c

S a
ft ^

£ a a -s » „

g; c3 a> ^ 8 O)

® rt

e3 OB

C5 a

a g

tS 03

^ "J

o S

2 i §

&I lU

a .s

O H

g S b (1h

a <ii u .

a> '-* a >

'^ r< ^ \S O

2 -SjH Ph o

'^ (-1 " "^ a
2 .-M H « §

o "^ "« "3 !§

O H »^

o

§•

d

rt

<u

<u

a

(i>

O

2

>»o

3

-o

H

O

I I I

I I I (N 05 O
I I I CO IM

t^ 00 O IM O

I I I I I

I 1 I I ^ I

I I I I I

CC <Z> ^ "^ y-* It-wSCOcO I I I I
Tj(r-I'<1<<>'IN Q0<O«3«5

I I t

I.

i^ I I I I I

ft t

I I

Qj Q) S g; QJ a

I I M I I I I I I I i I

I I I I I

1^

C5 w

la

& s 3

H o a

a S 5 5 ^

t» O 00 o

C5 o

I 1 I ^

s

I -S

1-^

<M O O O

iH o o o o

i I 1:2;

s

t» CD (M U5
«5 -* Oi O

iH 00 lO (M

S

I ^ I I I

1-1

S

a

[S4

US

08 00 CO O

>J >5 P>1 >>

a a a a

c8 CO cd c3

s s s ^
p^ >^ t> >^

o4

■* t» O t» CO

CQ Tjt CD o eo

I I

S

rO to

Si

1

O

Pi

Pli

fP {1.

s s s s

B
: S

• -a

O P4 >H

(4

a

o

s

292 Self-Sterility in Plums, Cherries, and Apples

T)(OCD050<M<M«5C<1

O C<1 ic W5

I 1 I I ^ ^ = ;2! ■=> <=^

"3 1-1 CC IM fO >Ci «

1 1 I

O o ns O >-• r-l
N C(5 «0 >0 X5 (M

a "S

'^

SO

9 fl

O 2

S3 &

o o

B >-. I»( O

0) i 1 :h;

o

o

I I I I I

a. ^ Ph
(3 "^ g >>

S m ^ H

O 2

-^ a S

,i«) •?

6C OJ 5 «

T3 ;z; fc( O ^ -T3

rfl . . ■« O .

^ S 3 ^ S S

5~

cd cS ^ cS cd

CO 03 CO w Co CO

s s § s s g

> > > >

I I

h

O IM 05 I rH

o o o

S 00 <M

Oi-lr-|IMIMlMi-HrH

« ec S !^ -^ 'x

^ r-l <M S S «5

Ph

Q Soft

a A >?

O

?0 ® -« > a s ^

•-H K t4 tn

1-s O O Ph P^ Ph O

a 5 I

bo ^

!§ a

CO ki

^ PM

a a

^ :^2«s

<B o a '"'

13 ^ fc! '^

ea eS O

2 -a

<X> 1-4

b o

o Q 0) S
^ •:3 '^ (^

O cd

Ml.> S

CO o O 5 OC 00 CO

m S o ph S S ^

Ph
" pq pq pq H

:^

s 3
§o

a o) a)

<'^ -^

0) o W
a ^ >^

o s s

1^

a-S

W) ti)

te iM lO o o

O 00 t- CO 00

03 fO

s

s

s

«2

P4

a

OS

n

too

1^

a

•

CQ

o

09

o

H

d

:=!

t-i

n

o
•S

^

o

S

H

2

>s

»

^

o

Pm

»

f3i3

-fi

<u too

cri

a> >ij

^

-c «

o

S^

ni

<{M

m

a
•c

eS

m

^

eS
too

g

toocjj

Ida Sutton

293

i-l -H «3 •*

O Ci CO O

I I

a

(MiHTjli-IOSeOt-OO

t~ i-l 00

o o o so

00 e« 05 <M

i I i I

?OC^»OCC«OiOOsOO
iH i-H (M crj

H

r8 5 ;z3 1:1;

