The American Naturalist

MSS. intended for publication and books, etc., intended for review should be

sent to the E Editor of THE on NATURALIST, Garrison-on-Hudson, New York,

articles cont summaries of research work bearing on the

problems ns of or organic Siete eo especially welcome, and will be given preference

One hundrea reprints of contributions are supplied to authors free of charge.

Further reprints will ed Hp sa at cost

Subscriptions and advertisem aati should be = to the publishers. The

erten S dollars a year. Ai hi ostage is ps cents and

ye Ci 5

— > THE SCIENCE PRESS

— | _ NEW YORK : Sub-Station 84

Entered i ‘ , April 2, 1908, t the Post Offi ter, Pa., under the Act ot

= Soe = “FOR SALE a | TENTH EDITION.

ARCTIC, ICELAND and GREENLAND

ii oe te li ae — THE Se Ore,

THE AMERICAN NATURALIST

TAE

AMERICAN NATURALIST

A MONTHLY JOURNAL

DEVOTED TO THE ADVANCEMENT OF THE BIOLOGICAL SCIENCES

WITH SPECIAL REFERENCE TO THE FACTORS OF EVOLUTION

VOLUME XLVIII -

NEW YORK

THE SCIENCE PRESS

1914

THE

AMERICAN NATURALIST

Vout. XLVIII January, 1914 No. 565

A GENETIC ANALYSIS OF THE CHANGES PRO-

DUCED BY SELECTION IN EXPERIMENTS

WITH TOBACCO!

PROFESSOR E. M. EAST axb H. K. HAYES

BUSSEY [INSTITUTION OF HARVARD UNIVERSITY

THE PROBLEM

In 1903 Johannsen announced that continued selection

of the extreme values of certain quantitative characters

in successive self-fertilized generations of a number of

strains of beans had produced no changes in the. mean

values of the characters. He concluded that these par-

ticular strains were homozygous for the gametic factors

whose interaction resulted in the characters investigated,

that these homozygous characters may be properly de-

scribed by one or more gametic factors nonvariable in

transmissible qualities and properties, and that the varia-

tions observed in the characters of any single fraternity

were due entirely to the action of environmental condi-

tions during ontogeny and were not inherited. Funda-

mentally, these conclusions were a recognition of the gen-

eral value of Mendelian description for all forms of in-

heritance through sexual reproduction, combined with an

1 These investigations were conducted with funds furnished by the Con-

necticut Agricultural Experiment Station from their Adams’ appropria-

tions, by the Bureau of Plant Industry of the United States Department of

Agriculture, and by the Bussey Institution of Harvard University, and the

writers desire to take this opportunity of expressing their sincere appre-

ciation of this hearty cooperation which made the work possible.

6 THE AMERICAN NATURALIST [Vou. XLVIII

admission of disbelief in the inheritance of ordinary

adaptive changes. The latter conception was Weismann-

ian in that all inherited variations were held to be changes

in the germ cells. It was not necessary to suppose it im-

possible for the environment to produce such changes and

therefore to have been of no value during the course of

evolution, but merely to suppose that during the compara-

tively short period of experimental investigations no gam-

etic variations have occurred traceable to such a cause.

For his first conclusion to be justified, it was assumed that

the changes which every biologist knows do follow the

continuous selection of extremes under certain conditions

are to be interpreted entirely by the segregation and re-

combination of hypothetical gametie factors which are

constant in their reactions under identical conditions.

Numerous investigators working on ‘‘pure lines’’ with

different material corroborated Johannsen’s conclusions,

and, as it was seen to be possible to interpret in the same

manner changes made by selection in experiments where

self-fertilized lines were not used, such as those of the

Vilmorins and others on sugar beets and those of the

Illinois Agricultural Experiment Station on maize, many

biologists accepted them and considered them a great ad-

vance over former conceptions of the mechanism of

heredity. On the other hand, there were those who main-

tained a skeptical attitude, the chief criticism directed

against the conception being that all progress due to

selection must have a limit, which in many of these ex-

periments had already been reached, and that even if re-

sults were being obtained action might be too slow to be

detected.

THE MATERIAL

These criticisms were reasonable when applied to cer-

tain specific cases, and in 1908 the experiments reported

in this paper were designed with the hope of testing their

validity, using the species ordinarily grown for commer-

cial tobacco, Nicotiana tabacum, as the material. This

plant satisfies the conditions which are requisite for

No. 565] CHANGES PRODUCED BY SELECTION 7

material used in pure line studies. It has characters that

can be estimated readily and accurately and which are

affected only slightly by external conditions. It is easily

grown, is naturally self-fertilized, reproduces prolifically,

and is known in many markedly different varieties. In

fact, it is an ideal subject for work of this kind.

The investigations were not patterned after the stand-

ard type set by Johannsen wherein the constancy of suc-

cessive generations of pure lines grown from selected

extremes were tested, since even if it were possible to

gather a quantity of data at all comparable to that col-

lected by Johannsen (:09) and Jennings (:08) in their

brilliant investigations, the criticisms mentioned above

might still be made. The plan chosen was that of cross-

ing two varieties of tobacco which differed in a character

complex easily and precisely determined, and of selecting

extremes from a number of families of the F, generation.

If Johannsen’s views be incorrect, such continued selec-

tion should affect each family in the same degree. If his

conclusions be justified, selection should reach an end-

point in different generations in different families, and

there should be no relation between the number of genera-

tions required to reach this end-point and the progress

that is possible.

There should be no need of a historical summary of the

previous investigations that have been interpreted as cor-

roborating or refuting Johannsen’s conclusions. Such

summaries have been made in other papers. It should be

mentioned, however, that the classical researches of Pearl

(:11) on the inheritance of fecundity in the domestic

fowl have been so planned and executed that certain of

the criticisms directed against Johannsen mentioned above

are not justified, yet Pearl finds himself thoroughly in

accord with the Danish physiologist’s position.

Several hundred varieties of Nicotiana tabacum exist

which differ from each other by definite botanical char-

acters, yet only two general characters suitable for our

purpose were found. We desired to confine our observa-

tions to quantitative characters that were influenced but

8 THE AMERICAN NATURALIST [Vou. XLVIII

little by environment, and number of leaves and size of

corolla were the only ones that satisfied this requirement.

Such character differences as height of plant and size of

leaf, while undoubtedly transmissible, are influenced so

strongly in their development by nutrition that work with

them is exceedingly difficult. For example, if a certain

variety of Nicotiana tabacum is grown under the best of

field conditions, the longest leaves are about 28 inches and

the total height about 6 feet, but a portion of the same

seed fraternity may be grown to maturity in 4-inch pots

without reaching a height of over 16 inches or having

leaves longer than 4 inches. On the other hand, several

experiments conducted in the same manner have shown

no difference between the frequency curves of variation

in number of leaves or of size of corolla, whether starved

in small pots or grown under optimum conditions. The

character complex number of leaves was chosen for this

investigation rather than the size of corolla because vari-

eties that differ greatly in number of leaves are common.

TABLE I

FREQUENCY DISTRIBUTION OF NUMBER OF LEAVES PER PLANT WHEN

ARVED IN SMALL POTS

(Compare with frequency distribution under normal field conditions at

Forest Hills, Massachusetts, in Tables VII and XI)

No. of Leaves per Plant

Plant No.

22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | 32 | 33 | 34 | 35 | 36 | 37

(6-1) SSO AS) SE Tih:

(6-1)-1 1 SOG TI2 ph tee eres

(6-2) bara 1 Sp o¢ (347161 14) 8S) Sai,

(6-2)-2 RER E S OL bh Oi, op o AD e ee

(56-1) solk Pe SiR OE TT 2a a

(56-2) EPE a a Lie beh Ol a EE

Previous Work or THE ‘‘Havana’’ X ‘‘SumatrRa’’ Cross

Several crosses have been made between varieties of

tobacco that had a mean difference of seven or eight

leaves, but the majority of the data reported here were

collected from the descendants of a cross made by

Shamel between the types known in Connecticut as

‘*Havana’’ and ‘‘Sumatra.’’ The ‘‘Havana’’ parent was

No. 565] CHANGES PRODUCED BY SELECTION 9

from a variety that had been grown for a number of years

at Granby, Connecticut. It averages about 20 leaves per

plant although ranging from 16 to 25 leaves. The aver-

age height is about 1.4 m. and the average leaf area about

7 sq. dm. The ‘‘Sumatra’’ parent was a type specimen

of a variety that had been introduced into Connecticut to

be grown under cloth shade. It averages between 26 and

27 leaves per plant with a range of from 21 to 32 leaves.

The average height is nearly 2.0 m., but the average leaf

area is only about 3 sq. dm.

According to Shamel, the first hybrid generation of

this cross developed somewhat more vigorously than the

parent types and was uniform in its habit of growth.

The second generation, he thought, was hardly more vari-

able than the first. Several F, families, the progeny of

inbred F, individuals, were grown in 1906 and proved to

be a variable lot. One of these plants produced 26 small,

round-pointed leaves with short internodes between them.

This plant was thought by Mr. E. Halladay, upon whose

farm the experiment was conducted, and Mr. J. B. Stewart,

of the U. S. Department of Agriculture, to be worth sav-

ing from its promise of producing a desirable commercial

type.

In 1907 the Department of Agriculture made an agree-

ment with Mr. Halladay to grow two acres of tobacco for

experimental purposes, and on his own initiative Mr.

Halladay grew a number of plants from inbred seed of

the one that bore 26 leaves. This selection, numbered 2

h-29 in accordance with the department nomenclature,

was comparatively uniform in appearance and several

plants were selfed. In Mr. Halladay’s absence, how-

` ever, all of the plants were ‘‘topped,’’ except one that

happened to be rather late. This plant was selfed. It

had 26 medium-sized, round leaves and grew to about the

same height as the Connecticut Havana.

In view of Mr. Halladay’s high opinion of the type, the

seed of this plant and the remaining seed of its parent

were planted in 1908. The plants of this generation pre-

sented a uniform appearance and promised a high grade

10 THE AMERICAN NATURALIST [Vou. XLVIII

of wrapper tobacco, but the crop when cured lacked uni-

formity. Some leaves of exceptionally high quality were

produced, but the crop in general lacked that characteris-

tic known as ‘‘grain’’ and had too large a proportion of

heavy leaves—the so-called ‘‘tops.”’

From this 1908 generation 100 seed plants were selfed,

their leaves harvested, cured and fermented separately,

and data on quality recorded. The type was also grown

commercially on a large scale. The commercial results,

however, have been reported in another paper. Weare to

consider only the results of the selection experiment that

began in 1908, through the cooperation between the U. S.

Department of Agriculture and the Connecticut Agricul-

tural Experiment Station, a joining of forces that in 1909

included the Bussey Institution of Harvard University.

Shamel ( :07) considered the strain produced by this cross

to be the result of a mutation. From a study of the

data from the previous work on the cross it seemed to the

writers that a different interpretation of the results might

be made. While it was not impossible that the many-

leaved type that had been isolated was the result of a

mutation, it appeared much more probable that it had

arisen through a recombination of Mendelian factors.

The type had the habit of growth and size of leaf of the

pure ‘‘Havana’’ variety and the number of leaves of the

‘‘Sumatra’’ variety, a combination that might reason-

ably be expected to be the result of the Mendelian law.

RESULTS ON THE RECIPROCAL Cross, ‘‘SumMaTra’’

X “HAVANA”

To test the hypothesis that the new tobacco was the

result of such recombination and could be reproduced

whenever desired, the reciprocal of the original cross was

made in 1910. The female parent, ‘‘Sumatra,’’ was the

direct descendant of a sister of the plant used as the

male parent of the original cross by Shamel in 1903

through seven generations of selfed plants. The male

parent, ‘‘Havana,’’ was from the commercial field of the

Windsor Tobacco Growers’ Corporation at Bloomfield,

No. 565] CHANGES PRODUCED BY SELECTION 11

Connecticut. It was a descendant in a collateral line of

the plant used by Shamel in 1903 as the female parent in

his cross.

Table II, giving the frequency distribution for the num-

ber of leaves of the two parents and the first and the

second hybrid generations, is a complete justification of

our prediction as to how the hybrid type produced by

Shamel originated. The ‘‘Sumatra’’ and the F, genera-

tion were grown at New Haven, Connecticut, in 1911, the

‘‘Havana’’ was grown at Bloomfield, Connecticut, in 1911

from commercial seed of the same variety as the plant

used for the male parent, while the F, generation was

grown at New Haven, Connecticut, in 1912. The F, gen-

eration, producing an average of 23.3+.14 leaves per

plant, is intermediate in leaf number, since the ‘‘ Havana’’

variety shows an average leaf number per plant of 19.8

+ .08 and the ‘‘Sumatra’’ variety 26.5+.11. The varia-

tion as determined by the coefficient of variability is some-

what less for the F, than for either parent. The value

for the ‘‘Sumatra’’ variety is 6.64 per cent. +.28 per

cent., for the ‘‘Havana’’ variety 6.98 per cent. + .27 per

cent. and for the F, generation 6.24 per cent. +.41 per

cent. Taking into consideration the probable error in

each case, one may say that the variability of the three

populations is almost the same.

The variability of the F, generation, however, is greatly

increased. This is shown by the high coefficient of vari-

ability, 10.29 -+ .23 per cent., although a glance at the fre-

quency distribution with its range of from 18 to 31 leaves

brings home the point without recourse to biometrical

calculation.

The appearance of the plants in the field corroborated

the data of Table II in other characters. The F, genera-

tion was intermediate in the various leaf characters, such

as shape, size and texture, that distinguish ‘‘Sumatra”’

from ‘‘Havana’’ tobacco, and in these characters it seemed

as uniform as either of the parental varieties. On the other

hand, the F, generation was extremely variable. Some

plants could not be distinguished from the pure ‘‘Suma-

12 THE AMERICAN NATURALIST [Vor. XLVIII

tra,’’ others resembled ‘‘ Havana,’’ although of course the

majority were intermediate in various degrees. Several

plants combined the leaf size and habit of growth of the

‘‘Havana’’ parent with the leaf number of the ‘‘Suma-

tra’’ parent. In other words, plants were produced in

the F, generation by the recombination of Mendelian fac-

tors that exactly repeated the type which Shamel had ob-

tained in the F, generation of the reciprocal cross made

in 1903 and which he thought was due to a mutation.

This fulfilled adequately the prediction made by us in

1908.

RESULTS or SELECTING ror High NUMBER anp Low Num-

BER OF LEAVES IN THE ‘‘Havana’’ ‘‘SumatTRA’’

Cross

In describing the reproduction of Shamel’s hybrid with

numerous large leaves by a reciprocal cross, there has

been a chronological inversion. This was done simply to

show that the original hybrid known commercially as

“The Halladay’’ was actually a recombination of Men-

delian factors in which the ‘‘Havana’’ and the ‘‘Suma-

tra’’ varieties differed. We will now describe the effects

of selection on the original ‘‘ Halladay hybrid.”’

It will be recalled that the selection experiment which

is the principal subject of this paper began with the self-

ing of 100 seed plants of Shamel’s Halladay hybrid in

1908. These plants were the F, and F, generations of the

cross ‘‘Havana’’ X ‘‘Sumatra.’’ Plants numbered from

1 to 49 were the F, generation; those numbered from 50

to 100 were the F, generation. They were apparently

breeding true for the short habit of growth and large-

sized leaf of the ‘‘Havana’’ parent and the goodly num-

ber of leaves of the ‘‘Sumatra’’ parent. The casual ob-

server either would have said with Shamel that here was

a mutation breeding as true as any tobacco variety, or

that a fixed hybrid, a hybrid homozygous in all of its

gametic factors, had been produced. Accurate data

taken on the progeny of those of the F, and F, seed plants

which it was possible for us to grow in our limited space,

No. 565] CHANGES PRODUCED BY SELECTION 13

however, show that such judgments would have been

superficial. The general type of the plant did appear to

be fixed, but the frequency distribution for number of

leaves of the F, and F, populations were not the same.

Strictly speaking, they were not fixed. What would be

the result of selecting (and selfing) extremes from these

different families for a number of years? A tentative

answer to this question is to be obtained by examining

the remainder of our tables.

The tables are arranged roughly in the order of the

effect that selection has had in changing the mean of the

various families that were the starting points of this part

of the experiment. The selections were grown near Bloom-

field, Connecticut, on the light sandy loam of that region,

soil typical of that which produces the famous Connecti-

cut River Valley wrapper tobacco. Duplicate experi-

ments with several of the original families were made at

New Haven, Connecticut, however, on an impoverished

soil not fitted to grow a good quality of tobacco even after

supplying large quantities of tobacco fertilizer, and in

the condition used not fitted to grow good crops of any

kind. Two families were also grown in triplicate, the

third selections being planted at Forest Hills, Massachu-

setts, on a very fine type of rich garden land which brought

out maximum luxuriance of growth, but which did not

produce good tobacco quality. These experiments were

not true repetitions of the experiments at Bloomfield,

Connecticut, since aliquot portions of the seed from the

selfed plant grown there were not sent to the other places

to be grown. But they were duplicates in that each

family came from the same F, or F, mother plant,

although, beginning with the F, or F, population, differ-

ent selfed seed plants furnished the starting point of selec-

tions carried on independently. In this way there were

afforded a greater number of chances to see what selec-

tion could do.

Table III shows the results obtained from family No.