^ ^ O) <c

2 S o o

<; <fj S §

■ S ^ V

« 0 w

rl

>»!z; ^

0

^ ^i^«

y

w S, m

2 >H TS

fi oj ^

2 «^

S ^ -2

(i ^ 1

m ^ «

t I i. I I III

m

>H ^

PQ

r^g o pq

1^ -2 ^

:^ « :5

.SP S o S S .SP cs o
eqMgH<ipq^S

-*C<)C(5iO«0-^COC3cOOOOi«0

a -^

o fi

CO CI;

a

cS O O ^ O
^ CC I-H

3

II 1 1

> ;=;

flS o o •^

^ J(5 t- to

I I I

H

.5 3^

« 3

O

"^ >» 5" 2

a « S W H S

: ^ r^ a 5

j«; 03 M) ^ W
.2 I Q ^ [^ ft [S

I I I

£ ft Q

^ >.

w

Mil

H

cS

0

ci3

0

<X)

a

E3
0

0

tao
0

z

fo

C3

S

13

c3

03

eS

(U

a>

a> >)

^

^

h

^

eS

eS

oS c3

tax)

tUD

bApq

CQ

00

n

t

P^

^

o
o

H

Journ. of Gen. vii

20

294 Self-Sterility in Plums, Cherries, and Apples

■^io-^«ooosco-^««o

IN CO t- fO

O «0 CO o

te iM

P4

rt(000'®OCO(M-^"SO O OO
CCOiOO-^QO-^OOiftt- lO (NO

eo iH Oi 05
CO C- 1-H (N
1-1 (N C<1 i-l

05 "J lO 5

.2 ^

0)

Q .2

■ 5

,5 5P

H5 pc.

03 §^ «.iyo ^oo

.1-4 f-4 '^ k_J

w w c5 ^

H

s

fl

-+^

o

s

Ti

Ol

CD

s

0}

P3

,d

s

o

60
' '3

o

pq

i-H to i-H IC CO o o

CO -«J( [-. O CO

-* 00 05 "* CO

S S

I I

rH tH CO CO CO
t- CO 05 CO Tj(
IN CO CO IM IN

PH ' Ph

Q ;S .2 fi P5

>^ 2 a ^ g

g S PL, Hi S

^ i

O H

: ;^

C4

a

3ll0

ish
;mor

1

0.2

o g S 2 "g .2

g M O O J -^

*-5 >s *^ 'S I^

• 2 >. S .

60 5 S^ <B 60

S O ^ S M

60

60

.2 ft

5 P5

M ^ S M S H P5

'^

■^ S S ^

-* '5 05 o

00 o c- o>

^ (N CO

SQ

Ic

1-H 05 C-

O I le-< I 1—1 CO
i-l ^ i-l iH

Ph

03

It
go

£ S)
> .rt

O P

CJCB

73

a

a>
•M

M

^

E3

0

n

Q

<u

>,

^
S

o
9

^

?

a

5 w

^ 0)

cs

1:53

H

^^

Ida Sutton

295

I I

o 00 1-1 o

+ 0> 00 o o

^ +

1-1 CO

o » o 00 e*i 05 «

a

o _ -t-

I— I I— I »H •— I 1-H r-< CO

I I 2

* I 00 t^ .H
g I 00 1-1 r-l

Oh J :=!

c

S£

J

Wh

C3

02

1 03

o

>i

u

a>

( >

3

s

_00

pq PQ

S o o

I g I I

^ >C O IN
CO lO (M iH

•s '^ S >»

02 _g _C -f^

c

^^ Oi

?5h

Oh'

o

<1 o

a
•a

O w
■ S 03

s

^

m

s

^

a

«o

o

r--

o

>.o

(T)

o

r^

1

«»

Ol

1

us

i-t

CO

■*

§

fH

rH

r-

1

on

w~^

1-1

CO

i-H

eo

C^

1-1

1-1

<2

Tl

ci

-a

OD

0)

.9

a

P4

p<

(U

a

&<

o.