77. This family arose from an F, plant having 23 leaves,

one below the modal leaf number if we may judge from

[Vou. XLVIII

THE AMERICAN NATURALIST

sooovaoy, (TIVW) Vulvwoag ANV (TIVNA) VNVAVH NIIMLIg

SSONQ NI INVIG Ud SHAVA'T JO YAAWON AO NOILAJINLSIA AONTNOAAT

Il @TaviL

12’ * 8F'L | 90°+ £61 | 80° + 8S [08z) T | T| F |£ | st| 6c SF] v9] OS | FFI See |T "Tl" "1 Q8(ZI6T | 8H] = T-(Z—-22)

02° * Z6'9 | GO" F8'T | 0° 9°9G |T8Z| T| T|? | ZL| Z| 1F| 6F| 89| SE} Fe) OT! Z LATE Se" 1 Oe TOT 1 Ae I-(Z-L2)

OEF ZEG | LOF SES | Il = OFS PEZ) T|o/¢ |9 | SL} 22] OF | 68} SE} FE} ST) IL) F ; ri "|tt t] 26 1OT6L | 4A (€L)

EEF 8E'9j LO F SST | Il FEISS | we ree (ate emma patie cms Oe Gan yt ak ee a ole ek CARR sige CEOE Cee GL | tae LL

LZ = L76 | 90° = $0°Z | 80° 6'13 (993); Zz |z |6 | 12| 02| 28| ZF| €S| 6F| OZ|¢ |T “°°! OZ |OT6T | 4a (1-22)

8T°'+09'9 | F0°= TFT | 90°F SIs 908} ° eS TS Peet Sot 261881 Oe re oe **) OS |TIGT | 84 I-(1-L4)

96° LES | 0° = FST | LO°F FST [OES * : T IT |g |F 19 | Te} se] 29] 9F| OT] © | T | Of\zr6T | SH) Z-1-(1-44)

e ze te) jae |se | az | om |s [ve |se |ze| ra| oz [or or] ar or [or ni Bal |

= =

A0 ‘a's vy AE SE 2 S ON

que[g sad seave'y jo Iəquny JR S 3

‘NNOO ‘CTAIANWNOOTgQ LV NMOUDY

‘LL ATIWV NI SUAVA'T JO WAAWAN AO NOILNSINISIA AONTAOTYT

OI WIdVib

EZ = OZOTT | 90°= 19°C | 80°+ F'F2| O8F |S 12 | SL} SS| 6F| 6S) ILI: 99 SF] 88i IS 6 jl POC “tq ‘vuvaey X seung

lesen | Oe ort) stl egec| cf fl ee ee ee ee i te e r. tease ak Suvas X Sa

82'=F9'9 |80=9LT|IT'=S97| SZI |T |+ | ST|9T| 12) 9s] zjete pop P ee ees veageumg

2z°=86°9 looser sorger) ost cll ell ct te lo lorlecleplepiee| epee Mets Fs : ` + + ouea sH

Ts | og | 6z | s3 | 23 | 9% | sz | #3 | eg | zz | 13 | OZ | GE | St} LT

AD ‘a's y IL verser FN

quvig 13d seavay jo soquinyy

CHANGES PRODUCED BY SELECTION

No. 565]

‘Buyyqnop Jeo] pomoys syuerd zo requmu y p

*{NOTPOUNOD PPPYUWOO Je UMOIH g

st +8z9 | 90 +691T | LO*69Z | S22 | 1 | Z| tl 2e| $9169] 6G seoir I 6z Z161 Ti 1-1-(2-92)

Za’ * 889 90° = 08'T 80° = 19% OZ | T| 119 | 02] #2) S| 87i eio |Z 8% TIGI 3 1-(2-92)

12’ * 129 SOEST oF FP | 603 ||" | 1 |1 19 | OT| 22109 \#¢ | OF| OT! F lz O16 y (z-94)

Pe = F9'9 80° = 09°T IU +1 | 68 ‘eeft lp letleries istie | welt! gz 6061 94

Zo" = L0°9 el = EST etraz | Te reige die P je jo1olr| ee OT6I i (1-92)

FL + 82'S £0° = 92'T GO’ ¥ 6'&% org T 16 |8I 99 SOT CL 96) 911 1 e TI6I Ki T-(1-92)

ze | 1e | og | 6z | sz | 23 | 9% | s | Fe | sz | zz | IZ | og mozeg Jo

sy ‘as Vv 1810, | poet Lag TOD | o ‘ON

qud 19d savory Jo Iəquny | 1əquny 5

‘NNOO ‘CTIIAWOOTE LY NMOUD a

L XIUNV,T NI SAVI'I JO YAAWON AO NOILAJINLSIA AONTAdAUT

AI ATSAViL

¥0' * IL'Z | OL = 62'S | FI = 2°63 [81] 3| 0 | 2 | Z]E€|OT|IT|ZI|OE|Eege| IT] Tt |T “T1711 OB 1ST6L | S| ATD

62’ ZL | 80° Z6'T | TT = 9'°9Z PRT] || g |9 | ST] SS TS fe OS Sti 2 1B ci iol es iter | tax I-(Z-LL)

18" * $2'6 | 80°= G8" | IL" FEZ |102|’ 1/%2{|% {ie |Z | OL} St| ez] ve} 2e| ge] te] ort lt "| 2g |O16T | 4a (-LL)

ESF 19'S SO FEST Le LIZ Pee € |9 |Sz| orl IF] 211% ` `| Og |TT6T | 34 T-(1-44)

GZ’ = LI'9 | S0'= 62'T | 80°= 6'02 (SET) IT |6 |8| FF] 6%| OT) 2) 1] TZ \ster |51| T-1-(1-42)

e oe | oe | ve | ee) ae | we | ee) oz | g | oe [os lve) os [o| wa | os or orien Bad |g

E 5

“A°0 ‘a's v <3 se $ 3 ON

a’ La |

que} q 19d seavery Jo soquinN Fe sis

[=]

‘NNOO ‘NUAV]T MIN LV NMOWÐ

16 THE AMERICAN NATURALIST (VoL. XLVIII

the F, generation of the reciprocal cross where the mode

was at 24 to 25 leaves. The F, fraternity that it pro-

duced was somewhat smaller than one would wish if

he were to, be confident of the calculations made. The

mode is 22 leaves and the mean nearly the same, 22.4

+.11 leaves. From among these plants, a minus variant

having 20 leaves and a plus variant having 27 leaves were

selected to produce the F, generation. The modes in this

generation are 21 and 25 leaves, respectively, a difference

of 4 leaves; and the means are 21.9+.08 and 24.9+ .11

leaves, respectively, a difference of 3 leaves. Progress in

both directions continued when a 20-leaved plant was

selected to carry on the minus strain, and a 30-leaved

plant was selected to carry on the plus strain. The modal

classes of the F, generation are 21 leaves in the minus

selection and 26 leaves in the plus selection, while the

means are 21.3 + .05 leaves and 26.6 + .07 leaves, respect-

ively. In the F, generation the plus selection was lost,

but the minus selection grown from a 20-leaved plant had

the mode dropped to 18 leaves and the mean to 18.4 + .08

leaves. In order not to lose the plus selection entirely,

however, more of the F, generation seed was grown in

1912. The mode is the same as in 1911, but the mean

dropped slightly to 25.8 + .08 leaves.

Here one notices what is very common throughout the

experiment ; the extremes selected for mother plants were

not members of the most extreme classes. This means

simply that vigorous healthy specimens were always

selected as the mother plants, and often the most extreme

variants did not come up to the standard. It is hardly

just to criticize this procedure, however, for with the best

care that it was possible to give, the experiments with

several families were terminated on account of non-

germination of seed or for some similar reason, it being

impossible, on account of the pressure of other work, to

self many plants in each selection. Even where seed

from several mother plants was collected, it did not in-

sure the continuation of that selection. The necessary

space and care involved in growing so many seedlings in

CHANGES PRODUCED BY SELECTION

No. 565]

“PIOFMOCT, YU UMOIH g

oc" ELO |S "= ZOC|IL'+OO8| OOL | | 9] Z| St| 18/98) 0E6166 |F IT |T ZE ZI6I ta Z-I-(Z-61)

8z°* 10°9 | 80'= OL'T | II" 88%} 80T 1] 0/8 |£ | OT) 20)\ 28] 13| Ft1z 6% TI6T | 4a 1-(-61)

61° = PSS | GO'F OST | 20°F T'LZ| OGT "l'""1S |L | 28|OP| 4h) PE/ 9s 2 l 6% OI6L i (-61)

LG * 189 | LOF 6L°T | OTF E92; LPI Z | Sl| S| PE] 6Z| S| 21/9 | FL TI’ 9% 606T g s61

SIT'= 82'S | S0'=F SE 90'F S'S) 66I "esiet 2 OTOP 29) 8h/ ST) Si Ti] we oter | *a (I-61)

Zh + 628 | Il’ + 80° | STF 1'SZ) 06 outteesie**i@ T 78 1OLGT ZEIGE o 1 £t rt) e TIGI 4a T-(1-61)

6T'=P8'F| SOF ZET | SOF ELS) SET treg |F | 20/08] 12] 62S |8 hoe TIGI ‘x | iiA.

eg | +e | ee | ze | te | og | 6z | sz | zz | 92 | sz | +z | ez | zz | ta | LI uamea jo

“A‘O ‘a's ‘v PL soavoy jo, OOD | on ‘ON

queg tod seave'y yo soquiny - | coquiny A uap

"NNOO ‘NIAYVH MIN ZV Nowy

0% * IT'9 | 90° = SLT | 80°F S63) SIZ S12 | Fl SS] C9 ee re Ors Fit ss oe TIGI gi 2-(2-61)

92° FIL | 0° GOS | OTF 2°8Z| GAT | T| T | 1 | S| 2 | Sl) 8S| OF] 8Z| HST |S p po 62 OI6I i (Z-61)

12° = 18'9 | L0 6L°T | OTF 8°92 | LPT peer eS Sl eel eel eres Chie | Ft 9% 606T “fT 61

IZ’ = $9°9 | 90°F OL'T | 80°F 69% | LES "13 13 | GL| 92) €h| 89) T| 9S) 6T) Se || #3 OI6I i (1-61)

SZ'= 97'4 | L0`= L8°T | 60°F 897| OGT I|T IT |F |4 91|9|90 oelet ejt] z TIGI 1 1-(1-61)

9g | ce | +e | se | ze | te | og | 6g | 83 | Lz | 9 | 9z | FZ | sz | zz | IZ |JUeABqJO

AO a'S Y IPL soawoy jo; TAND | TOR ‘ON

qux[g ted soave'y yo oquiny qoquinN A aÐ

*NNOQ ‘TIMIANO LY NMOUH

‘6[ ATINV,[ NI SAAVTF'I JO SAAWAN dO NOILAIINISIQ AONATNOay,T

A WTavVib

18 THE AMERICAN NATURALIST [Vor. XLVIII

isolated seed pans filled with sterilized soil made it im-

possible to start more than two sets of plants for each

plus and each minus selection. Generally both sets grew

perfectly, but occasionally both failed, and in that case it

was usually too late in the season to start a third set even

if it were available.

The second part of Table III shows the results obtained

on the poor soil of New Haven, Connecticut, with the same

family. There was continuous progress in both direc-

tions. The minus selections during the three generations

show a constant reduction of mode, the figures being 23,

22 and 21; the plus selections show an even greater in-

crease in mode, the figures being 25, 27 and 28. The same

decrease and increase occur in the means until in the F,

generation there is a difference of nearly 9 leaves, the cal-

culated means being 20.9+.08 leaves and 29.7+.14 leaves,

respectively.

Figs. 1 and 2 show typical plants of the plus and minus

strains of this family as developed by 3 years of selection.

Fig. 3 illustrates an interesting change of phyllotaxy in

some plants of (77—2)-1-1 as grown at New Haven in 1912.

Passing to the data on Family No.76 (Table IV) there

is the same evidence of the effectiveness of selection, ex-

cluding the minus strain in 1910, of which only 31 plants

were healthy. This effect is markedly less than with the

other family. The mode of the minus selection remained

at 24 leaves and the mean was reduced only from 24.1

+ .11 leaves to 23.9 +.05 leaves,—hardly a significant

figure. The mode of the plus selection crept up to 26-27

and the mean to 26.9 + .07 leaves, there being here one

more generation than in the case of the minus strain.

Table V gives the data on plus and minus selections of

Family No. 19 at Bloomfield for two generations. The

original family stock of the F, generation has the mode at

27 leaves and the mean at about 26 leaves. A 24-leaved

plant of this generation became the parent of the minus

strain, giving in the F, generation a population with the

same mode and a slightly higher mean (26.9 + .08 leaves).

Continuation of the strain through a 24-leaved plant gave

No. 565]

CHANGES PRODUCED BY SELECTION 19

an F, population with the mode one class lower and the

mean at 25.8 + .09 leaves. Whether this slight reduction

really means anything we are unable to say.

it yields at all to selection,

the progress is very slow.

On the other hand, a con-

siderable gain has been

made in the plus selec-

tions. The mode rose im-

mediately to 29 leaves

when the progeny of a 29-

leaved plant were grown,

and went up to 30 leaves

the next generation, the

modal condition being the

same as the number of

leaves of the parent plant.

The means are 26.3+.-10

leaves, 28.7+.10 leaves

and 29.2 + .08 leaves, the

amount of progress being

—as may be seen—2.4

leaves and 0.5 leaf in the

two successive genera-

tions. This result appar-

ently indicates a slowing

down of the effect of selec-

tion.

The continuation of the

table gives the results ob-

tained at New Haven on

this same family. Here

there are data from three

generations, and these

data modify the conclu-

sions based on the results

At least, if

1 X F HALLADAY HA-

vANA Tosacco (77-2)-1-1, WHICH AV-

PLANT IN 1909.

Haven, 1912.

obtained at Bloomfield. Both plus and minus strains

nearly parallel the Bloomfield results for two generations,

20 THE AMERICAN NATURALIST [Vou. XLVIII

the F, generation means being 28.3 + .11 leaves and 25.1

+ .15 leaves, respectively, but in the F, generations they

differ. Selecting minus extremes for the first two genera-

È PLANT OF HALLADAY HAVANA Topacco (77-1)-1-1, WHICH AVERAGES

20.9 LEAVES PER PLANT. IT IS THE RESULT OF THREE YEARS OF SELECTION FOR

Low Lear NUMBER IN FAMILY 77. New HAvEN, 1912.

tions reduced the mean of that line from 26.3 + .10 leaves

to 25.1 + .15 leaves, but the third selected generation (F,)

had a higher mean than the original family (27.3 + .08

leaves). The parent plant of this F, generation produced

No.565] CHANGES PRODUCED BY SELECTION 21

24 leaves, and as the strain indicated that it was hetero-

zygous for a number of factors by showing a coefficient of

variability of 8.29 + .42 per cent., it is possible that the

selected parent plant may have belonged gametically to a

higher class than was indicated somatically ; nevertheless,

it can not be denied that three generations of selected

minus extremes have produced no results. This conclu-

sion is not valid for the plus strain. Starting with 26.3 +

.10 as the mean number of leaves (F,), the succeeding gen-

erations had means of 27.1+ .07 leaves, 28.3 + .11 leaves

and 30.0 + .11 leaves. The differences are 0.8, 1-2 and 1.7

leaves, respectively. Progressive change has certainly fol-

Fic. 3. CHANGE OF PHYLLOTAXY IN SOME PLANTS OF (77-2)-1-1 GROWN IN New

HAVEN IN 1912.

22 THE AMERICAN NATURALIST [Vou. XLVIII

lowed, and unless one considers that the results of 1912 are

somewhat too high (probably a valid assumption), the

change has increased instead of decreased. Naturally

there must be a decreased momentum in change of mean

time, but this decrease is not yet shown by the figures.

Fig, PLANT OF HALLADAY HA- Fie. 5. PLANT oF HALLADAY Ha-

VANA ToBacco (19-2)-1-2, WHICH Av- vyAana Tosacco (19-1)-1-1, WHICH Av-

E E 0 LbBAvEs PER PLANT, IT ERAGES 27.3 LEAVES PER PLANT. THREE

IS THE RESULT OF THREE Years OF Su- YEARS OF ŠELECTION FOR Low LEAF

LECTION FOR HIGH LEAF NUMBER IN NUMBER HAVE PROVED UNSUCCESSFUL,

FAMILY 19, WHICH IN 1909 AvmRracED New Haven, 1912.

26.3 LEAVES PER PLANT. New HAVEN,

1912.

No.565] CHANGES PRODUCED BY SELECTION 23

Representative plants of the plus and minus strains of

family 19 as obtained by three years of selection at New

Haven are shown in Figs. 4 and 5.

Family No. 5 (Table VI) shows a decrease in mode

from 28 to 26 leaves, and a similar decrease in mean from

28.1 + .06 leaves to 26.6 + .09 leaves as a result of the first

minus selection. A second minus selection, however, in-

dicates either that the future progress is to be very slow

or that the entire effect of selection was manifested in the

first selected generation.

With the three parts of Table VII we take up the re-

sults on Family No. 6 at all three stations. The minus

strain was carried on only two generations at Bloomfield,

but with this exception there are data upon three genera-

tions. At Bloomfield the two generations of selected

minus extremes resulted in 0.6 leaf decrease in the mean,

but at New Haven the results were negative, the means

advancing from 25.8 + .06 leaves to 27.9+.12 leaves in

three generations, while at Forest Hill the mean remained

practically the same. Surely selection was unprofitable

ere.

The first year of selection from the other end of the

curve, however, resulted in marked progress. The mean

advanced nearly 5 leaves in each case. The original F,

mean is 25.8 + .06 leaves, but the three F, means are 30.7

+ .09, 29.6 + .08 and 30.8+.12 leaves. This is a remark-

able concurrence of results. The means in the two suc-

ceeding generations were about the same in the Bloomfield

and New Haven experiments, but there was another defi-

nite advance at Forest Hills. Such a result should not

be unexpected. If the F, generation were almost but not

quite a homozygous lot, and if one assumes that selection

of extremes from homozygous population has no effect

in shifting the mean, it would frequently happen that

some individuals selected to continue the line would be

homozygous in all factors and some heterozygous in one

or more factors.

The cause of the peculiar distribution of the population

(high variability) of the F, generation grown in Bloom-

THE AMERICAN NATURALIST [Vou. XLVIII

0&8’ = $6'6 | 60° €0°E | STF SOE EPZ Z | 9 | ST| Z| SF) OS | SF| Z| ST/9 |Z IT [O |S | SIL T| si ays Te ZI61 Fi ¥-1-(Z-9)

El = LIP | FO’ IS'T | $0" 1°62 |6EZ T |O |F | 06| 19| 02) #9} ZT1Z eee Cees (es Shae Beat Bee het ZE TI6I 44| 1-(-9)

TZ'=8T'9 | 90° 06'T | 60°F 2°08 |903| Z | Z | OL] S| 98| GE) FH) GS/STIE |S IT Pp Topo og | OT6T | (Z-9)

LT = 91'9 | FO" 6ST | 90° 897 |LOE)**)° "Tt EAD EE OS SO) 22 ee oe Obl Bit a 8% | 6061 | ‘A 9

ZZ = LO'9 | SO" EST | 80°F SSS IST || 110 0 10 11 0 [88198] IF PF OT! ti eT SZ | Ot6T | A (1-9)

61°+ 909| GO" ZEIT | LO’ TSS SEs 10 |0 |¥ [6 | | 2h) dL) 8h 108) FI Toc e | tter ta) Go

9g | og | 8 | 88 e | 18 | os 6z | 8% | 22 | 93 | 9z | FS | ez | zz | IZ | og | GT | ST | }ueTEG

3 JO | umoiyy Q

“AO ‘a's $i = se = 30x F ‘ON

que[q ied saatey Jo roquiny ana? 3

*NNOQ ‘CTHIANOOTgQ LY NMOLDY

‘9 ATIWV, NI SAVTI JO WAWOAN AO NOILAIIISIA ZONTAÙTHT

IA Fay L

8T + F8'9 | GO" FST | LO°-F 06S | Z8Z | I |9 |sT|1e|s9|£9|zr|zeloz|ř |Z |T Eg ZIGI 84 €-1-(z-¢)

LI'*G0°9 | GO’ ELT | 80°F 9°8Z| EFS | T | T|8 |0Z/ OF) S| HS) 2E/ Ste |S ZE TIGI 4a 1-(Z-9)

SI'= GZS | SO’ 89'T | 80° 367} 92S | £ | F | ST|9S/ 9F/ 99| E| 9| L T | TE OT6I WT -¢

ST'= FES | FO'F OST | 90°F T87) P63 T |Z |ST| 68] S2/€2/29| 90/8 jg | T 8% 6061 sE g

PE = SIL | 90°F I6'T | 60°F 99% TIZ TIT |9 |8 | 9T| | SF] So) 18) ZZi 0]|z GZ O16T I 1-3

UIV,

is k 5 Poa | ve | se | ze | te | os 6g | 8% | LE | 98 | SB | FZ | 88 | Zz | IZ y Co uor nen E

| quviq 19d səavag Jo requiny JoquinN A 9

‘NNOO ‘GTaIaNWOOTG LV NMOWH

‘G ATINV,, NI SHAVATT AO UAAWOAN AO NOWMAaMIsIg AONTAÒTAA

IA WTAViL

CHANGES PRODUCED BY SELECTION

No. 565]

*plyMoorg 4B UMOIH »

IA’. o'ra

OnT

FI = L8'E | 90" = LE'I | L0'= 8°ZE |89T si ‘1g IT ege IO EEs Oe Oe SIGI | "a |cg-@-0)

ST SOF | 90°F FST | 60°F TSS (ZF | T | S| 2 | St) 221 2h) 90919 Tt Ae ee SE TIGI LI g-(7-9)

LZ°* ggg | 80°F OL'T | STF 8°08 BFO Tg |e oTi oioi Is) eiA eo baci Tana tame 0g OI6T | %@ (Z-9)

ALT= 91'9 | POF 6G'T | 90° 8°SS |208" E SEES OT IT ET EGS E Gee ore 83 6061 sI 19

GZ = p09 | 90° = OET | 60°F 8'97 |€6 3l A "1S |8 |81103287 gi | gs O16T | PA (1-9)

6l°= I8'F | 90 F63 T| 201897 EST oT” ALAL OG TIEI 69) IE GiS ii a 9% TIGI <i 1-0-9)

SI'S LLP | SO°F P'TIT | LO°F 09S SGT) I’ OW Peel ae + OG) Peer ea 4” VG SIGI 8a T-I-(1-9)

i ae t 3 6g | sg ze | 98 | os | +8 | sg | ze | te | og | 6z | 8g | Lz | 9% | 9g | FZ | ez | zz nere cea) Gon a

LS ueq 19d seavey jo requinyy zoquiny | 79A [-BreUeH

‘SSY ‘STHH Lsau0,g NI NMOUDH

03° 00°S | 90° TS°T | 60° 7°08 SPI "| |’ ‘} FOL] QELS +17 | Sai ZEIS if fon Wt i) Fae ZE ZIT6I a itrio

PEF OEO | LOR PSL OV SZ 8S IPS l TTI S18 O lig OS er vertii te it ace re 0€ TIGI ta I-(Z-9)

61° T9°S | 90° 9'T | 80° = 9°62 |102) T | O| T| 0} 0/8 | IL] $8] 28/08! €8\ ati iT IT gon las Beek og OTGI Bf (Z-9)

LI’ ITO | $0' = 691 | 90°F BSS (L208) ee OB a Be 80) £21 99) 08 Ol) Pi 8% 606T i 9

66° = 09'0 | 00° 29 T | SOF LEC (SGT ae T e e TE PE ee a oe ey a GS OT6T 7 (1-9)

OEF €9°L | 80°F SI'S | SL F 6'42 |SFT| EAER JO IE ILLEG IE 46106) 04 Sle g GS SIGI ‘a timo

ie oo £ 3 8g | 48 | 98 | se | #8 | 88 |- ze | 18 | og | 6B | 83 | Lz | 92 | Gz | FZ | £z | zz | Iz Dea daup denke ee

5 que[g tod saavery Jo 1əaqunyN Joroquiny| 789A | -199

‘NNOO ‘NEAVH MIN LV NMOUD

[ Vou. XLVIII

THE AMERICAN NATURALIST

“PPYyMoolg 4e UMOIH 9

13° G'9 | 90° = 68'T | 80°F 0°6Z| TES | 1 | 0)0/8 1/9 | b€| SF| 89! 68) EZG |9 E | T | 0€ SIGI Aq 1-(Z-21)

03° = T9°S | SO°F OST | LO°F 8937| 88I Com OM De er ie ee ee G S oi Oe TI61 t 1-(Z-Z1)

96°F S'S | LOF SIS OTET] SIS | | "FI. {T | 81] 61] $3} 83] 68) 28) OF | 9T| | I 8% O161 rA (@-gi)

63° = €9°9 | L0°= 09'T| OL FGFS) ETI eno A a SR a ee So oe ee VE 606T at | gI

secos sejossoososoejoooocs..’ 9 . I g g ele eeleweeleeetoerefesefoeeetien &Z OI6T Ki (I-21)

GG =ZL'9 | 90° €6'T | 60°F 2°82} OIZ G | G | IT} 02} S| 8F| OF| TE) csi g¢ jT md Be TT61 4a t-ren)

GT'=19'9 | 90°= P6'I | 80 =FT'637| 962 IT | 3| 8 | #2! E| 99) 69 | Shi FF) ST) |T Pt ae 161 Si I-(1-@1)

9g | og | 8 | sg | ze | Te | og | 6g | sz | 23 | 9% | ez | + | sz | zz | t3 |JUeIBGJO

A D ‘a's Y 1810; , seavery joj TAOID Pisis ‘ON

quv[q rod soavey Jo səquny azequnN A

‘NNOO ‘ATaIGWOOTG LV NMOUDH 5

‘SI ATINV,[ NI SHAVE'T JO UTAWON JO NOWMAATULSIG AONTOAOAY,T

XI GIAVL

0G’ = 67'S | GO’ SFT | 80°F OLS] OAT | Z| LT | G | 6L| 62] FS] SP) GT] FP Tt] Dey 83 TIGI 4q G-(Z-#8)

86°F FL | SOF SLT | 80°F TH] 82S [Le 1S «18 | £3) OF| 89) 6S) 4T;8 |S |" 9% OT6I cf (Z-¥8)

66° = IT'9 |20 = OF'T | 60°F 62S] OOT “oA Oe aed ek eee ek el PS 606T at | FE

6T°+IT'9| SO EFL | 90°F FES] IPS eerie 1 ST] PE] 249] 29) 28) 12) F | T BS O16. dt | (1-¥8)

03° = G8°S | SO°= LIT | L0°F SSS; 86I TIF )6 | 6S) tr Se GP oren n 0% TIGI t g-(T-78)

ze | 18 | og 9 +3 | ez | zz | IZ | 03 | GI | quaa8a Jo

‘A‘O a's ‘Vv 18101 pat Mal Bod Bees! səavə' JO pared Ba ON

quv[q 19d seavory Jo oquny aoquny x D

‘NNOD ‘aTaIaWOOTG LY NMOWD

‘pe ATINV,[ NI STAVTI AO WAWOAN AO NOILAJINISIA AONTAÒTAA

IIA YEV

No.565] CHANGES PRODUCED BY SELECTION 27

field is not clear. It is possible that the plants having

from 18 to 23 leaves were diseased, but no such condition

could be recognized in the field. Again, it is possible

that a few Havana plants were mixed in by mistake,

although as the leaves of the selection are characteris-

tically different from Havana and as the plants with low

leaf numbers resembled the remainder of the row, this

supposition is improbable. The most likely explanation

is that mutation occurred in a few gametes of the mother

plant, a condition that did arise, or that we assume to

have arisen, in Family 41 (see Table X). At any rate,

the change did not follow the path of selection.

In Figs. 6 and 7 are shown typical plants of Family No.

6 obtained by three years of selection in the effort to pro-

duce strains of high and low leaf number, respectively.

Family No. 34 (Table VIII) is peculiar—although this

is not the only time the phenomenon occurred—in that

the F, population grown from a 24-leaved F, plant seems

not to have given the true mean. Plants with a low num-

ber of leaves (22 and 20) were selfed to carry on the

minus strain, but both gave means higher than was shown

by the F, generation. Perhaps further selection will

produce results, but the case is not a hopeful one. The

only evidence for such an assumption is the increased

mean of the F, plus strain. If it is assumed that 24.0 is”

nearer the true mean of the F, population than the 22.9

actually calculated, then the jump to 27.0 + .08 leaves in -

the F, generation gives us a basis for expecting results in

F, in the minus strain.

N othing can be said as yet about the minus strain of

Family No. 12 (Table IX), for it happened that the first

selection was a complete failure. Six plants were ob-

tained, but the lowest number of leaves was 29. One of

these plants was selfed and gave an F, population having

a mean of 28.7 + .09 leaves. Unfortunately the selections

from this fraternity did not germinate and in 1912 we had

to fall back on the reserve seed from which the 1911 crop

came. The crops of 1911 and 1912 are therefore dupli-

cates. The plus strain made an advance from 24.5 + .10

28 THE AMERICAN NATURALIST [Vou. XLVIII

leaves to either 26.8 + .07 or 29.0 = .08 leaves. The first

advance is 1.6, the second 0.7. We can give no explana-

Fic, 6. PLANT OF Pegi sek joer ANA ToBAcco (6-2)-1-1, WHICH AVERAGES 30,2

LEAVES anes NT. IT IS THE RESULT OF THREE YEARS OF SELECTION FOR HIGH

LEAF NUMBER IN FAMILY ig WHICH gredis 25.8 LEAVES PER PLANT IN 1909.

NEw aans 1912.

No. 565] CHANGES PRODUCED BY SELECTION 29

tion of the failure of the results of 1911 and 1912 to dupli-

eate. This is the greatest deviation obtained in the course

of our experiments. The results of 1912 are probably

too high. It is yet too early to say whether or not this

. T. PLANT OF poean HAVANA pirn as 1)-1-1, WHICH AVERAGES 27.9

ase PER PLAN THREE YEARS OF DECREASE THE LEAF NUMBER

OF THIS TYPE ret PROVED anion tor gy ter: Haven, 1912.

30 THE AMERICAN NATURALIST [Vou. XLVIII

strain is decreasing in the average annual shift of the

mean.

Family No. 41 shown in Table X gave perhaps the most

peculiar results of any of the selections. It may be that.

no great shifting of the mean toward the minus end of the

curve should have been expected, because the minus

mothers were each rather high in number of leaves. There

was one with 25 leaves and one with 24 leaves. This was

unfortunate, but was made necessary by the number of

late and diseased (mosaic) plants in the selection. Never-

theless, each of these plants was below the mean of the

previous generation and if a marked change would have

followed the selection of extreme individuals, some change

should have followed the selections of the individuals that

were the actual mothers. But in spite of this fact the

mean persistently rose from 23.9 + .07 leaves to 26.3 + .08

leaves, then to 28.1 + .07 leaves, although the duplicate of

this selection grown in 1912 went down slightly to 27.4

+ .07 leaves. In the plus strain successive generations

of mothers having 28 and 30 leaves caused a small upward

shift of the mean; it became first 25.7 + .09 leaves then

25.6 + .14 leaves, although the 1912 duplicate of the last

population had a mean of 26.9 + .08 leaves.

The extraordinary phenomenon to which we wish to

call particular attention, however, is not this behavior of

the minus and plus strains in the regular selection ex-

periment, but rather the origin of a few-leaved strain

from a single individual that appeared in the F, genera-

tion of the plus strain. Referring to the table, it will be

seen that in this generation a 12-leaved plant appeared.

This is really a peculiar phenomenon, for we had never

before observed a normal 12-leaved plant among the many

thousands that have come under our observation. They

do not occur. In this population the plant with the next

lowest numbers of leaves had 20 leaves, and in classes 20

and 21 there was only a single plant of each. This 12-

leaved plant was selfed and gave rise to a population

ranging from 8 leaves to 30 leaves, and having a vari-

ability of 23.50 per cent. + .11 per cent. The mean of the

No.565] CHANGES PRODUCED BY SELECTION 31

distribution was 19.8 + .28 leaves. A 10-leaved plant of

this lot was selfed and gave a progeny with a mean of

17.9 + .08 leaves and a variability of 11.24 per cent. + .33

per cent. What interpretation can be given these facts?

We believe a distinct mutation occurred, a mutation

different from those of DeVries. At least DeVries be-

lieves that the mutations that he has observed always

breed true. If the following hypothesis as to the origin

of the 12-leaved plant be true, it is unnecessary to sup-

pose with DeVries that mutations always breed true or

even that they often breed true. Of course DeVries be-

lieves that his @nothera mutations obey laws different

from those of whose mechanism we know a little. He be-

leves that species crosses always breed true; that they

do not Mendelize. ‘This belief we hold to be unfounded.

Species crosses have never been shown to breed true.

There have been statements to the effect that crosses be-

tween Rubus species breed true, but no good evidence has

been submitted in their support; while the data of Tam-

mes (:11) on Linum species crosses, Davis (:21) on Œno-

thera species crosses, and of East (:13) on Nicotiana

species crosses, concur in showing that species as well as

varieties obey Mendel’s Law of segregation and recom-

bination. Furthermore, we think that Heribert-Nilsson’s

(:12) beautiful experiments on DeVries’s own material

show that the latter did not collect sufficiently exact data

on his own crosses to find out whether they bred true or

not.

If one is to believe that a mutation in a hermaphroditic

plant breeds true he must suppose that constitutional

changes occur both in the male and the female gam-

etes, or that the change occurs after fertilization. But it

seems more probable that such a change will take place

either in the one or the other gamete and not in both. This

we believe to be the explanation of the appearance of the

12-leaved tobacco plant. A mutation occurred in either

an egg cell or a pollen cell. It does not matter in which

one it is assumed because there is no evidence favoring

either case to the exclusion of the other. This cell with

THE AMERICAN NATURALIST [Vou. XLVIII

‘suOT}ejNduI0d UT pesn Jou ərəm sT, “ApOAT}odseI SOAVOTTZ PUB QT ‘OT 010q Orqa MOI STY} UT punog oIəMm syuud vueaRyy OI, rr

‘suorzejndurod ur pesn JON ot

“SOARS IUU pue qro Burrveq youve ouo ‘OsTY e

Smaa [10s puorpynsur 0} enp ews pue mores AIOA OOM squeyd oy} PUB MOI PU UB SEM SII, 8

8T'= 09'S | 0°= OST | L0 =F 8°93 ZEZ 9 | 2z] Iv] 09] 2¢|8z\o1!/z l0 |T L% TIGI 8a I-(@-99)

13° = 8T'9 | 90° SOT | 80°F L°9Z | 903 = | 8 | ¥S| 8G| 8F| IS | Z| 61S : 83 OT6TI 4a (3-99)

ST'=98°9 | SO°+ 99°L | 90°F SVFZ) SIE ~T 74S Jg | OF1 Shi 62162) SF tE St] Fy Tt 9% 6061 "A 1199

IZ’ = FF'9 | SOF 6ST | LO°F LFS | EIS TTS LO TE] PE PF | OS BS Otis 1 te ez OT6I tA (1-99)

Of = 19| SOF LTI | LOF EHS) 033 “1313 | 2 | 08) 98| FO| 8h; 1816 |S "ye" ez TIGI Sa I-(1-99)

CS FILS | PVF PPT | Gloss) Gs i a eed ae Wee Se AEB Sey tees: bake e ZI61 wf €-I-(1-99)

ze | Te | og | 6z | 8z | 23 | 9% | 9z | FZ | z | zz | IZ | oz | ET

‘AO ‘a's ‘Vv 101, Gone oe x Pros ‘ON

que[g red seave'y Jo 1oquiny zequmnN

"NNOO ‘TIMANI LV NMOUD

‘99 ATINV,[ NI SHAVE] A0 LAINAN AO NOMLAGIULSIG AONTAÒOTAT

IX WIaViL

€°+ FST | 90°+ 00'S | 80°F SLT | 843 [° ETT E Oa we ak e es 9e oT Or SIGI | a4 Z-8-(6-1)

T@°*OL°L | 90° 20° | 80°F 6'9Z| OLE | T| H/F [IZLE 6S O9|PSIGE/9TIOTIE |T |Z. x Be Bees ieee Wee os og ZI6I ta 1-(Z-19)

IT’ 09°EZ | 06° SOF | 8° 8'6T| 9T |") T 10 10 IF IA i8 16 19 PIS ITiS 18 |8 0 1€ |S) S| €89) St LOE: "a s8—-(Z-19)

ESFLYL | 90°F ZOT | 60°F 2°92 | FES ee eo L STITSTPIPPFISPIOE FSF T T || °° ort|"? 8z OT6I Le i (Z-19)

I3°+99'9 |90'F GLT | 80°F E°9S | PEZ F |TZ|SS\ LG FSi SE/O1 91s Ae te ie 5 L OTGI i (1-1})

8ST'=ZF'9 |SO'FOLT | LOF FLZ) 88S °° Elg [PZI9IEZIPLITSILTS E P T : , o Sie SIGI i g-(1-1F)

toe low ee teh R ueiBg Jo

yatq d T Jo 1ə3quny haps . Đ

‘NNOO ‘@IgIaNOOTG LV NMOUD

‘TP ATINV, NI SHAVAT JO UTAWON AO NOMMaMisiq AONTAÒTAT

X WTavi

CHANGES PRODUCED BY SELECTION

No. 565]

“PPYUOC[, 48 WMI gr

‘PIPpWOOJg 48 UMOID zt

LU CYP | ogros] POT r 18 eeraa eA Pe ye ae siet | “I | ceog)

61° = 16'F | 90° FT | 80°F 6'SST| GFT || T| 9 | eT] FS/T9|08/6L\ JT | cl’ Ze TIGI sa g-(6-99)

0G’ = 92°F | SOF OTL ISOS Le] SOL | A T e 2) ae OF St 2 E 3 8% OTGI ta (z-99)

ST'= 98°9 | GO°+99°T | 90°F SHS) STE Py Tg e | SL Sh] SL) 6L 6F| TE| ST) F | T 9% 6061 A 299

PS = 10'S | 90° L2°T | 60°F E'SZ| TOT SEEPI TYS Oe eee OP te ie ee E3 OTGI ia (1-99)

LU OS) SOF LUT | 90° 0900 | TST O Ye Or e ree eee ee ee TIGI "a T-(1-99)

66°F PTO) 90°F 6ST | SOS GSS) SAT OTT Tg 19 18i Zo) Shi tri R 12 1° 7° F SIGI 67 t-1-(1-09).

re | se | ze | te | og | 6z | sz | 23 | 9% | SZ | +3 | sz | zz | IZ | og | 6I |yueeq zo

SA'D ‘a's vV 1810, HIN No? Pend “ON

queja 19d seavary Jo LequnN ZOQ UIT

"SSV ‘STII, LSTAOT LV NMOAD

6S F OTL SOK LET FCS TL] Set rl EY) ea 8 SE Ie Oe ae eee ee aa a 0€ S161 ‘a | 2-1-(Z-99)

80° ZE°L | 80°F OT | I= F'97Z| TST | 1} 0} 0) 1) Z|? | OL! 03/63/78] S| 221g T Tror. 83 TI6T eg 1-(Z-99)

OS" = 8°9 | 90-+ OO T| SFE LET fd S | OL SB a oe ee eee et a 83 OT6I Aq (3-99)

ST’ 98°9 | S0°=99'T | 90°F SZ) GTE : AS 1G | 6L| &Fi S2Z| 62 Ori Iei ST] F | T 9% 606T 7 2199

VG = 9Z'9 | 90°F OST | 80°F OFS) BST || "yee"! 18 | FL18S|9r| 981 8TI9 | T 1°" w nori a T-(1-99)

02° = POS | S0 F EZT) LOF HHS) FHT mie ate ee Se e ee Se SG e €@ gwri Aa | T-1-(1-99)

3 (U ued o

EN B z pa se | +e | se | ze | 1e | ng | 6g | sz | zg | 9 | sz | +3 | ez | zz | tz | og | GT ren ze R a ion i

qus[q 1əd seave'y Jo Iəaquny

Iəquny Jgəðg -Bisuey

‘NNOO ‘NIAVH MIN LV NMOUDH

34 THE AMERICAN NATURALIST [Vou. XLVIII

a changed gametic constitution, —a loss of gametic fac-

tors,—was fertilized by an unchanged cell. The un-

changed cell may have had any of the gametic possibil-

ities open to the germ cells of the 28-leaved plant of the

F, family in which the mutation arose, and we know that

certain factors in this plant were heterozygous, for pro-

gressive change followed the selection of a plus extreme

in the next generation. The 12-leaved plant was there-

fore a hybrid. It resulted from the union of a mutating

germ cell of the mother plant that furnished the F, gen-

eration with an unchanged germ cell. We can even as-

sume that the mutating germ cell, if fertilized by another

of the same kind, would have produced a plant with less

than 12 leaves. The reasons for believing this are simple.

There is experimental evidence (Hayes, 1912) that the

F, generation of a cross between varieties differing in

their number of leaves is intermediate in character. Our

12-leaved plant is the lone representative of such an F,

generation. The F, generation therefore should give

plants with less than 12 leaves, and in fact such plants

did occur. The distribution marked Fa in the table is

the F, generation, and this accounts for its extreme vari-

ability. The distribution marked Fs is the F, generation,

and its variability is less than half that of the preceding

generation.