^

^

^

S

fr

<A

0)

a

as

:;

^

p:i
o

3
OS

m

o

a

Oh

1

^
s

a
o

a

QQ

a>

i

a

o

a
o

O

O

3

2

130

00

Q

T1

o

no

g

0)

a

n1

^<

;-•

fl

O

-

•^

a

<1)

o

rri

O

O

^.

<

O

<

«

:z;

i-:i

^

rt

o

O

+

g a s o

32 02 02

1-1 05 CO 1-1 tH O

t^ iH CO O -* OJ

I I I |,-(00oa>^-^-=0'M■*lO■*

S ft s^

= s a £ '^ ■" a

COcOCD>OC<li-li-ICO^

^ -^ "v

^ 05
CO lO

©q (N

I I O rH

t~ tH Its C- (M

MM

00 Cd

CS OS r-j
leH rH iH

03

-S a

04

3 *• '-3 ij

O ^ Q O

.3 <

■t: .s

>> CM

<v

Oh

5P %

3 (Q

o f=5

■ 3 a - a

tH cS O 3

P5 J O -?

CO .

J Oh

- .2 8

w I

.2 §

- 3 -o
' 3 .

< Oh

O 9

3

1 . . <o ® «8 S « i''

"S O

(B .-I

-PQ

O 3 0)
O <! P5

u a

PLh

M

5 3 ^ o a -a

;z; ^q ^ H^ pq o

t^ 05 >o X e<i «o o

o -^ ^ «5 ■* «o

s s

t- 1-1 t^

00 t~ 1-1
iH IM

03 :

<5

»

a
a
<

e3

Cm

O

m

a>

&4

a

s

dn

eg

A

S!

Sh

too

5*

03

o

o

:§

o
»4

CD

oa

a>

o

^

o

O

o

O

o

296 Sdf-Sterility in Plums, Cherries, and Apples

(1A

(M ^

iM

o

o

1 2

ee

lO

i-H

o

^"^

§

CO

o

00

S

i

s

g

s

ss

g

s

»o

o

CIS

1 t^

1 tr-

8

00

OS

1

0)

1

1

1

o

: P'

P<

a

<s Pu

§

3

a "S

tS 60

O .S

O M

a "^
o <u

h5 :zi

'^ -i

a .S

■a a

^5 M

Pm :-

3

a oi _
^ '^ 9
3 O &

<x) o

^ 05

0^

o -5

2 *

a %

3. s

5 S -5

a ■

a. -a

CO

HftccfiOoa CSOOh

. O a Q
O O 1-^

o

:; a

a

'ft .

S >.

a o)

tc O J

a o o
<J o ;z;

O

50 05 t> "^ f-* t~

+ O 05 O O 00

rH r-l <M >-•

O O rH -^ !?« »-l

+ + +

rH CO <M O?

^ CO >C
O r-l rH

13 r:! «3

fM ft

a TJ -" ^

a 2

o -a p^

.? = 2 S

5 £ i o I g^ ^

• a « i^c s : ^ a

> .9 g a 3 .£3 08

i^ o o M 5 ^ m

- a
o

a

fe

211 Mini

fi ft

O a^

a

'ft • :
ft

Ph

® 0)

60 *-i .j^

a cu ^

o _o o
O O i-:i

ft

ft

Xi

C

ft

ft

a

Ph

Ph

.^

1)

Tl

a

00

n

«

60

S

O

4)
tc

^

t3

0)

o

O

60

a

o

no

- J3

3?

ft

CO

a

tx)

»H

>4

(Tl

)H

ra

C

*j5

O

!-l

o

OS

O

02

w.

o

W

m

I-)

t-l

a>
"ft

CO

<

d

o

:^

o
00

CO

ra

>i

a

2

o

a>

^

o

-rs

u

CO

cs

c8

o

Q

o

O

Ida Sutton

297

i I I I I I -^ i

^ ^ ^. « ^ I I I I

+ +
th th e« -* >si

M O t- C3

I I I

«0 I t> r^ rH

lllllli|3SIS§^ig§ll|ll§iM|S§§llllglo|

r-^ en Qi iH THi-lrH i-(S'

Ph

m

o -5

::> a

"5 e i :S

.o A, "S 2
OS <c . . o o

U <1 fq (l< ftn ;zi O

m

SO

«? ^

to

c

2

taO "S ._

fl it a> ?

o -^ o 3 a) o

O CC O O >H Q

a>

I-c • —

- !S B

= o £

2 !»« i^
(S o o
O O iJ

60

S
cS

o c3

P-I <d
o

& . «^

g t: Si

O Q >.