Family No. 56 was the second family to be grown at all

three of the experimental stations (Table XI). It arose

from a 26-leaved plant of the F, generation which pro-

duced an F, progeny with a mean of 24.2 + .06 leaves and

a mode at 24 leaves. The three generations of the minus

strain grown at Bloomfield remained practically the same.

The last generation did indeed show a mean 1.0 leaf

higher than the original population, but no dependence

ean be placed in data from only 25 plants. The data on

the minus selections grown at New Haven are for this

reason a little more dependable. They show a fluctuat-

ing mean, but no progress due to selection, the F, genera-

tion having a little higher mean than the F, generations.

The three minus selections grown at Forest Hills also

No. 565] CHANGES PRODUCED BY SELECTION 35

resulted in higher means, those for F,, F; and F, being

25.3 + .09, 26.0+.06 and 25.9+ .08 leaves, respectively.

This peculiar result implies only that the mean of the

original F, population which was grown at Bloomfield

was lower than it would have been if grown on the Forest

Hills’ soil. This is not a direct effect of environment on

the growing plant. It has been shown conclusively in

our pot experiments, as stated before, that starvation or

optimum feeding has scarcely any effect on the number of

leaves, although it has a marked effect on the develop-

ment of many other characters. On the other hand, en-

vironment does appear to have a marked effect on the

number of leaves that a plant is to develop, if it acts

during the development of the seed. It is well known by

plant physiologists that the environment produces many

of its effects very early in the life history of the indi-

vidual or in the development of the organ concerned. For

example, the so-called light leaves of the beech with two

layers of palisade cells are differentiated from the shade

leaves with only one row of palisade cells by the amount

of light that falls on a branch during the season preceding

the development of the leaves: that is, it is determined

during the laying down of the bud from which the next

season’s growth of twig and leaves comes. This period

during which a particular change is possible is called the

critical period for that change by plant physiologists.

Thus a plant may have hundreds of critical periods in its

ontogeny, each marking an end-point of development be-

yond which a certain feature is irrevocably fixed. For

example, the critical period for that cell division that de-

termines leaf size in the beech is much later than that

which determines the number of layers of palisade cells.

Now the critical period for influencing the number of

leaves of the tobacco plant is practically at an end when

the embryo plant goes into the resting stage of the seed.

Before that time the number of leaves may be influenced

by the external and the internal influences that form the

total environment of the mother plant; after that time

environment has little influence on the number of leaves.

36 THE AMERICAN NATURALIST [Vou. XLVIII

The rise in the mean of the population of the F, genera-

tion of Family No. 56 is due partially to the effect of en-

vironment, therefore, in that the mother plant was grown

under better conditions, but is probably not to any great

extent due to the conditions under which the plants them-

selves were produced.

The better environment of the mother plants does not

account for all the rise in the means in populations F,

and F, but it accounts for part of it. It will be noticed

that all of the populations grown at Forest Hills had

higher means than those grown at Bloomfield and New

Haven, although the F, mother plants were grown at

Bloomfield and not at Forest Hills. The greatest shift

of the mean, however, comes in the F, and F, generations,

for the mother plants of both of these populations were

grown on the more fertile soil. There is a simple ex-

planation of these facts, an explanation that is of great

economic importance to practical tobacco growers. A

part of the rise in mean at Forest Hills was due to set-

ting the plants in the field there when they were in an

earlier stage of development than those at Bloomfield and

New Haven. They were not set earlier in the season (at

least, one year they were set early, one year they were set

at the average time and the third year they were set late),

but they were set as small plants. When small plants

(about 4 inches high) are set in the open the root system

is equal to the task of supporting the aerial parts and the

plants start right in to growing normally. There is no

period of passivity. The plants produce leaves spaced

with normal internodes and these leaves develop suff-

ciently to have a commercial value. But when the plants

reach a height of 8 or 10 inches in the seed pans or seed

beds and are then set in the field, the normal metabolism

is likely to be upset for a time. The plant takes some

time to recover its equilibrium and start a normal growth.

During this period basal leaves begin to develop, but the

internodes are so close together that they do not obtain

their aliquot share of nutriment, hence they grow only to

one quarter or one third their normal size and soon wither

No. 565]

CHANGES PRODUCED BY SELECTION 37

and drop off. The leaf scars are left, but they are so

close together that it is difficult to make a correct count of

the number of leaves.

Fic. 8. ANT OF HALLADAY HA-

VANA Topacco (56-2)-1-1, WHICH Av-

ERAGES 27.5 EAVES LA T

IS TH ESULT OF THREE YEARS OF SE

CTION FOR HIGH EAF UMBER IN

Fa 6, WHICH IN 1909 AyPRAGED

24.2 Leaves Per PLANT. New HAVEN,

12.

But more important than this,

Fic. 9. PLA HALLADAY Ha-

NA [iepa (se B: 1-1, WHICH Av-

GE NT. THREE

Low AF

VED eiliseen

New Haven, 1912.

[Vou. XLVI

THE AMERICAN NATURALIST

PE" BTI | 60°F SOT | ETSL) 92 ||” Tc" Fe PSL IS Sr erie 11 | tia See sie | a oera

22° = LE°9 | 90°F OS’ | 60°F S8S| SET | I | Z| S| Il] Fs mise te GT ITs | | TI6T | E €-(Z-€2)

G3 = 899 | 90°F ELT | SOF E'9S| 10S |)" ieS | PT] 63| PH] SF] FE} FT/6 |01/0/0/T 63 OI6T | 4H (Z-€2)

SI’ 68'S | 40°F SFT | 90° 6°93} 008 |) | TIT |P | 66) #2) 8819910 STIS | cyl] 8 6061 | %4 (JA

GZ = F94 | L0°* 103| 60°=£'9Z| 9IZ IJJI |8 | 81] 68) 9F| SF] 62) Sc) OT] | 2] 1 I" 9 OI6T | 4a (T-€2)

£6°= ZTA | 90°F E81 | 80°F 2°93) Tes |||" IT | S1/€%| Ze] 9} 68/081 6T) Si ell] cter | ta | 1-@-(1-€2)

FE | E£ | ZE | TS | OF | 6B | 8% | LZ | 9 | 93 | FS | ES | 33 | TZ | OS | 6T

AO ‘a's Yo |n convey jo| TA0TĐ | uone OX

yurd aed S8ABaT JO qequinn Jequin ny A °

‘NNOO ‘ATIANOOTg LY NMOUD

‘¢L ATINVG NI SHAVE'T JO AAAWOAN dO NOILAJINISIA AONTAOAY,T

IIX WIaViL

EZ'= GT'L| S0 = SOT | 40°F 8'SZ| ESS | T | 0} TT! 0/9 | 18\9F| F9|8F| 62) STIT |T] 9% ZI6I 8 9-1-(@-)

02° = £0°9 | GO°F SFT | LO°F OFS | P03 ‘| TL | t |] 8 |8t| 6r| Zo; Io gre |O |O |; T/T 9% TIGI ta 1-(2-S1)

Ge * 86'L | 80°F ZBL | II" 8'ss| OFT || T| 3] S/T |6 | 68) 8) zisio |] oz OI6T si (wD

OF + 29°F | 60" SOT | ET FEZ) TE RD te te Se eee 9% 6061 Ci M

LZ = 6'9 | L0°* 69T | 60°F FFS| SET || Z| & | 6 |E zeltei STi 2 p i ~" (2)08 OI6T si (1-9)

Zg =Zg'9 | GO’ EST | LO°F SES) Zog || T | T | S| 9T| 83|9t|Z9|6g|71|E Bed ica Z3 TIGI 4 I-(I-y)

92° = 11'2| 90° = L9°T | SOF SIZ) 603 YT] og |9 | ST] 98] S&F] SS] shi ST] c |Z #3 ZI6I sq &-I-(1-) ©

08 | 6% | 8% | £3 | 93 | 9 | FS | 88 | ZZ | TS | OG | 6T | ST | LT | JUeIB Jo

A'O a'S y IIOL SOABO'T JO D RA “ON

quva 13d seavey Jo sequin TequInN A 9

‘NNOO ‘CIMIAWOOTgJ LY NMOUH

‘SJ ATINV,[ NI SHAVU JO WAWAN AO NOITAIINISIA AONTAdTA,T

IIX WTaviL

No.565] CHANGES PRODUCED BY SELECTION 39

the tobacco grower loses an average of from one to two

of his most valuable leaves.

The plus strain of Family No. 56, which we were dis-

cussing when we digressed to speak of the critical periods

of development, did show a considerable shifting of the

mean following the selection of high-leaved mother plants.

In the Bloomfield selections the mean went from 24.2

+ .06 to 26.7 + .08 leaves, then to 26.8 + .07 leaves; in the

New Haven experiment the mean shifted to 27.4+.08

leaves,—a gain of 3.2 leaves,—and then dropped to 26.4

+.11 leaves, recovering again in the F, generation to

27.5 + .11 leaves; in the Forest Hills experiment the suc-

cessive means were 27.2+.08, 28.9+.08 and 26.7 + .06

leaves. Summing up the data from this experiment, it

may be assumed to be reasonably certain that no progress

resulted from the selection of minus extremes, but that

there was a slight effect gradually diminishing in quan-

tity when plus extremes were selected.

Representative plants of Family 56 obtained by three

years of selection in the effort to produce strains of high

and low leaf number, respectively, are shown in Figs.

8 and 9.

Family No. K (Table XII) was grown on a farm near

the Bloomfield experiments, in 1910. The records of the

F, generation consisted of the number of leaves of only

31 plants. From among these individuals two plants

were selfed to become the mothers of the F, generation.

Since no dependence can be placed on the F, distribution

by reason of the few plants and since it is not absolutely

certain that the mother plants of F, had 20 leaves each,

the selection really began in 1911 with the F, generation.

There is a difference between the minus strain and the

plus strain in 1911 and 1912,—0.5 leaves the first year and

1.3 leaves the second year,—however, so that one may

assume the possibility of a slow shifting of the mean in

both directions.

The data on Family No. 73 are shown in Table XIII.

This family came from a 28-leaved plant, one of the

highest of the F, generation. The F, progeny of this

[Vou. XLVIII

THE AMERICAN NATURALIST

FI’ = 80'S 40° = LE'I G0" = 9°93 263 T |T | 91] 09/16] 2/88] ST} Z 0g 606T | %Z z8

GS" = EEL 20° 96'T 60° = 9°92 40Z € | &1| GZ| OF | 9E| 1F| 1/13} F |S) o0)\T 83 OTt6T rr (Z-@8)

LV LG'S SO = OF'T 90° + 9'93 LIZ T | ¥ TOi Ze) 28} 89) Si Tk) Sp 0€ TIGI 1-(Z-@8)

IF PLO 90° 08'T 80° 2'93 Sco || £ |8 | 02| SF) 09| 8F| 12) €T; 9) ¢)T oE ZI6I a 1-(Z-@8)

ze | 18 | og | 6% | sZ | 23 | 9% | SB | HZ | SB | zz | T3 | 03

A3 ‘a's Y PL moa wot 50 ar f ‘ON

quu[g ied seave'yT jo requiny aaquny

‘NNOO ‘ATaIaWOOTG LY NMOUD

‘48 ATINV NI SHAVE'T JO UAAWON dO NOWMAaTALsIq AONTNOAY,T

AX WTAViL

8I'* 89°9 G0" O'T 90°+ O'S | SOE | T | 0/0 |S | FT \6S | $2| Z| TS| 88) FT| 9 8% GIGI | 4a 1-(Z-22)

61° + 8Z'¢ GO" OF'T L0°* 0°92 Soe |" |) SEL 28 168 | 88i Zi esi ec |r | t 8% TIGI | 4a 1-(Z-23)

IG = L¥'9 SO’ E'I L0°* BSS GES Te | QOT/IE | Sh) 99] 6887|22 |8 83 OT6T 7 (3-13)

PI’ £0'°S ¥0' = OF'T 90° 8'LE | 008 T |6 | 02| 9F|60T| 29| SE) IT} % |Z |" ** 63 606I | %4 Lz

CSF BEL 90° = F6'T 80°F S'9S | #9 P | 31 | Z| 6Z G9 | OF | 18| €Z| ST} e |T gz otet | I (1-22)

9T'= 10'S POZEI 90897 | FIS [°°] 11S 1s e A SE) OL F e °° Z3 1161 | ʻa t-

8I'’= 9'9 SOFII L0°*6°SZ | OIE € |Z | 9T| 48/19 | 69/ €9| FF) 2/8 OJT d SIGI | 4a 1~(I1-22)

ze | te | og | 6g | 82 | 43 | 93 | 9z | #3 | £z | zz | IZ | og | uerBg Jo

AD ‘a's V T80L s ; asiani 1 O T ea BR ‘ON

queja 19d seavery Jo 1əqunN ae a z

‘NNOO ‘ATIIANWNOOTg LY NMOUDH

‘L3 ATUWV, NI SAVTI dO UAINAN AO NOILAIIISIA ZONTADTAIE

AIX WTavViL

No.565] CHANGES PRODUCED BY SELECTION 4]

individual showed a mean of 26.9 + .06 leaves, and from

among them plants having 25 and 29 leaves, respectively,

were selected to start the minus and the plus lines. These

two mother plants gave F, populations alike as to mean,

but differing by one class as to mode. The minus line

had the higher mode. The extremes of this generation

used in carrying on the experiment differed by 8 leaves,

and the resulting progenies apparently followed the selec-

tion. The means are 25.6+.07 and 28.2+.09 leaves.

Whether these shifted means represent a permanent

change or not we are not prepared to say. The minus

mean is probably somewhere near the correct figure for

in the F, generation it was practically the same, but in

the F, generation of the plus strain the mean dropped

from 28.2+.09 leaves to 26.7+.13 leaves. This is a

slightly lower point than that of the original F, distribu-

tion, but it was calculated from only 76 individuals. A

conservative estimate of the significance of the results

would probably be as follows: the mean of the minus

strain has shifted slightly but permanently and is now

fixed, while the mean of the plus strain has not changed

but has shown evidence of some heterozygosis in one gen-

eration.

We come finally to consider Families No. 27 and No. 82,

the data on which are listed in Tables XIV and XV. Two

generations of both plus and minus selection were re-

corded for Family No. 27, but only plus selections of

Family No. 82 were grown. There is no necessity for

considering either in detail because a simple inspection of

the tables shows that selection has accomplished nothing.

CONCLUSIONS

The cumbersome and no doubt dry details of the ex-

periments to the close of the year 1912 having been de-

scribed, let us give a brief résumé of the conclusions that

we believe may reasonably be drawn from the data that

have been offered. There can be no doubt that the orig-

inal ‘‘Halladay’’ type of tobacco, isolated and propa-

42 THE AMERICAN NATURALIST [Vou. XLVIII

gated by Mr. Shamel and Mr. Halladay from the cross

between ‘‘Havana’’ and ‘‘Sumatra’’ tobaccos, arose

through the segregation and recombination of the Men-

delian factorial differences of the two plants, and not as

a mutation. It is simply a union of the factors that stand

for leaf size and height of plant in the ‘‘Havana”’ variety

with the factors that bring about leaf shape and high

number of leaves in the ‘‘Sumatra’”’ variety. It hap-

pened that the somatic characters of these varieties ac-

count for all the characters of the hybrid. At the same

time one must remember that strains were obtained by

selection that averaged higher in number of leaves than did

even the‘‘Sumatra’’ parent. We can only conclude from

this fact that the difference between the ‘‘Havana’’ and

the ‘‘Sumatra’’ varieties in leaf number is greater fac-

torially than somatically. Besides certain factors com-

mon to the two varieties, the factors for leaf number in

‘‘Havana’’ tobacco might be represented by the letters

AA, and those of ‘‘Sumatra’’ tobacco by the letters BB,

CC, DD, EE. By recombination, this would give plants

with a smaller number of leaves than the ‘‘Havana’’

variety and plants with a greater number of leaves than

the ‘‘Sumatra’”’ variety. Both combinations were ob-

tained; and further, the theory has been shown to be cor-

rect by the results of other crosses where both types ap-

peared (Hayes, 712). It is probably unwise to suggest too

concrete a factorial analysis of the cross, yet the factorial

difference assumed above will account for all of the facts

obtained, by simple recombination. We assume a factor

in the heterozygous condition to account for the produc-

tion of one leaf and a factor in the homozygous condition

to account for the production of two leaves. The mean

of the ‘‘ Havana’’ variety is about 20 leaves and the mean

of the ‘‘Sumatra’’ variety about 26 leaves. Somatically

there is a difference of 6 leaves or three factorial pairs

for which to account. But in order to have the theory

coincide with the facts there must be at least one (pos-

sibly two or three) factorial difference that does not show

in the two varieties. The meaning of this statement can

No.565] CHANGES PRODUCED BY SELECTION 43

be shown best by an illustration. The 20 leaves of the

‘‘Havana’’ variety and the first 20 leaves of the ‘‘Suma-

tra’’ variety are represented by 10 pairs of factors, of

which nine are the same and one different in the two

strains. The ‘‘Havana’’ variety is nine leaf factors plus

AA, the first 20 leaves of the ‘‘Sumatra”’ variety are nine

leaf factors (the same as those in the ‘‘Havana’’) plus

BB. The additional leaf factors of the ‘‘Sumatra’’ are

CC, DD. EE. With these assumptions, the recombina-

tions of a tetra-hybrid will represent our facts fairly

accurately. But, as was stated above, it does not seem

wise to take this interpretation of the facts too literally.

That some such factorial combination will represent our

facts superficially there can be no doubt, but in reality if

one could grow hundreds of thousands of individuals and

follow the behavior of each he would likely find himself

constrained to represent his breeding facts by a much

more complex system. There would probably be gametic

couplings and factorial differences whose main effect

would be on some entirely different character or complex

of characters, but which would have some slight jurisdic-

tion over leaf determination. To become diagrammatical,

the unit characters of a house are its cornices, its win-

dows, its floors and what not, but a collection of these

components is not a house. We may even exchange

dormer windows with our neighbor, but we can exchange

them only if they fit. Again, we may put on a coat of

paint, a color unit, but this color unit affects the appear-

ance of many other parts that are just as truly units.

The essential part of our conception of the origin of

this hybrid type is that recombinations of characters

quantitative in their nature can be expected and predicted

in crosses in exactly the same manner as is done with

qualitative characters. On the other hand, it must be

borne in mind that here was a hybrid type that appeared

to be breeding true to the general characters that we have

described, in the F, generation. That it was not breed-

ing true is clear from the results of the selection experi-

ments, yet out of the small number of F, and F, families

44 THE AMERICAN NATURALIST [Vou. XLVIII

taken under observation at least two were found to be

breeding true for all practical purposes in the F, and Ë,

generations. We were able to reproduce the ‘‘Havana’’

type by continued selection in Family 77 and were able

to produce strains breeding approximately true to 30

leaves or so by the selection of mother plants in several

families. But can we say that any of our families are

now fixed so that no progress can be made by selection?

We can not. But we can say that some of them are so

constant that it would be a loss of time for selection to be

continued for economic results. It is important to know

whether plant or animal populations can reach such a

state of constancy by inbreeding that no profitable results

can afterwards be obtained by the practical breeder. We

believe it demonstrated by even these few data that such

a state, a homozygous condition, occurs in a definite pro-

portion of F, offspring, and can be propagated commer-

cially at once if a sufficient number of families are grown

to be relatively certain of including the desired com-

bination.

As to the problem of theoretical importance, the ques-

tion of the true constancy of homozygotes generation

after generation, we believe it to be fair to conclude that

a state so constant is reached, that even for the theoret-

ical purposes of experimental genetics it may be assumed

as actually constant. Further experiment and larger

numbers may show that selection can always cause a shift

in the mean, but will necessarily be a shift so slight that

it can be detected only by a long-continued experiment

and enormous numbers. Assuming for the purpose of

argument that this is the case, the matter would affect

only the question of the trend of evolution. It may come

to be believed, from evidence now unknown, that evolu-

tion may progress slowly in this manner, but if it does,

its course can hardly be demonstrated experimentally be-

yond a reasonable doubt. The problems of experimental

genetics can be attacked, however, from the standpoint

that experimental evidence of the shifting of the mean of

a homozygous population by selection is negligible.

No.565] CHANGES PRODUCED BY SELECTION 45

Mutations may occur. We have shown the origin of

one family by a very wide mutation. In this particular

case it was not difficult to show that a constitutional

change took place in a single germ cell of the mother

plant. It was only by a lucky chance that this fact could

be demonstrated, for with smaller changes such proof

would be impossible; but there is no reason to believe that

this phenomenon is unique or even rare. It is much more

reasonable to assume that mutations usually arise in

single gametes than that the same change occurs simul-

taneously in many germ cells. One should expect the

somatic result of a mutation in an hermaphroditic plant

—the sporting plant itself—not to breed true, therefore,

but to behave as an F, hybrid between a mutating and an

unchanged germ cell. It is true that the mutations ob-

served by DeVries in @nothera Lamarckiana are sup-

posed to have bred true, but this is sometimes question-

able even from DeVries’s own data. The Lamarckiana

‘‘mutants’’ that did breed true are much more reason-

ably explained as segregates from complex hybrids.

They can be interpreted by Mendelism with no essential

outstanding facts, but if they are to be interpreted as

mutations, several discrepancies between what actually

occurred and what should be expected on DeVries’s own

theory must be explained. It must be shown why the

changes took place in numerous germ cells,—in both the

male and the female gametes,—and why these germ cells

always fused at fertilization; for the changed germ cells

must have fused with each other because many Lamarck-

iana plants were produced by the same mother plants that

produced the mutations, while the mutations are sup-

posed to have bred true. On the only other possible theory

of mutation, that the change occurred in the developing

zygote after fertilization, one would have to explain why

the mutants did not often appear as bud variations, in-

stead of these being much rarer than the supposed muta-

tions, as is actually the case.

We do not deny the theory of mutation as modified to

46 THE AMERICAN NATURALIST [Vou. XLVIII

assume only that constitutional changes usually occur in

the germ cells, but on this belief the sporting plants must

often be F, hybrids, and the plant breeder must resort to

selection to isolate his pure mutation. And by the same

reasoning one gametic change may produce many new

creations, for there is a chance to recombine it with all the

known gametic differences in the species.

No one can say how often mutations arise. It is likely

that changes other than the one observed took place in

our tobacco experiments, but it is not likely that they

are sufficiently numerous to base a system of selection

within a pure race on the possibility of their occurrence.

The fact that no changes ensued that could be detected in

several of our selected lines is an argument against it.

The comparatively large jumps are the ones likely to

have the greatest economic importance, and these are

easily detected without refined methods of procedure.

Small jumps can be economically important only if they

are numerous, and, as there are absolutely no data to

show either that they are numerous or that changes can

be produced rapidly within homozygous pure lines through

any other cause, it seems unwise to recommend that the

practical breeder expend time and money to bring about

results that either can not be expected at all or that are

so slow and so trifling that they can not be detected in

carefully planned and accurately executed genetic inves-

tigations. On the other hand, the results of the last de-

cade show that important economic results can be ob-

tained easily and surely by selection from artificial hy-

brids or from the natural hybrids that occur in cross-

fertilized species by the recombination of Mendelian

factors. We believe, therefore, that the isolation of ho-

mozygous strains from mixtures that are either mechan-

ical or physiological, that are either made artificially or

are found in nature, offers the only method of procedure

that the practical plant breeder will find financially

profitable. —

Finally, we should like to call attention again to the

No. 565] CHANGES PRODUCED BY SELECTION 47

practical importance of determining the duration of the

period in the course of which particular plant characters

are responsive to the action of environmental influences.

The character complex that has been the basis of this

study is a striking illustration of how results from such

investigations may be applicable to farm practise. One

may plant a portion of the seed from a self-pollinated

tobacco plant on poor soil or on good soil and the average

number of leaves per plant and the general variation of

the plants in number of leaves will remain nearly the

same in both cases.? But seed selected from mother

plants grown on the good soil will produce plants aver-

aging slightly higher in leaf number than the plants com-

ing from seed on mother plants whose environment is

poor. Consequently, it is better to select seed from well-

developed mother plants—mother plants whose environ-

ment has been good—than from mediocre mother plants.

There is no question here of the inheritance of an acquired

character or of continuing to raise the number of leaves

by cultural treatment. One simply takes advantage of

the fact that during seed formation there is a period of

mobility at which time the potential number of leaves of

the young plant are practically fixed. Pending the end

of this critical period, the number of leaves can be in-

fluenced by external conditions within the limit of fluctu-

ating variability.

In the same connection, the effect of time of planting

on the tobacco plant should again be mentioned, as this

also emanates from environmental change. The actual

number of leaves is, of course, practically fixed at the

time of setting the plants in the field, but this is not true

of the number of leaves that will have a commercial

value. For example, a seedling with 26 potential leaves

is planted. If it is planted when about four inches high,

the general physiological disturbance due to transplanta-

tion is negligible and the plant continues its normal cycle

of development without a pause, bringing to maturity

2 Garner’s (:12) results on Maryland Mammoth are an exception to this

statement because this variety is indeterminate in growth.

48 THE AMERICAN NATURALIST [Vou. XLVIII

about 22 leaves. If planting is delayed until the seedling

is eight or ten inches high, there is a different state of

affairs. Development is arrested, the plant pauses to ad-

just itself to the change. It soon recovers and continues

its normal ontogeny, but the period of reduced growth

has left an ineffaceable record. Several of the leaves—

among them the more valuable leaves—have been so

' affected during this readjustment, that they develop to

only a fraction the size that they should attain because

the internodes between them are so short, due to the con-

stricted development that normal metabolism does not

occur. Thus there is a loss of one or two leaves, which

on several acres of tobacco may make the difference be-

tween profit and loss. Hence, the grower should not de-

lay setting his plants in the field until they have become

overgrown in the seed bed.

March, 1912

LITERATURE CITED

Davis, B. M. Genetical Studies in Œnothera, III. AMER. NAT., 46: 377-

427. 1912.

East, E. M. Inheritance of Flower Size in Crosses between Nicotiana

Species. Bot. Gaz., 55: 177-188. 1913.

East, E. M., and Hayes, H. K. Inheritance in Maize. Conn. Agr. Exp. Sta.

Bull. 167: 1-142. 1911.

Garner, W. W. Some Observations on Tobacco Breeding. Ann. Rpt. Amer.

: 458-468. 1912.

Jennings, H. S. eredity, Variation and Evolution in Protozoa, I. Jour.

Exp. ee 5: 577-632. 1908.

redity, Variation and Evolution in Protozoa, II. Proc. Amer.

Phil. See 47: 393-546.

rtive Mating, Variability and Inhéritance of saree in the Con-

fagation "at Paramecium, Jour, Exp. Zool., 11: 1-133. 1911.

Johannsen, W. Uber Erblichkeit in Populationen und in reinen Linien.

Jena, Gustav Fischer, pp. 1-515. 1903.

Pearl, Raymond. Inheritance of Fecundity in the Domestic Fowl. AMER.

Nart., 45: 321-345. 1911.

Shamel, $ D. New Tobacco Varieties. Yearbook U. S. Dept. Agr., 1906:

907.

Tammes, Tine. Das Verhalten fluktuierend variierender Merkmale bei der

Bastardierung. Rec. Trav. Bot. Néerl., 8: 201-288. 1911

GYNANDROMORPHOUS ANTS DESCRIBED DUR-

ING THE DECADE 1903-1913

Proressor WILLIAM MORTON WHEELER

Bussey INSTITUTION, HARVARD UNIVERSITY

In 1903 I described six gynandromorphous ants and

reviewed the previously recorded cases, seventeen in

number. Although many thousand ants have since passed

through my hands, I have failed to find any additional

cases. Other observers, however, have been more for-

tunate and have described seven within the past decade.

As these are all very interesting, it seems advisable to

give a brief account of them as a sequel to my former

paper.

1. Lateran GyNANDROMORPH OF CARDIOCONDYLA BATESI

FOREL. VAR. NIGRA ForEL.—Santscui (1903, p. 324,

Fig. 5, 7)

This specimen is female on the right and partly male

on the left side. The male portions are sharply marked

off from the black female portions by their testaceous red

color. The line of demarcation, very clear in front, starts

at the anterior clypeal border and divides the head into two

nearly equal parts, but leaves the median ocellus on the

male side. It then divides the pronotum down the middle

and the three anterior quarters of the mesonotum. Thence

the line fades out on the right side so that the whole pos-

terior border of the mesonotum is male. Three quarters

of the prescutellum and the anterior half of the scutel-

lum are male. The epinotum and the abdomen are female

throughout, but the female genitalia are slightly asym-

metrical on the left side. The fore and middle legs on

this side and a portion of the mesosternum are male.

There are wings on both sides, but the anterior one on the

female side was lost after capture. Those on the left

50 THE AMERICAN NATURALIST [Vou. XLVIII

side are well-developed, with distinct venation and pale

pterostigma, and are inserted in a distinctly male area.

The specimen was not dissected.

Santschi found this ant in a nest with females at Kai-

rouan, Tunis, but without males, either of the winged or

of the ergatomorphic type, which is peculiar to this and

some of the other species of Cardiocondyla. His atten-

tion was attracted by the bizarre movements of the speci-

men, as it turned around rather quickly in circles about

10 cm. in diameter, with the male portion inside. In

other words, owing either to the asymmetry of its brain

and visual organs or to differences in the length of the

legs on the two sides of the body, it made circus move-

ments like a normal insect which has had one of its eyes

or optic ganglia injured.

2. LATERAL GYNANDROMORPH OF ANERGATES ATRATULUS

SCHENCK.—ÅADLERZ (1908, p. 3, Fig. 1, a, b, c, d and f)

An imperfect lateral gynandromorph, with the head

largely male on the left, female on the right side, the light

color of the male being sharply marked off from the dark

color of the female only anteriorly. Thorax in front

female, with wings equally developed on both sides (the

male Anergates is wingless and pupoid!), but with pale

(male) coloration on the left and dark (female) colora-

tion on the right side, the line of division between the

two neither sharp nor straight and the whole postscutel-

lum blackish brown. Abdomen with irregular arrange-

ment of color. Petiole black on the right, grayish yellow

on the left; postpetiole mostly blackish brown, but with

a large grayish yellow spot on the left side of its anterior

surface. Third dorsal tergite blackish brown on the right,

grayish yellow on the left side. Remainder of gaster

grayish yellow, tinged here and there with pale brown.

Third tergite with a median longitudinal groove which

runs back on to the succeeding segment as in the virgin

female. The left side of the abdomen has seven com-

plete segments and well-developed genitalia; the right

No. 565] GYNANDROMORPHOUS ANTS 51

side has only six complete segments and a membranous,

incomplete seventh. The genitalia on the right side are

imperfect, the volsella being represented only by a piece

corresponding to its dorsal portion and the stipes is com-

pletely lacking. The legs are of the female type, except

the left fore leg, which is male, although the tibial spur

(strigil) is pectinate as in the female. This spur is known

to be nonpectinate in male Swedish, but pectinate in male

Swiss Anergates specimens.

On dissecting this specimen, which he took from a large

Anergates-Tetramorium colony near Arkösund in Oster-

götland, Sweden, Adlerz found on the left side a well-

developed vesicula seminalis, receiving a vas deferens

half as long. No traces of female reproductive organs

nor of the poison gland and vesicle could be detected.

Of particular interest was the behavior of this gynan-

dromorph, because, as Adlerz says, it evidently felt itself

to be a male but was treated by the normal males in the

colony as a female. Its movements were somewhat live-

lier than those of normal males, and it at first made feeble

attempts to copulate with the females and was treated

with indifference by the males. A few days later it be-

came more energetic and persistently attempted to copu-

late, especially with one particular female, although

always unsuccessfully while it was under observation.

It was evidently inflamed with the insatiable sexual appe-

tite so characteristic of the normal Anergates males,

- which, being wingless, always mate with their sisters be-

fore they fly out of the parental nest. On the following

day, however, a normal male made the most persistent

efforts for several hours to mate with this same gynan-

dromorphous individual. Adlerz concludes that

this indicates that the males regarded it as a female. Of course, we

may suppose that its wings made it seem like a female and attracted the

male, but from the fact that males attempt to mate even with female

pupae and therefore with a stage which has not yet developed wings, it

18 more probable that the male was attracted to the gynandromorph by

Some female odor. At any rate the double nature of the gynandromorph

52 THE AMERICAN NATURALIST [Vow. XLVIII

is even more strongly indicated by the facts just recorded than by its

morphological peculiarities.

3. LATERAL GyNANDROMORPH OF ANERGATES ATRATULUS

ScHENCK.—ADLERZ (1908, p. 5, Fig. 2, a, b, c, d and e)

An imperfect lateral gynandromorph, male on the left,

female on the right side, resembling the preceding speci-

men, but with the dark female color more pronounced on

the male side of the head. There were well-developed

wings on both sides of the thorax, which was of the female

form though dark on the right and pale on the left side,

except the epinotum, which was grayish yellow through-

out. Abdomen in color and form almost typically male,

with the genitalia well-developed on both sides, but with

a feeble mid-dorsal impression recalling the condition in

the virgin female. Legs of the female type, except the

left fore one, which is somewhat shorter and thicker as in

the male and with the tibial spur (strigil) cleft but not

pectinated.

Dissection showed the reproductive organs to be in the

same condition as in the preceding specimen; i. e., they

were present only on the left side and ontot of a

rather large vesicula seminalis with its vas deferens. No

traces of female reproductive organs, nor of a sting or

poison apparatus were to be found.

This specimen was taken from the same nest as the

preceding.

4. LATERAL GyNANDROMORPH (ERGATANDROMORPH) OF

FORMICA SANGUINEA LATREILLE.—DONISTHORPE (1909,

p. 464, Fig. 1)

A nearly complete lateral ergatandromorph, with the

right antenna, mandible and eye, and right and median

ocellus male and the left antenna, mandible, eye and ocel-

lus of the worker type. Head black, except the left

mandible, left half of clypeus, left cheek and a small patch

in front of the eye, which are red. Thorax and petiole

No. 565] GYNANDROMORPHOUS ANTS 53

male on the right, worker on the left, the line of division

running to the left of the median line so that the black of

the right side of the mesonotum encroaches on the red

color of the left side. Petiole and gaster sharply divided

into black right and red left halves, the right half of the

latter also with male pilosity and sculpture. External

male genitalia and anal sternite on the right side. The

red and black coloration is sharply divided on the venter,

but the coxe are all black and red as on the male, and the

legs on both sides are somewhat infuscated. Tarsi longer

on the right (male) side. Wings well developed, on the

right side only, with pale veins and stigma and more like

those of the female. Length 7 mm.

This specimen was taken by Mr. Donisthrope July 20

or 21 from a large colony in Bewdley Forest, England.

5. LATERAL GyNANDROMORPH OF FORMICA SANGUINEA

LATREILLE.—DonsTHORPE (1909, p. 464, Fig. 2)

A nearly complete lateral gynandromorph, male on the

left, female on the right side. The head is of the female

type, rather small, with both of the antenne and the ocelli

female and the left eye a little larger than the right.

Head black, clypeus and right mandible red ; thorax evenly

divided into a black left and red right half, but only the

upper right corner of the epinotum red. A piece of the

scutellum and postscutellum red on the left side where

the wing is inserted. Petiole sharply divided into a red

right and left black half. Gaster black, the pilosity and

sculpture on the right half female, on the left half male,

the color being sharply defined on the venter. Legs and

cox female on the right, male on the left side. External

genitalia of the male type present on the left side. Both

pairs of wings fully developed, but the stigma and veins

darker as in the male. Length 9 mm.

This specimen was taken from the same colony as the

preceding,

54 THE AMERICAN NATURALIST [Vou. XLVIII

6. FRONTAL GYNANDROMORPH OF SOLENOPSIS FUGAX

LatremLLe.—Santscut (1910, p. 649)

The head and thorax in this specimen are female, the

pedicel and gaster male. The head is somewhat smaller

than in normal females. The copulatory organs are those

of the normal male. Santschi remarks that it ‘‘ would be

interesting to observe the sexual behavior of such an indi-

vidual possessing a female brain and male genitalia.’’

7. LATERAL GyNANDROMORPH (EDRGATANDROMORPH) OF

MYRMICA SCABRINODIS NYLANDER.—DONISTHORPE

(1913, p. 44, Pl. I)

A nearly complete lateral ergatandromorph; worker on

the right, male on the left side, the former being blackish,

the latter reddish yellow. Right half of head larger than

the left, but with a smaller eye, striatorugose; right an-

tenna yellow, with a three-jointed club, its scape with the

usual strong lateral tooth at the basal flexure. Right

half of thorax yellow, its epinotal half with a strong spine ;

right half of petiole and postpetiole yellow, rugose and

punctured ; right half of gaster pale fuscous yellow. Legs

on the right side of the worker type, yellow. Left side of

head blackish, punctate, not striatorugose, with a larger

eye and the median and left ocellus; its antenna fuscous,

with four-jointed club. Left half of thorax blackish, its

epinotal portion unarmed; left half of petiole and post-

petiole smooth, fuscous black. The greater part of the

left half of the gaster had been eaten away but the re-

mainder was darker fuscous than the right. Legs on left

side of the male type, fuscous ; wings on the left side only.

Donisthorpe remarks that this specimen, which was

picked up dead by Mr. Dollman at Ditchling, England, ap-

proaches the var. sabuleti Meinert in having the left

antennal scape longer than in the typical male scabrinodis

and the tooth on the right antenna large.

In conclusion I would call attention to a peculiar ant

described by Mayr (1868, p. 60) from the Baltic amber

No. 565] GYNANDROMORPHOUS ANTS 55

and designated as a ‘‘Zwitter’’ (gynandromorph) of

Hypoclinea constricta Mayr, or Iridomyrmex constrictus

as we must now call the species. Through the kindness

of Prof. A. Tornquist, of the University of Königsberg,

I have been able to examine this specimen in connection

with many other amber Formicide. The general struc-

ture of the head, thorax and gaster is that of a worker,

though the thorax is not typical, as the base of the epino-

tum is less convex and less abruptly elevated, so that the

angle between it and the declivity is less pronounced in

profile. Mayr does not mention that the eyes are decid-

edly larger and more convex than in the normal worker

and therefore more like those of the male. There are a

few small white spots or bubbles on the vertex, which re-

semble small ocelli, but these organs seem to be actually

absent. The antenne are 13-jointed and very long, as in

the male; the scapes, however, are like those of the worker,

but extend well beyond the posterior borders of the head,

whereas joints 2-11 of the funiculi are cylindrical, sub-

equal and fully three times as long as broad, the terminal

joint being somewhat longer than these, the first shorter.

In the gaster, which is shaped as in the normal worker,

there are five distinctly visible segments, but the tip shows

clearly the small, hairy, external genital valves (stipes)

of the male. The legs are also more slender than in the

normal worker and therefore more like those of the male.