.- ^ §

v< 1^ «

o o ?

h5 O

C<l 1-1 O CO o

o .-H 00 to 00 -^

i I :i

t- O 05 '^ 1— <

2 I

(M <o

■<* O ul

i-t ■»»< CO r-l

^, ^ O rH 50 00 Tj<
<I> HI ^ CO ^ 00 -.!j»
piH p^ iM (M (M OJ ijq

I I I ! I I I I I I I S I

X CO
r-l

I I I

s s

I I

3 C

PLi

o

a a

- a ^ 'S S 'S

" o 0^ o) (V a>

,. a .„ a „

"S
m

-2 ^ Vi ^ '>> e o e
ooO <dO_5 o !>^v>

2 -" a -=" a « 3 «

^ O M O «

I I I

PLi 3 'O

P3 a;

a : ?i= a

.3 ." Cu

a (2^

* -'" a

o ll^ g o g>
00 ^ q ^ "^ :a
"■ g § fe S .S

3 ij o o V3

:z; o a:

So "S
a 08

c3 O

O ^ i-

u Q m

u

3

.a
a

O

O

3

a
'p-

01

• a

c8

O

04

a

S^ .5-

a j«i a «

flS qj C8 ^

a ±! 3

■ O 02

s » >.

_5 O «8

O o iJ

O

O

■ft
P.

I «
a

o3

O

o

• a.

• &

<»

S* a

•P !5

>-• r

Q «='

TS -"

O O

h-3 Q

i-H o -* o

tH O rH

1^

o <o

o

CO

1 "*

CO

50 «0

•»1<

oa 1

1 <3»

^

1-4

rH

i-l

CO

Si

a:
'o

a

_ '^
+j as &.

w -« P^
§ O 2

go®

2 a ^

o '^ -S
O C5 W

a>

13
CO

OS

.a

n

o

bo

-«-3

*

G

<o

$M

•5

0

13

^

Pi

a

<o

fi

ss

0)

Oi

-«->

o

o3

pq

O

a

.a

CO

O

J<3
1

o

00

"3

<D

o

o

9)

CQ

Jz;

!zi

a

P^

Pi

298 Self-Sterility in Plums, Cherries, and Apples

M< r-t rH 05

I I I I 5J

i I I I I I I I I >^

I I I I I I I I I

o +
OQ eft

£ S

'^

.2 ^

35

^ I I i

l>>

p

a

•d

a

N

Cm

TA

pH

S 1

.2 1

1 d

o

^

>

o

^

a,

s
o
>

5?

'a.

PLI

fee

a

S
pq

a,

a}

J3

o

o
a

o

a
o

CO

O
60

o a 3 'G -3
O <J « O M

■«1» CO «5 O

^ ^ V ^

Ir- o O ^

t- e«s (N »o

■-I IM (N

■* US 05 S*" 'N

;S t^ (M «2 "o

C<1 ^ rH cq

O ■* 1-1 »0 i-H ?o

05 i-H ^ -* «D
>« 00 •^ "-H

S, a

(0 g

.2 ^

03 ^

>> a

C O

2 02

o a

'A S

cQ .— I " ;z3

J m o o

M

J

r-l (M OS p lO ^ O

05 «5 O ?0 O Ol t-

(M WS CJ «» 00 rH

1-1 rH (M

O

«f-l

OS

a

d

^ d

a, ;
a

60

a

: 60

• .a

'ol

Ti

^i^

Ti

^ f^

®

%

OJ

^ ^

-« 0,

60

^

02
a:

I'

d

s

^ 02

Js e

>j

ni

e

3 >>

60 ji;

O §

00 §

q

Q

o

>^ a

d O

« >«

»ri|

;i^

u

M

Ti g

r^S l><

O

o

o

O Bh

Wh

m

1-1

CJ

Ws'

i-:i m

O Q

m

1 -g

>« -t-

-J-

HH

-)-

•f-

-i-

1

o o

lO CD

oi 00

CO

t-

<M

05 Oi

o

00

t^

00

t-

IM

1

o

CO

fH

'I

S =«

rH iH

iH

tH

i-H

'

:^^

<»

^

b

OQ

1

1 '-I

§^

(N I

1

o»

1

1 cc

■^

05

C^

1

•*

-*

CO

«o

iH

o

1 CO

1^

«5 1

1

^

1

1 o

S

•^

"*

1

o

iH

1-1

1-1

eo

(N

ig

<N

i-(

(N

<N

(M

(N

C<l

5<l

eo

1-1

a

•

;

-S

m

2,

.g

o

H

53

OQ

a

93

-♦J

Pi

<

O

1

»-l

o

1

e3

<u
o

o

Yellow Ne'

s

CQ

^

^

p4

Ida Sutton 299

"* o o 00 eo

s s g g s s s

«o -^ ?o o "^

IM CC 5C C~

a)

2

a

1

o

IB

o

1

n

J3

Tl

I-»

IXl

«

tc

T3

Tin

^

^

u

>

1 fl

a

o

a

w

^

1 •«

P^

^

^3

P.

s

3
O

T3

[i<

^

o

P

O

^

SOtO-^JtOOCsOO-^OSiMCCio

tH (M r-l rl

ssssssss

I— li— I CCt-C<l kOi— It^lM

«0 C 9)

• _c •

Oi

§

.s

Pi

•F-4

o

A '3

^

'5

p^

!h

"S S >i

'S

N

o

m

n

pq 02 ^

Oh

Q

6C

^1

^

TS

C3

Q

"S)

0)

;-!

"Sb

P^ ^ 03

_s

^

^i

a

>

o

a s

O g >i

'o

c

^
C

P=q

o

OQ

•E3 W

S § -c

C!

•^

n3

04

"><

5 •«

9 V< CS

O

s;

s

o

o S

PQ m (-5

Q

Ph

fl^

O

u

ft O

I ^

(M 1-H «C <N -^

t3

g -3 1

H s c a ^-^i

•^ f^*- -o S S*

300 Self-Sterility in Plums, Cherries, and Apples

EXPLANATION OF FIGURES.

Fig. 1. Jefferson. Tree was covered with flowers ; different branches were crossed with
pollen from Coe's Golden Drop, Coe's Violet, Crimson Drop, and Pond's Seedling.
No fruits set by the pollen of Coe's Golden Drop, or Crimson Drop, only two fruits
set by Coe's Violet, whereas a heavy crop set on the one branch pollinated with
Pond's Seedling — see lowest branch on right-hand side of figure.

Fig. 2. Governor Wood. Tree in flower ; the flowers on one large branch (on left-hand
side of figure) were crossed with pollen from Early Rivers : the rest of the flowers
were self- pollinated.

Fig. 3. Governor Wood. The same tree as in Fig. 2. No self-pollinated flowers set
fruit ; a good crop set by the pollen of Early Rivers. Note that this photograph was
taken from the opposite side of the tree so that the crossed branch shows on the
right-hand side of the figure.

CAMBBIDGE: PBINTED by J, B. peace, M.A., at the UNIVEliSITTf PBESS

JOURNAL OF GENETICS, VOL. VIL NO. 4

Fig. 1. Jefferson.

PLATE XV

f

t.

m

<ali

J

Fig. 2. Cherry in Flower.

Fig. 3. Cherry iu Fruit.

n

QH Journal of genetics

AIJ64.

V.7
cop. 2

Biological
& Medical
Serials

PLEASE DO NOT REMOVE
CARDS OR SLIPS FROM THIS POCKET