At first sight this-singular insect seems to be a gynandro-

morph, as Mayr supposed, or more specifically, an erga-

tandromorph of the blended type, with worker characters

preponderating in the trunk and those of the male pre-

ponderating in the eyes, appendages and genitalia. It is

possible, however, to regard this specimen as an ergato-

morphic male, like those which occur normally in certain

species of Ponera, Cardiocondyla, Formicoxenus, Sym-

myrmica and Technomyrmex. Unfortunately we are not

in a position to decide between these alternatives, because

we are dealing with a single fossil specimen and are not

even sure that it belongs to the species to which Mayr

O o o THE AMERICAN NATURALIST [Vor. XLVIII

assigned it. Still the case is interesting if only because

it suggests the further question as to whether the ergato-

morphic males in the genera above cited may be regarded

as originally frontal ergatandromorphs, with worker

head and thorax and male gaster, that have become the

only males of the species. If this is true, the ergato-

morphic males may have arisen by mutation from patho-

logical or teratological forms and have been preserved in

certain species in which peculiarities of habit rendered the

fecundation of the virgin females in the nest by wingless

males more advantageous than the type of mating ex-

hibited by the nuptial flight. A moment’s reflection shows

that the nuptial flight is a highly advantageous institu-

tion in common ants that form large colonies, but must be

as decidedly disadvantageous in the case of very small,

rare ants whose colonies are very sporadic and comprise

only a few individuals. This is actually the condition

seen in all the species with ergatomorphic males in the

genera Ponera, Cardiocondyla, Formicoxenus, Symmyr-

mica and Technomyrmex, and may be supposed, theré*

fore, to account for the substitution of the wingless, erga-

tomorphic for the normal winged males in these species.

LITERATURE

1908. Adlerz,G. Zwei Gynandromorphen von rig sia atratulus Schenck.

Arkiv. för Zool., V, 1908, No. 2, 6 pp., 2

1909. Donisthorpe, H. S. J. K. Formica sanguinea, a at Bewdley, with

an Account of a Slave-raid, and Description of two Gynandro-

morphs, ete. Zoologist, 1909, pp. 463-466, 2 figs.

1913. Donisthorpe, H. S. J. K. Some Notes on the Genus Myrmica. Ent.

Record, XXV, 1913, pp. 1-8, 42-51, 1 pl., 10 text fi

1868. Mayr, G. Die Ameisen des batischen ppor Bélir. Naturk.

nt sens, 1. Königsberg, 1868, pp. iv + 102, 5 pls

1907. Santschi, F. Fourmis de Tunisie Capturées en ier Ea. Suisse

Zool., XV, 2, 1907, pp. 305-334, 7 figs.

1910. Santschi, F. Contributions à la Faune Entomologique de la

Roumanie. Formicides Capturées par Mr. A. L. Montandon.

Bull. Soc. Sci. Bucharest, XTX, No. 4, 1910, pp. 648-651.

1903. Wheeler, W. M. Some New Gynandromorphous Ants, with a Review

of the Previously Recorded Cases. Bull. Amer. Mus. Nat. Hist,

XIX, 1903, pp. 653-683, 11 figs.

SHORTER ARTICLES AND DISCUSSION

ON THE RESULTS OF INBREEDING A MENDELIAN

POPULATION: A CORRECTION AND EXTENSION

OF PREVIOUS CONCLUSIONS?

IN a recent paper by the present writer on inbreeding,” the con-

clusion was reached (loc. cit., p. 608)

that no increase in the proportion of homozygotes in the population

follows inbreeding save under one or the other of two special condi-

tions, viz.:

(a) Continued self-fertilization.

(b) Some form of gametic assortative mating which increases the

natural probability of like gametes uniting to form zygotes

This conclusion is entirely correct as it stands, but also barren,

for it overlooks the very essential fact that any sort of inbreed-

ing involves in greater or less degree ‘‘gametic assortative mat-

ing.” The mathematical demonstration on page 608 of the paper

referred to is also entirely correct so far as it goes, but it stops

too soon. Up to the third generation of brother X sister mating

starting from a population of complete heterozygotes there is no

increase in the proportion of homozygotes beyond that prevailing

in a general Mendelian population. In the fourth and later gen-

erations there is, however. The blunder, kindly pointed out to

me by Professor E. M. East, which in retrospect seems altogether

too stupid even to be possible, was in the failure to recognize that

after the second generation the constitution of the family would

no longer be the same as that of the population. This is the point

which makes illegitimate the extension by induction of the results

up to the third generation to the generations beyond.

The general conclusion of the former paper quoted above,

should then be as follows: An increase in the proportion of homo-

zygotes in the population will follow inbreeding of any sort,

though at different rates for different types of inbreeding,

because any inbreeding involves homogamy (or assortative mat-

ing) in some degree.

Having made clear the location and nature of the error I desire

now to show in some detail exactly what results follow from con-

1 Papers from the Biological Laboratory of the Maine Agricultural Ex-

periment Station, No. 54.

2‘*A Contribution Towards an Analysis of the Problem of Inbreeding,’’

AMERICAN Naturatist, Vol. XLVII, pp. 577-614, 1913.

58 THE AMERICAN NATURALIST [Vou. XLVIII

tinued brother X sister mating in a Mendelian population. To

this we may now proceed.

THE DISTRIBUTION OF A MENDELIAN POPULATION IN SUCCESSIVE

GENERATIONS WITH CONTINUED BROTHER SISTER

MATING

Let us start with a population composed entirely of complete

heterozygotes. We shall consider a single character pair, A

denoting the dominant character, and a the recessive. The com-

plete heterozygote individual will then be Aa, and will produce

in equal numbers A and a gametes.

In making an analysis of the effect of inbreeding on the popu-

lation it will be necessary to deal not merely with the distribu-

tion of individuals in each generation, but also with the distribu-

tion of families of the several types. Each mating will produce

an array of families, as well as an array of individuals. The

standard family throughout this discussion is taken as including

32 individuals, of which 16 are males and 16 females. It is

further assumed that there is no sex-linkage of characters, and

that in any family there will be an equal number of brothers and

sisters of each zygotic constitution represented. One family of

16 pairs of brothers and sisters will make 16 matings and pro-

duce 16 families of 32 individuals each. This constant rate of

fertility is assumed throughout the discussion.

Every mating made is of a brother with his sister.

With so much by way of preliminary definition of the limita-

tions of this investigation, let us proceed to the actual analysis.

First Generation

Constitution of the Population—By hypothesis all individuals

are Aa.

Proportion of Homozygotes in this Generation—0O per cent.

of the whole population.

Matings to Produce the Second Generation.—Start with one

brother X sister pair of individuals from this population. The

mating will be Aa X Aa. This will produce one family, 8AA

+ 84a + 8aA + 8aa.

Second Generation

Constitution of the Population—8AA + 84a + 84a + 8aa.

Proportion of Homozygotes in this Generation.—50 per cent.

of the whole population.

No.565] SHORTER ARTICLES AND DISCUSSION 59

Matings to Produce the Third Generation—The matings of

the one family of this generation will be as follows:

cae á (9) cA (8) aA

Waa ee (10) ad m

Mad (9) 4A (is a UDe

(AJAA (8) AA a (18) F

GI as (G) 4s (13) a (A) a

(6) Aa (6) Aa (14) aa (8) aa

(7) 4a (10) Aa (15) aa... (12) aa

(Oy da... (14) Ad (16) aa “(107 dä

Third Generation Families Produced—(Note: the numbers

in parenthesis are to identify matings and their consequent

families. )

AA Aa aA aa

(1) RO. Set ee te ms Svs bs 4 oc ea dee ce a da

(2) 16 AG Gs ow Moe wa kde’

(3) 16 TR link E hbo Nha vi he cea cee

Oe eee SBA Sh ee ie

(5) 16: o o Avie Sis 1 ob SR

(6) 8 8 8 8

(7) 8 8 8 8

eee ieee as Se, os he, oe ee 16

(9) 16 16

(10) 8 8 8 8

(11) 8 8 8 8

G Boarea e E E D E T 16

NOME aa aaa a 32

RM ete cy a aa a eee 16 16

D ee ec, wi eit) ae ees 16 16

OO aA a ee ea ee 32

Constitution of the Population. —

Third Generation

128AA + 128Aa-+ 1284A + 128aa.

Proportion of Homozygotes in this Generation—50 per cent.

of the whole population.

Fourth Generation Families Produced.—The third generation

families, when mated, will produce families as follows: —

Summarized this gives the following fourth generation families

produced :

(1), 36 families like (1);

(2), 24 families like (2);

(3), 4 families like (4

(4), 24 families like (5

(5), 80 families = (6

(6), 24 families like (8

4 families like (13

(9), 36 families like (16

THE AMERICAN NATURALIST

[Vou. XLVIII

AA Aa Aa aa

iani (1)s will produce ..16 families of constitution! 32 |..... fesson fesses

Families (2) will produce 4 families of constitution) 32 |......)......)......

(3)s, or will ae a . +4 families of constitution) 16 16 aire ee

(9)3 will oe . +4 ai constitution}; 16 |...... ts gol eee

each will produce....... +4 constitution 8 8 8 8

Families (A)s will produce 16 rae tr of constitution 8 8 8 8

m hae Se Ray are ecto is Rie ees

es (6)s will rani 1 family of constitution| 32 cece Wipe ee ee oes

Oe. net will produce 2 families of constitution) 16 16 cle eee

and (11)s will produce. . +1 family of constitution|...... pi) RS a NG ED E

2 families mstitution| 16 |...... lG hona

4 families of constitution 8 8 8 8

2 families titution|...... IG heri 16

+1 family o SEAE TAS i a REE SRE E E Oe. R

2 ilies o HAtUIONL. eio verae 16 16

+1 family tut EA E RA TEN ERT 32

es (8)3 will produce 4 families of constitution 8 8 8 8

i oh cee produce. . +4 families o titut cE A gs eee 16

and (15); will produce. . . +4 families of constitution|......)...... 16 16

each will produce....... +4 families of constitution|......|......)....6- 32

Family (16)3 will produce 16 families E cy Vinal eee 32

Fourth Generation

Constitution of the Population.—

25604.A + 15364a + 1536aA + 2560aa.

Proportion of Homozygotes in this Generation.—62.5 per cent.

of the whole population.

Fifth Generation Families Produced.—The third generation

families, when mated, will produce families as follows:

(1), will produce 36 X 16 — 576 families like (1),

(2), will produce 24 X 4= 96 families like (1);

+ 96 ilies like (2),

+ 96 families like (5),

+ 96 families like (6),

(3), will produce 4 X 16— 64 families like (6),

(4), will produce 24 X 4= 96 families like (1),

+ 96 families like (2),

+ 96 families like (5),

+ 96 families like (6),

No. 565] SHORTER ARTICLES AND DISCUSSION 61

(5), will produce 80 families like (1),

+80 X 2= 160 families like (2),

+ 80 families like (4);

+ 160 families like (5);

+80 X 4= 320 families like (6),

+ 160 families like (8);

+ 80 families like (13),

+ 160 families like (14),

families like (16),

families like (6);

families like (8);

+ 96 families like (14),

+ 96 families like (16),

(7), will produce 4 X 16= 64 families like (6);

(8), will produce 24 X 4= 96 families like (6);

+ 96 f i

for)

+ 80

(6), will produce 24 X 4=

+ 96

+ 96 families like (14);

+ 96 families like (16)

7 (9), will produce 36 X 16 = 576 families like (16),

Summarized this gives the following fifth generation families

produced:

(1); 848 families like (1);

(2), 352 families like (2);

(3), 80 families like (4);

(4), 352 families like (5),

(5), 832 families like (6),

(6); 352 families like (8),

(7); 80 families like (13);

(8), 352 families like (14);

(9), 848 families like (16),

4096 (== 16 X 256).

Fifth Generation

Constitution of the Population.—

44,7364A + 20,480Aa + 20,480aA + 44,736aa.

Proportion of Homozygotes in This Generation—68.75 per

cent. of the whole population.

Sixth generation families produced:

17,216 families like (1);

4,480 families like (2);

11,520 families like (6);

4,480 families like (8);

832 families like (13);

4,480 families like (14);

62 THE AMERICAN NATURALIST [Vou. XLVIII

17,216 families like (16),

65,536 (—=16 X 4,096).

Sixth Generation

Constitution of the Population.—

786,4324 A + 262,144.40 + 26214404 + 786,432a0.

Proportion of Homozygotes in This Generation.—75 per cent.

of the whole population.

From this point on it will not be necessary to carry out the

work in detail. The final results are given in Table I for four

more generations.

TABLE I

SHOWING THE CONSTITUTION OF THE POPULATION AFTER 7 TO 10

GENERATIONS OF BROTHER X SISTER MATING

ogo pues oy

ozy-

— AA Aa aA aa peek d

; Whole Pop-

ulation

$ 13,369,344 3,407,872 3,407,872 13,369,344 79. vod

8 224,395,264 44,040,192 44,040,192 224,395,264 83.5

9 i La 541,952 570,425,344 570,425,344 | 3,724,541,952| 86. a

10 337,501,696 | 7,381,975,040 | 7,381,975,040 | 61,337,501,696 89.26

It is evident that the proportion of homozygotes is approaching

100 per cent. in the same manner as in the case of self-fertiliza-

tion, worked out by East, Jennings and others, but at a slower

rate.

In a later paper I hope to take up the problem of the general

formule for finding the constitution of a Mendelian population

after n generations of inbreeding of the different types, and at

the same time discuss the relation of these results to the coeffi-

cients of inbreeding described in my former paper. It should be

specifically mentioned that, in the light of the data here set forth,

those criticisms of the conclusions of East and Hayes made in my

former paper? which were based on the erroneous assumption of

a fundamental difference between self-fertilization and all other

forms of inbreeding in respect to homozygosis, have no validity

whatever. It scarcely needs to be said that the blunder on the

theoretical side here corrected in no wise affects the usefulness of

inbreeding coefficients. RAYMOND

3 Cf. Pearl, loc. cit., p. 606, 609 and 610.

ISOLATION AND SELECTION ALLIED IN PRINCIPLE

THERE are those who fully recognize the influence of natural

selection in transforming the hereditary characters of a species,

but are unable to see how isolation should have any effect of that

kind. They say that you may divide a species into two branches

between which all possibility of crossing is completely prevented,

but if the environment surrounding each branch is the same, the

natural selection to which each is subjected will be the same, and

no divergence of character will take place. They forget that the

separate branches, if prevented from crossing for many genera-

tions, are sure to develop different types of variation, and in due

time different methods of using the same environment, and are

therefore liable to subject themselves to different forms of selec-

tion. Again they forget that when the power of dispersal is

highly developed in a species it may be exposed to diverse en-

vironments and therefore to diversity of selecting influences, and

still remain one harmonious species, because free erossing is

maintained between all parts of the species. As long as there is

no isolation of different branches, that is, while free crossing con-

tinues, there is no permanent divergence resulting in diverse

races or species, even though the one species is exposed to differ-

ent forms of selection in different parts of its habitat.

Diversity of evolution, producing many divergent forms of

animals, could never have arisen without continuous isolation be-

tween the different forms.

Again there are those who maintain that selection unaided by

isolation can not produce transformation. It is true that diver-

gent groups can not be produced and intensified without isola-

tion; but a given race may be transformed by selection without

being divided into two groups by isolation.

Heredity with variation is the active cause of transformation;

isolation and selection are the conditions that shape the forms of

heredity and variation.

It is a law of heredity, that, if those of a given stock that are

most alike in hereditary characters mate with each other, there

will be a tendency in their offspring to a stronger emphasis of

that character,

64 THE AMERICAN NATURALIST [Vou. XLVIII

Another law of heredity is that as long as free crossing is main-

tained between the different forms of a species these forms can

not become widely divergent. The elephant and the mouse could

never have been developed from one original stock while free

crossing continued.

Now there are many ways by which the free crossing of one

variation with others of the same species may be prevented, but

they all.come under two groups.

Under selection are classed all the influences enabling certain

variations to reproduce more successfully than other variations,

and so preventing free crossing between the successful an

unsucessful. Under isolation are classed all the influences that

prevent living, and sexually reproducing creatures, from freely

crossing.

Under normal conditions there is no crossing between the ass

and the horse, though there is reason to believe that the early

ancestors of each were of one stock freely interbreeding and pro-

ducing fertile offspring. If isolation had not existed for ages be-

tween them, they could not have become the separate creatures

that they now are. Heredity can combine only compatible char-

acters. In some cases, incompatible characters arise between

creatures of the same race preventing any crossing between them,

as when a dextrally twisted mollusk produces a sinistrally twisted

one; but, in most cases, such incompatibility arises only after

imla kon, through geographical separation, for many generations.

In view of these facts, is it not plain, that, in the case of a

variable and plastic organism, races more or less divergent will

be produced, if for many generations the organism is divided

into branches that are prevented from crossing? Is not such a

result just as sure as the gradual transformation of the race

under a slow change of climate, when the successful variations

are prevented from crossing with the unsuccessful variations?

JoHN T. GULICK

HoxNoLULU, T. H.

VOL. XLVIII, NO. 566 -£ FEBRUARY, 1914

THE

AMERICAN

NATURALISI

A MONTHLY JOURNAL

Devoted to the Advancement of the Biological Sciences with

Special Reference to the Factors of Evolution

CONTENTS

Page

Some New Varieties of Rats and Guinea-pigs and their Relations to Prob-

lems of Color Inheritance. Professor W. E. CASTLE

Il. “Dominant” and “Recessive” Spotting in Mice. C. C. LITTLE - a

IHI. On Differential Mortality with respect to Seed a eaer in m

sum sativum. Dr. J. ARTHUR HAER

IV. The scones of a Recurring Somatic —— in ‘Varierated Ears of

Professor R. A. EMERSON

y. S ern 6f Edaphosaurus ee Cope. Professor E. C. CasE- - Ht

VI. Shorter Articles and Discussio

Humidity—a Neglected Pastor in Bairen Work. Dr. Frank E. LUTZ 122

THE SCIENCE PRESS

LANCASTER, PA. GARRISON, N. Y.

NEW YORK: SUB-STATION 84

Two Important Biological Journals

The Journal of Genetics

Edited by

W. BATESON and R. C. PUN shit

A periodical for the Ss me of records of original research in contd Dan and allied topics.

The JOURNAL also contains, from time to time, ar ticles summarizin as ex himig = ate of knowledge in the

colored plates are freely employed. The JouRNAL i s issued in par rts as ma iial à ccumulates, and a volu

(appearing, “or e STS annually) consists of four such parts. pabesi iden price, $7.50 a ventas

single copies, $2.

Biometrika

Edited by KARL PEARSON

A journal se ee Statistical study of biological problems, appearing about four times a year. A volume

containing abou pages, with plates and tables, is issued annually. Volume X begins i in 1914. The

nine preceding skend es form a veritable encyclopaedia of biological statistics in the fields of ee

craniometry, routine, heredity, ete. Subscription price, $7.50 a volume : ; Single copies, $2.

Prices on back numbers of these journals vary, and postage from London is charged on back volumes

and single copies

Published in America for the Cambridge University Press of England by

The University of Chicago Press

Chicago, Illinois

University Control

By J. MoKeen CATTELL, Professor of Psychology in Columbia University

Together with a series of Two Hundred and Ninety-nine Unsigned Letters by Leading Men of Sciencs

holding Academic Positions and Articles by Joszra Jastrow, Georam T. Lapp, Jomn J. Srzvenson, J. E

Cruicuron, J. McKeen Carretn, Guorncn M. Srearron, Stzwart Paton, Joun Jay Cuarman, James P.

Munron and Jacos Goutp

Sic Beir ooh of questions dnon general university administration are dealt with in an original and

way.—

These quotations and examples are taken from Professor Cattell’s informed and thorough discussion of ae

pe i of E control, a subject upon which h he has had much to say of late, finding occasion for ca

ing paare oe which a man may enter without forfeiting his self-respect.— The

Sentences and paragraphs that betoken the expert, highly-trained Mioa, the suggestions that come to 1%

fresh a d tell us that a new day is about to dawn in educati onal writing.—The Boston Evening Transcript.

CIENCE AND EDUCATION

A series of volumes for the promotion of scien tifie research and educational progre ont

VOLUME I. The Foundations of Science. By H. Porcar. Containing the authorized English tran

7 Gzogas Beocom Hatsrap of “Science and Hypothesis,” “The Value of Science,” and “Scien?

Method.” Price, $3.00 net.

VOLUME II. Medical Research and Education. By Rıcmano M. Paancs, Wim H. Watos, W- B

H m P. Matt, Lewmuys F. Barker, Cmarims S. Minor, W. B. Caxnon, W. T.

Councitman, THEOBALD Surm, G. N. SrawarT, C. M. Jackson, E. P. Lron, James B. Herrick, J08"

M. BE peaa C. 7 aR W. Opntizs, B. J. Meurzwr, Jamas Ewin, W. W. Kuen, Haney H. Dona”

R, and Hewry P. Bowprrcu. Price, $3.00 net

VOLUME Iii. Givat s G trol. Now Ready. Pages z-4-484. Price, $8.00 net.

GARRISON, N. Y. THE SCIENCE PRESS LANCASTER, PA.

o SUB-STATION 84, NEW YORK CITY

THE

AMERICAN NATURALIST

Vor. XLVIII February, 1914 No. 566

SOME NEW VARIETIES OF RATS AND GUINEA-

PIGS AND THEIR RELATION TO PROBLEMS

OF COLOR INHERITANCE

PROFESSOR W. E. CASTLE

HARVARD UNIVERSITY

THE marvelous color variation of the domesticated ani-

mals has been recognized as a capital illustration of evo-

lution from the time of Darwin to the present, and much

study has been devoted to the question of how it has taken

place. Nevertheless we have very little positive informa-

tion as to how existing color varieties originated, and

theories differ concerning the matter.

It becomes therefore important to record carefully any

contemporary events which may throw light on the sub-

ject. This is my excuse for calling the attention of scien-

tists to the recent appearance in England of two new and

striking color variations of the common or Norway rat.

I say ‘‘appearance’’ advisedly for it is impossible to say

how long these variations may have been in existence

within the race, cropping out perhaps from time to time

sporadically. Certain it is, however, that they have only

quite recently come to the attention of ‘‘fanciers,’’? who

have taken them up with great enthusiasm.

The new varieties are known to the fancy as (1) pink-

eyed yellow, fawn, or cream; and (2) black-eyed yellow,

fawn, or cream. From the evidence at hand it is clear

that each of the two variations has originated in the wild

66 THE AMERICAN NATURALIST [Vou. XLVIII

race and as a mutation or unit-character variation, retro-

gressive in nature (i. e., due to loss of some normal con-

stituent from the germplasm). Each is a simple Mende-

lian recessive character in crosses with wild race, and

with certain at least of the tame varieties. The two varia-

tions have not as yet been combined by intercrossing, but

this will be attempted soon and, I doubt not, with entire

success.

My first information about the new variations was

obtained from Fur and Feather, the official organ of the

English fanciers, in which appeared advertisements of

‘‘the new variety” of black-eyed yellow rat. Now as

long ago as 1903 Bateson had commented on the singular

absence of a ‘‘yellow’’ variety among rats, noteworthy

because nearly all mammals kept in captivity have such

varieties; and I have since been bold enough to publish

some speculations as to why this variation had not made

its appearance. Consequently I was much excited to

learn that it actually had appeared. Miss M. Douglas,

one of the editors of Fur and Feather, and secretary of

the National Mouse and Rat Club (of England) very

kindly answered my inquiries about the new varieties and

put me in communication with the ‘‘originators,’’ who

have given so clear and full accounts of their procedure

in establishing the new varieties that even the genetic be-

havior of the variations is fairly certain, though I pur-

pose to confirm this fully with experiments which are

already in progress.

The pink-eyed variation made its’ appearance first, so

far as known, about 1910 or 1911, but it had probably

been in existence for some time and become rather widely

diffused throughout the central part of England, for at

about the same time pink-eyed wild rats were caught at

or near Preston and at Chesterfield, cities some 65 miles

apart. I am informed that Mr. T. Robinson at Preston

and Mr. W. E. Marriott at Chesterfield independently

established the ‘‘pink-eyed fawn’’ variety, or what would

better be called the pink-eyed agouti variety, since appar-

No.566] NEW VARIETIES OF RATS AND GUINEA PIGS 67

ently it differs from the wild gray (or agouti) variety by

the pink-eyed variation alone. It is not a true yellow

variety at all genetically, though (like the pink-eyed gray

mouse) it resembles one superficially because of the yel-

low ticking of the agouti fur.

It is also quite distinct genetically from the albino

variation seen in white rats, yet its ‘‘dirty white’’ color is

enough like the appearance of the albino to permit mis-

taking one for the other. Possibly this is why the pink-

eyed variation may have been for some time overlooked.

Mr. Robinson has not answered my inquiries, but Mr.

Mariott writes in detail about his observations and ex-

periments.

Under date of October 11, 1913, he says:

The first rat with any semblance of fawn in it that I had was caught

in a trap on a provision merchant’s premises in Chesterfield. You could

scarcely call it a fawn, but more of a cream or dirty white. I have also

had four others similar to this one, 2 caught at the same place and 2

caught at a malt-house in close proximity to the other premises, [in all]

3 bucks and 2 does, but the only one that I was able to get to breed was

the first brought to me, which was a buck. When first caught it was

very wild, in fact it appeared to me to be more wild than an ordinary

wild rat. It was a source of trouble getting it to mate, killing no less

than 20 does before mating. I eventually got it mated to 2 does, one a

pure white for at least 10 generations, and one black-and-white hooded-

and-striped, or Japanese rat. The result of the pure white cross was 2

young, a buck and a doe, which were agoutis with no white at all. The

result from the Japanese cross was 7 young, 5 does and 2 bueks, which

were the color of Irish agoutis being agouti color with a white stripe

running underneath. These results naturally caused me great disap-

pointment as I was expecting a fawn colored young one. When the

young were old enough I mated father and daughter, result mil; mother

and son, result nil; brother and sister. The brother and sister mating

from the pure white cross produced 2 fawn colored rats, a buck and a

doe, and 5 agoutis.? The brother and sister mating from the Japanese

cross produced 2 fawn-and-white Japanese, 1 cream-and-white Japanese,

1Italies mine. Note the reversion to full wild color. This shows the

pink-eyed variation to be entirely different in nature from the ordinary

albino variation

2 Note the return of ‘‘fawn’’ (pink-eyed agouti) as a recessive character

in approximately 1 in 4 young.

68 THE AMERICAN NATURALIST [Vou. XLVIII

1 black-and-white Japanese, and 4 agoutis. The fawns and fawn-and-

whites resulting from these crosses were much’ deeper in color than the

wild grandsire. Mated one with another they gave a proportion of

about 2 fawn colored or fawn-and-white in 7 young.* I may say in

conclusion that the original wild rat was in shape of body, skull, ete.,

as the ordinary brown or agouti rat that we have running wild in

this district.

Mr. Marriott sold a ‘‘fawn-and-white’’ (pink-eyed

hooded agouti) buck to Mr. E. F. Tilling, of Hessenford,

who also ‘‘originated’’ the second variation, the ‘‘black-

eyed yellow,’’ or true yellow variation. His results from

the pink-eyed variation confirm those of Mr. Marriott.

Mr. Tilling writes under date of October 18, 1913:

I see by Fur and Feather this week that you are interested in the

yellow and cream varieties of rats. I am also much interested in these

and have produced the latter variety within the last few months. We

have 2 kinds over here, the yellow-and-white hooded with pink eyes and

the self yellow (and cream) with black eyes. Both are quite distinct.

The first mentioned was produced some 2 or 3 years ago. Mr. Marriott,

of Chesterfield, bred the first I heard of from a wild caught fawn. He

bred a couple of yellow and white hooded bucks of which Miss Douglas

bought one and I the other. I mated mine to about 15 does of various

colors and definite strains. He was a splendid breeder and got some very

fine youngsters, but not one of his own color from the first cross® I

subsequently mated him to some of his daughters and they produced a

ood proportion of yellow-and-white young.® ese are now fairly

plentiful over here and are in the hands of several fanciers.

Of the other kinds, black-eyed fawns and creams, the first one ex-

hibited and from which all mine are descended, was a very fine wild

caught, deep colored, fawn specimen. I got her partly tame and ex-

hibited her at the National Mouse and Rat Club’s annual show at Bristol

on November 27 and 28, 1912, where she won first in the self class and

3‘‘ Fawn- and white Japanese’’ means (to me) pink-eyed agouti with the

‘‘ Japanese’? color pattern (hooded). The formation of this class of young

shows the hooded pattern (‘‘Japanese’’) to be independent in transmission

of the pink-eyed variation. ‘‘Cream-and-white Japanese,’’ I interpret as

pink-eyed black ~ agouti) hooded. ‘‘ Black-and-white Japanese’’ is the

familiar black hooded. We should expect this mating to produce also self

pink-eyed agouti and self pink-eyed black which are not mentioned.

+The Mendelian expectation is 2 in 8

5 Italics mine. Note again the recessive nature of the variation.

® Not real yellow-and-white, as already explained, but pink-eyed agouti-

and-white or black-and-white.

No.566] NEW VARIETIES OF RATS AND GUINEA PIGS 69

was well commented upon in the fanciers’ papers. From this doe I have

built up my strain of black-eyed creams. I mated her to a self black

buck and she bred 8 youngsters all wild colored.* This is the only

litter I had from her, as shortly afterward, during my illness, my man

while transferring her from one cage to another let her get away and

was unable to recapture her. However, I have bred from her young-

sters, mating brother and sister, and the litters have invariably con-

tained at least 1 fawn or creamê each time. I have now just bred for

the first time from the 3 first does so produced, again mating them to

their brother and the result is litters of 7, 5 and 7, respectively, all self

creams.?

From the statements of Messrs. Marriott and Tilling,

it is evident that the two variations, which they, respect-

ively, have introduced into the rat fancy, are both reces-

sive in heredity, as are also the three previously known

Mendelizing color variations of rats, viz., (1) the albino

variation (with uncolored coat and eyes); (2) the black

variation (lacking the agouti ticking of the fur); and (3)

the piebald ‘‘hooded’’ pattern of white and colored fur.

Each of these is known to be an independent Mendelizing

unit-character. If the new variations are as supposed in-

dependent of each other and of those previously known,

they will make possible the immediate four-fold increase

in number of the previously known color varieties of rats.

If for the present we adopt a simplified terminology (as I

have elsewhere suggested) for the different color varia-

tions, employing small letters for such as are recessive

in heredity, we may use the following set of symbols:

White (albino) = w,

ack =: D;

Hooded a= My

Pink-eyed =p,

Yellow a a

7 This shows that the original yellow animal was potentially an agouti.

A pair of yellows which Mr. Tilling has sent me have light bellies and I

presume are also potentially agoutis. t

8‘‘Cream’? here probably means yellow not transmitting agouti. It

probably lacks the lighter belly as do yellow rabbits which do not transmit

agouti.

° This shows that extracted yellows breed true to yellow. Henee the

variation is recessive, as in rabbits and guinea-pigs, not dominant as in mice.

70 THE AMERICAN NATURALIST [Vow. XLVIII

By various combinations of these variations, if each is

independent of all the others, 32 varieties become possi-

ble. Half of these varieties will be albinos, white and so

visibly indistinguishable. The other 16, we have reason

to suppose, will look different from each other. Pre-

viously we had but four of these, the first four in the fol-

lowing list of the theoretically possible 16.

1; Normal or wla s.: c. agouti.

Bi Oise wees emcee ee black,

Be Roce opie ee eas hooded,

A DN 2 de P E E black hooded,

OP Weve O a pink-eyed,

O PO cerere Eas ae pink-eyed black,

Dice eed cee neues e pink-eyed hooded,

B. PON Sis. Us esns avec. pink-eyed black hooded,

B, Goo ees OF orate yellow,

IK PO co bes es ee ae yellow black (i. e., non agouti yellow),

Ble GR gies ae yellow hooded,

Megh Serer rere rer yellow pink-eyed,

ESPON o eho a as yellow black hooded,

De FOO: see e we yellow pink-eyed black,

EOS MON esas vee eke yellow pink-eyed hooded,

SOs POOR is cs tai reds yellow pink-eyed black hooded.

Varieties 1-4 have been known for some time; they

have constituted the fancier’s entire repertoire up to the

present time. Varieties 5 and 9 have apparently arisen

as wild sports obtained by Marriott and Tilling, respect-

ively. By crosses these gentlemen have apparently ob-

tained varieties 6, 7, 8, and probably 10. Varieties 11-16

are as yet unknown, but will doubtless soon be produced.

Corresponding with each of the 16 colored varieties, an

uncolored one should be possible of production, which

would transmit in crosses with any colored variety the

characteristics indicated by its formula. Albinos cor-

responding to colored varieties 1—4 are positively known

to occur; their symbols would be w, wh, wh and wbh, re-

spectively. Symbols for the remaining 12 expected varie-

ties may be formed in like fashion, by prefixing w to the

combinations already given.

All the five unit-character variations, which in different

combinations are responsible for the color varieties of

No. 566] NEW VARIETIES OF RATS AND GUINEA PIGS TI

rats, have their parallels in other mammals. Albinism

and white-spotting (which in rats takes the form of the

hooded pattern) are among the commonest. They occur

in practically all mammals from mice to men. Albinism

appears to consist in such a modification of metabolism

that the process of pigment-formation can take place only

feebly or not at all. That particular process which seems

chiefly affected is the production of yellow pigment.

Albinos, so far as I know, never produce genuine yellow

pigment, though they may produce considerable quanti-

ties of black or brown pigment, as in the case of the

Himalayan rabbit. An undescribed variety of guinea-

pig, which I obtained about two years ago in Peru, may

bear as much black pigment in its coat as wild cavies do,

yet it forms no yellow pigment at all. Further this varia-

tion behaves as the allelomorph of ordinary albinism, in-

dicating that it is probably of the same genetic character.

For this reason we may provisionally consider the albin-

ism of mammals as due to a loss of the ability to form

yellow pigment. This usually, if not always, involves a

lessened capacity to form other pigments also, so that it

seems probable that the same chemical process, which

produces yellow pigment as an end-product, is ordinarily

involved also in producing the higher oxidation stages

seen in brown and black pigment. In albinos this process

would seem to be omitted, or to be accomplished by some

step which does not involve the production of yellow

pigment.

The yellow variation is extremely common in mammals.

Yellow varieties, which at opposite extremes of intensity

of pigmentation are known as cream and red, occur

among horses, cattle, hogs, cats, dogs, rabbits, guinea-

pigs, mice and human beings. In this variation pigment

oxidation stops at the yellow stage, usually throughout

the coat but not in the eye. Described in negative terms

a yellow variety is one in which black and brown are sup-

pressed or restricted. Black and brown, though usually

restricted to the eye in yellow varieties, may occur also in

72 THE AMERICAN NATURALIST [Vou. XLVIII

small quantities in the fur. Examples are found among

horses (bay and dun varieties), cattle (the Jersey breed),

dogs (the common dirty yellow variety), rabbits (the

‘*tortoise-shell’’? variety), mice and guinea-pigs, and

probably red-haired human beings also.

Black varieties of mammals arise in two genetically

distinct ways. One is a quantitative increase or exten-

sion of black, the reverse of what happens in yellow varie-

ties, so that black encroaches on regions normally yellow

or may even obliterate them altogether. Examples are

found in black squirrels, in which the agouti yellow tick-

ing of the fur is almost, but not quite, obliterated by black

pigment. But the ‘‘black’’ variation of rats, mice,

guinea-pigs and ordinary rabbits results from a total

loss, not a covering up, of the yellow ticking of the fur

seen in agouti varieties. Genetically it is quite distinct

from the other kind of black. ye is a recessive variation

and so breeds true.

The pink-eyed variation is the rarest of all the five

enumerated as occuring in rats. It has been known here-

tofore only in mice, though I have recently obtained it

also in guinea-pigs from Peru, where it seems to be well

established.

In this variation the capacity to form yellow pigment

is unimpaired, but only traces of black or brown pigment

are produced. Consequently varieties which possess the

other genetic factors of normal yellow animals have fully

pigmented (yellow) fur, but with very faintly pigmented

(pink) eyes, when they possess this factor. If, however,

they possess the other genetic factors of black, brown, or

agouti varieties, along with this pink-eyed variation, then

both the fur and the eyes are very faintly pigmented.

From this results the seeming paradox that pink-eyed

blacks are less heavily pigmented than pink-eyed yellows,

so that in rats the fanciers have called the former

‘‘creams,’’ the latter ‘‘fawns.’’

When pink-eyed animals are crossed with albinos, off-

spring fully colored (eyes and all) result, as was first

No. 566] NEW VARIETIES OF RATS AND GUINEA PIGS 73

shown by Darbishire some ten years ago. This indicates

that the two variations are not only genetically distinct,

but are physiologically complementary. The albino has

defective metabolism for producing yellow (and in conse-

quence brown and black also); the pink-eyed animal has

the full mechanism for forming yellow, but its brown and

black producing mechanism is defective. Together they

possess the full mechanism of normal color production.

Hence the reversion on crossing.

White spotting is clearly due to neither of the above

modifications, but to a different change in the metabolism

so that no pigment at all is produced. For an albino rab-

bit or guinea-pig may, as already observed, bear consider-

able black or brown pigment, but a white spot either on

an albino, on a pink-eyed animal, or on a fully colored

animal is entirely devoid of pigment. The paradox of a

white spot on an albino is obtainable by crossing a white-

spotted colored race with an albino race, which develops

some pigment in the fur, as for example the Himalayan

race of rabbits and guinea-pigs. In this way English-

marked Himalayan rabbits and spotted albino guinea-

pigs have been produced in my laboratory.

Postscript: While this paper was in press, Mr. Tilling,

in reply to a further inquiry, wrote that his original black-

eyed yellow rat was caught on a ship at Liverpool. The

fact that the pink-eyed variety was found in the same gen-

eral region leads him to believe that both variations were

introduced on ships from some foreign country. It would

be of much interest to know from what country or coun-

tries. Any information on this point obtainable from

rat-catchers or others would be welcome.

“DOMINANT” AND ‘“‘RECESSIVE”’? SPOTTING IN

MICE ,

C. C. LITTLE,

BUSSEY INSTITUTION, HARVARD UNIVERSITY

INTRODUCTORY

Tue inheritance of spotting has long proved of interest

to animal geneticists. The nature of spotting is such

as to afford an excellent chance to observe quantitative

fluctuation and variations of very minute size. Further-

more, the fact that spotted varieties are found in all the

rapidly breeding smaller domesticated mammals has led

to a widespread investigation of its phenomena of in-

heritance.

One of the most clean-cut and constant types of spot-

ting which has been studied is that of the ‘‘hooded’’ pat-

tern in rats. This character was studied independently

by Doncaster (1905) and by Castle and McCurdy (1907).

All these observers agree that this form of spotting is

due to a recessive Mendelizing unit which gives a 1:3

ratio in crosses with self-colored races.

In mice there has been no such well-localized pattern

recorded and a series of spotted forms has been described

which vary from black-eyed whites on one end of the

series to heavily colored. animals having only a few

white hairs on the forehead or on the belly at the other

extreme.

Cuénot, who did considerable work on the inheritance

of spotting in mice, came to the conclusion that spotting

is due to a group of recessive spotting factors which he

describes as pl, p2, p3, p4, ete. His figures, however,

show a single unit character difference as 3:1 and 1:1

ratios prove.

Up to 1908 all the spotting in mice was classed as re-

cessive to solid-colored coat. At that time, however,

No. 566} SPOTTING IN MICE 75

Miss Durham described the appearance of dominant spot-

ting in addition to the recessive form which she also had

experimented with. Such a dominant form of spotting

is supposed, by Bateson, to be due to the addition of some °

factor for restriction of pigment formation in certain

areas. This produces a dominant form of spotting as

contrasted with the recessive type, which, he holds, is due

merely to the loss of the ‘‘self’’ factor.

Hagedoorn (1912) gives data to show that the domi-

nant form of spotting occurs in mice and in addition con-

siders it as produced by a factor analogous to that which

produces the dominant ‘‘English’’ spotting in rabbits.

The object of this paper is to present certain evidence

concerning the nature of dominant and recessive spot-

ting in mice; to discuss in its light the results of the above-

mentioned investigations; and to criticize one additional

point in Hagedoorn’s work with mice.

EXPERIMENTAL

Materials —Among several wild mice caught during

the spring of 1911 was one individual with a white spot

or ‘‘blaze’’ on the forehead between the eyes. This spot

or ‘‘blaze’’ was about one quarter of an inch in length

and one eighth of an inch in width. This mouse, an adult

male, was transferred to a breeding cage and a series of

experiments was started to determine whether the

‘blaze’? character was inherited and, if so, in what way.

As at that time no adult wild females were available

from unrelated stock the wild ‘‘blaze’’ male (S1) was

crossed with a female from a dilute brown race. In

many ways this dilute brown race was the best possible

material for such a cross. It was very closely inbred,

being descended from a single pair of animals, progeny of

which had been free from out-crossing for more than a

year. Further, it had never given, nor has it ever given

in hundreds of young, an animal with the slightest trace

of a spot, even on the tail, where white bands are fre-

quently seen in wild mice. Besides this the race was vig-

orous and active and yet easy to handle.

76 THE AMERICAN NATURALIST [Vou. XLVIII

RESULTS

As a result of mating S1 ‘‘blaze’’ with a female of this

dilute brown race, two litters, totalling eight young, were

‘produced. All these young were self-colored without a

trace of white, and, as expected, all resembled the male in

coat color.

The F, generation selfs were then crossed in two ways,

(1) inter se and (2) with animals of the dilute brown self

race to which their mother belonged. It is hoped that a

detailed account of all the matings made may be pub-

lished later, but for the present purposes certain of the

crosses under the first heading will suffice.

When F, was crossed inter se, two sorts of young were

produced, namely, those with white and those without.

While all of the latter type may be classed as self, the

former were of two general sorts: (1) those with a

‘‘blaze’”’ as large or larger than that of S1, these we may

call ‘‘blaze’’? animals; and (2) those with only a few

white hairs on the forehead, which we may call few white-

haired (f.w.h.) animals.

The exact numbers in this cross were

Offspring

Parents

Self F.W.H. Blaze

SO NBER ee; 11 3 3

BID OC Be a ea 10 13 6

Bis x Be ee 3 1 2

24 ee | 11

When the F, few white-haired animals were bred to-

gether they produced three types of young: few white-

haired, blaze and self, as follows.

| Offspring

Parents

| Self | F.W.H. Blaze

zoa x $098 | 11 6 ae

$06 re i. 2. ose, s 5 5 5 ioo

i | 16 | 11 beto

One further fact is also of interest. Various descendants

of F, ‘‘blaze’’ animals, which should breed as recessives,

No. 566] SPOTTING IN MICE 77

have given the following results. The generation num-

bers may be disregarded as they refer to another method

of classification. It is to be remembered that the parents

in the tabulation given below, are all ‘‘blaze’’ in

character.

Young Produced

Generation |

te | ae ae Self | Total

WES ols: 33 6 4 1 44

Bee ics 157 60 27 3 247

See 70 53 5 0 128

WB as is 9 6 0 0 15

i Se ee 125 36 ti d S

If the ‘‘blaze’’ is a true Mendelian recessive we should

expect all 434 offspring to have some white on them.

The figures show that 430 of the 434 are of this type;

that is to say, approximately 1 per cent. are self.

It is possible to account for the occasional production

of selfs even if the ‘‘blaze’’ character is a true recessive,

if we supposed that there are supplementary factors

which may influence color development; and it is quite

conceivable that such is the case.

The chief point of interest in the crosses given above

is that while spotting behaves in F, as a recessive, certain

of the F, spotted individuals fulfil the requirements of

dominant spotting by producing self offspring.

The spotting came from a single individual and can

scarcely be considered to be of two distinct types.

We may now consider the bearing of these results on

the work of Miss Durham and Hagedoorn.

Miss Durnam’s RESULTS

Miss Durham (1908) gives a detailed account of a re-

cessive type of spotting in mice. The numbers she ob-

tained are extensive, and the case seems well established,

coming as it does in corroboration of the work of Cuénot,

Darbishire and others. In the same papers she records

the occurrence of a dominant spotted type of mice. Bate-

son (1909), commenting on the case, compares it with the

78 THE AMERICAN-NATURALIST [Vou. XLVIII

dominant ‘‘English’’ spotting in rabbits but also agrees

that, in the case of mice, there is no criterion to enable one

to distinguish somatically between the dominant and re-

cessive forms. This, of course, is not the case in rabbits

where the ‘‘English’’ pattern differs visibly from the

“Dutch” spotting, which Hurst (1905) found to be re-

cessive to self. Bateson also considers that the case of

dominant spotting in mice, reported by Miss Durham, is

the result of a different spotting factor from that pro-

ducing recessive spotting.

In terms of the presence and absence hypothesis this

means that the dominant form possesses a factor for re-

striction of pigmentation which self forms lack. This

fact becomes of interest when Miss Durham’s experi-

mental results are closely examined.

In the race which gave rise to the dominant spotting

the following conditions are seen.

A sooty yellow spotted mouse of unknown origin was

crossed with a black-eyed white (spotted) animal (of

Atlee’s strain). Among other progeny was obtained a

black-eyed white mouse with ‘‘agouti ears.” This

mouse, No. 21 (spotted), was crossed with an albino (car-

rying chocolate), No. 35, and gave among its progeny No.

69, a black self mouse. This black animal, No. 69 was

crossed with an albino (carrying chocolate), No. 34, and .

from these two individuals came the dominant spotted

race.

Now: inasmuch as No. 34 and No. 35, the albinos, were

not supposed to carry spotting, the dominant spotting

must be considered:as probably coming from No. 69, a

black self animal. We know that this animal must carry

spotting as a recessive character since its parent, No. 21,

was spotted.

If, therefore, this animal was the progenitor of the

dominant spotted race, and if he carried a recessive spot-

ting, as it seems certain he did, we must suppose that one

of three things has happened to the recessive spotting

which he carried.

No. 566] SPOTTING IN MICE 79

1. It may have been completely lost, failing to manifest

itself in his germ cells.

2. It may have continued to exist and to be inherited

together with the dominant type of spotting.

3. It may have been changed to a so-called ‘‘dominant’’

type of spotting simply by the nature of modifying sup-

plementary factors which it encountered during ontogeny.

The first two cases necessitate the origin of the ‘‘domi-

nant’’ spotting by a mutation in no way connected with

the previous recessive spotting. In the first case, more-

over, we should have to suppose the disappearance of the

recessive spotting character in a manner entirely con-

trary to any principle of Mendelian heredity. In the sec-

ond case the occurrence of the two types of spotting side

by side in the same litters of young would so complicate

the experiments that analysis would be difficult if not im-

possible, on Miss Durham’s results.

There is good reason to believe that the third possible

explanation is the correct one. It accounts for the for-

merly ‘‘recessive’’ type of spotting. It presupposes no

fundamentally different appearance of the two types

of spotting. Moreover, it is very probable that the al-

bino race brings in the modifying factors necessary to

give the apparent change in the type of spotting. The

addition of a factor as presupposed by the presence and

absence hypothesis is not proved by the results obtained

nor is it necessary to account for them.

That the presence and absence hypothesis does not

apply to all cases of spotting is seen in the case of the

‘“‘blaze’’ mice in my experiments. Here, if F, animals

had been given me as a starting point for experimenta-

tion, I should conclude the spotting to be recessive, while

if F, spotted animals were given as a starting point the

conclusion would be inevitable, that spotting should be

considered dominant. Yet it is one and the same spot-

ting in both cases. It is certain that ‘‘self’’ and ‘‘blaze”’

are alternative conditions, but it is equally certain that

they differ from each other rather as two degrees of a

80 THE AMERICAN NATURALIST [Vou. XLVII

single process, one greater, the other less, than as the

presence and absence of one or more unit characters.

Hacepoorn’s Work

-= Hagedoorn’s work shows the danger of the modern

tendency to produce factors upon the slightest provoca-

tion. While adding, in experimental work, only a single

litter of young bearing on the problem, he gives a symbol

for a factor for dominant spotting in mice, and further

considers it as due to a factor similar to that producing

the dominant ‘‘English’’ spotting of rabbits. He refers

to Morgan’s work with black-eyed white and self mice

as being a study of this dominant factor in mice. Mor-

gan himself suggests that if black-eyed white mice repre-

sent the extremes of the spotted series the appearance

of spotted animals in crosses with selfs is due to a

strengthening of the spotting factor or to a change in

dominance. This is far different from supposing the

addition of an entirely new inhibiting factor comparable

to the English pattern in rabbits. Cuénot with mice and

Castle (1905) with guinea-pigs have shown that black-

eyed whites are the extreme of the recessive spotted

series and it is almost certain that Morgan’s explanation

of the results, as due to a change in dominance, is the

correct one. It is, of course, obvious that the presence

and absence hypothesis fails to explain any change of

dominance of a single character.

To treat ‘‘dominant’’ spotting in mice as due to the

presence of a definite unit-character is exceeding present

experimental facts, while to consider it similar in nature

to the ‘‘English’’ spotting of rabbits is still less justified.

One other point in Hagedoorn’s work is of such a

nature as to require further experimentation before it

can be accepted.

This is the case (on page 126) of ‘‘mutual repulsion be-

tween two factors.” In this case, Hagedoorn mated to-

gether agouti animals heterozygous in factor A (for color

production) and in factor G (for the agouti pattern).

No. 566] SPOTTING IN MICE 81

Such animals would ordinarily form gametes AG, Ag, aG

and ag in equal numbers. These by independent recom-

bination would form

1 AAGG

2 a) 2

9 agouti,

2 AAGg

4 AaGg

1 AAgg

2 Aagg

1 aaGG

2 aaGg ‘4 albino.

1 aagg

je black,

But Hagedoorn gives figures which show that the pro-

portion which he obtains is nearer 2 agouti; 1 black and

l albino. This he supposes to be due to the fact that A

and G can never go into the same gamete.

Now let us see what happens if this is the case. The

original heterozygotes will form only two kinds of gam-

etes instead of four, these will be aG and Ag. Now in the

recombination of these gametes the following result will

be obtained.

1 aG aG = 1 albino,

2 aG Ag = 2 agouti,

1 Ag Ag = 1 black.

So far, so good, but the trouble comes in testing the

albinos. Here I may quote from Hagedoorn, p. 126:

- . . thirteen of these albinos have been tested by mating with black.

Without exception they have given black or equal numbers of black

and albino young. ... But never has one of those albinos produced

a single agouti young in a mating with black. Counting together the

colored young of such families I get 89 black young.2 s

This result is indeed remarkable, for on Hagedoorn’s

own hypothesis the albinos should have produced in such

matings nothing but agouti young, ‘‘since they are all,

by his hypothesis, homozygous for the agouti factor.

The evidence is incontestable; no repulsion of A and G

can have occurred. Has there been any coupling of these

two factors? If such was the case only gametes AG and

82 THE AMERICAN NATURALIST [Vou XLVIII

ag would have been formed and this would have given

only agoutis and albinos in a 3:1 ratio, while Hagedoorn

reports ‘‘73 agouti, 37 blacks! and 32 albinos.’’

The case then is nothing so simple as ‘‘repulsion’’ or

‘coupling, it includes failure to segregate and com-

plete disappearance of a dominant Mendelian factor; G

the factor for agouti.

Since numerous investigators of color inheritance in

mice have never found the agouti factor anything but a

normal Mendelizing factor epistatic to black, and since

Hagedoorn himself seems to have become mixed in his in-

terpretation, it seems that the case proves or shows little

until a satisfactory answer can be found to the question

of what has become of the agouti factor.

CONCLUSIONS

The facts above given lead to the following conclu-

sions:

1. The so-called dominant type of spotting in mice does

not differ from ‘‘self’’ color by the presence of a unit

character which ‘‘self’’ lacks. The presence and absence

hypothesis fails to account for the shifting dominance

seen in spotting in mice.

2. It is misleading to describe, under the same symbol,

the so-called ‘‘dominant’’ spotting of mice and the Eng-

lish spotting in rabbits.

3. It seems probable that differences in ‘‘dominance’’

of spotting in mice are due to modifying supplementary

factors and such spotting might be termed ‘‘unsup-

pressed’’ and ‘‘suppressed”’ spotting rather than ‘‘domi-

nant’’ and ‘‘recessive’’ in the Mendelian sense.

4, Hagedoorn’s hypothesis of repulsion between the

color factor, A, and the agouti factor, G, is incorrect.

November 19, 1913.

1 Italics mine.

ON DIFFERENTIAL MORTALITY WITH RESPECT

TO SEED WEIGHT OCCURRING IN FIELD

CULTURES OF PISUM SATIVUM

DR. J. ARTHUR HARRIS

CARNEGIE INSTITUTION OF WASHINGTON

In two papers which have already appeared in these

pages, I have shown that for the dwarf varieties of

Phaseolus vulgaris the mortality of apparently perfect

seeds (failure to germinate or to complete the life cycle)

1s not random, but differential, or selective.

It seemed highly desirable to extend these studies to

other forms. Pisum sativum naturally occurred to me as

affording suitable experimental material—both because

of the wide range of seed characteristics and the conve-

nience with which it may be bred. I had no pedigreed seed

and consequently began work in the spring of 1913 with

commercial stock. About 1,000 seeds from each of ten

early (dwarf) varieties purchased from the Thorburn

Seed company were weighed, individually labelled and

planted in short rows scattered over one of the fields of

thé Station for Experimental Evolution. Conditions

were not the best, and the mortality was high.

Table I? gives the weights in units of .025 gram range®

1 Harris, J. Arthur, ‘‘On Differential Mortality with Respect to Seed

Weight Occurring in Field Cultures of Phaseolus vulgaris,’’? AMER. NAT.,

40; 512-525, 1912; ‘t Supplementary Studies on the Differential Mortality

with Respect to Seed Weight in the Germination of Garden Beans,’’ AMER.

Nat. [in press].

* For convenience the series may be designated by letters: A, Witham

Wonder; B, American Wonder; C, Premium Gem; D, Little Gem; E, Nott’s

Excelsior; F, Sutton’s Excelsior; G, Laxtonian; H, Little Marvel; I, Peter

Pan; J, English Wonder.

ass 1==0,000-.025 gram, ... class 4==.075-.100, elass 5==.100-

-125, and so on. Thus to obtain means or standard deviations of weights in

grams, deduct .5 from the values in the tables and multiply by .025.

84 THE AMERICAN NATURALIST [Vou. XLVIII

TABLE I

WEIGHT OF SEEDS WHICH GERMINATED

Series | 4 | 5 | 6 7| 8| 9 | 10 11 |12 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | Totals

A |1/ odia 75| 57| io alai the tee — me

B |—|—|—|—| 40 107/134/170/105| 35} 12! 1/—|—|—|—|—| 604

¢ |=] $] 51 53117173120 50 11) oe ee M9

D |t] 31 341061911671 80| 18! 4 i=l- | ~| — | 806

p E ||. E eh E dae

s sme ed pe Aaaa gp eee me

G |—|—|—|—| 1 71/114|1591132| 81| 23 16| —|—|—| 631

a — rs ER " 44/116|183|151| 63| 20, 6 2|—|—|—/|—| 603

I |—|2/—| 2| 6' 8| 16| 25| 51| 881127126 101| 55|17\ 8 | 1| 633

J eja oe 97 107,126 142/112 41| 13 | sae eae eS.

TABLE II

WEIGHT OF SEEDS WHICH FAILED TO GERMINATE

Series | 4 | 5 6|/7|8 | 9 | 10 11 | 12 | 13 14 | 15 | 16 | 17 | 18 | 19 | 20 Totals

A 4 |1649 63 102 90 93 78 42| 16 2 Pe iat DSPs OS) een ee ee

B |—|—|—| 2|—| 52/104/117| 84| 36 3|—| 2)—|—|—|—| 400

C. |—| 2| 1/12] 56/116|122| 95) 34) 18| 4| 1|—|—|—|—|—| 461

D =i] E Bl Oe ener) Wel Be BU Se ae ee el ae

E |—|—|—|/13/115\210180 91| 23 4 1 — — — — —|— 637

PI 46| 84125153114 49 | 16| 10| 3|—|—|— | 613

G |—|—|-—|— 43 5|70|42|13| 7|—|—| 1 | 376

H |—|—|—| 11} 17] 66).71/110| 93| 20/14| 3|—|—|—|—|—| 404

F iis 1/19 18 26 25 29 G0 65 02 3I 22 3 2 |4| 374

J |—|—|—] 18} 55} 88)125| 91] 53) 12) 2|—|—|—|—|—|—| 430

of the seeds which germinated.t| Table IT gives the same

distributions for the seeds which failed to germinate.

The physical constants® with their probable errors are

given in Tables III-IV.

Taking the differences, germinated less failed, in order

to have the positive sign if elimination tends to increase

mean weight or variability of weight and the negative

sign if it tends to decrease these constants in the popula-

tion of seeds which grow as compared with those which

fail, I find the differences shown in Table V.

4 When the plantlets were about three inches high the labels for seeds

which had failed to germinate were collected. The distributions for the seeds

which had germinated were then sano by subtraction from the weight

seriations prepared before planting. Some of the plants epoca! died.

5 Sheppard’s correction was applied to second momen

No. 566] DIFFERENTIAL MORTALITY 85

TABLE III

PHYSICAL CONSTANTS FOR SEEDS WHICH GERMINATED

Standard Deviation and Coefficient of Vittles”

Series | Mean and Probable Error | P ble E | and Probable Error

A 9.254 + .062 j f | 21.610 +.498

B 10.581 + .038 1.371 +.027 12.954 + .256

C 10.078 + .037 | 1.294 + .026 | 12.844 + 266

D 10.355 = 58 | 1.211 =.023 4 mis =+,230

E 9.790 + 1.193 +.030

F “i 568 = ron 1.571 + .038

G 3.090 = .04 1.612 +.031 t Per +,237

H Sr 186 =.037 | 1.362 = .026 | 12.178 +.240

I 4.269 + .057 | 2.134 + .041 | .958 =.

J 9.773 + .040 1.429 +.029 | 14.622 +.300

TABLE IV

PHYSICAL CONSTANTS FOR SEEDS WHICH FAILED TO GERMINATE

Series Mean and Probable Standard Deviation and Coefficient of Variation

Error Probable Error and Probable Error

A] 8.993 +.057 2.003 =.041 22.286 + .472

B | 10.898 =.041 1.236 +.030 11.346 +.274

C 9.913 =.045 1.439 + .032 14.512 +.329

D 10.048 = .041 1.229 + .029 12.234 +.

E 9.488 = .030 1.122 + .021 11.826 + .227

F 11.726 1.653 + .032 14.097 +.

G 12.816 + .062 1.787 + .044 13.945 +.350

H 10.869 =.049 1.447 +.034 13.317 +.322

I 13.225 = .089 2.552 = .063 19.298 +.493

J 10.009 + .044 1.376 +.031 13.749 +.311

TABLE V

COMPARISON OF PHYSICAL CONSTANTS FOR SEEDS GERMINATING WITH THOSE

FoR SEEDS FAILING TO GERMINATE

Differe in Coefficient

Series re Probable Error of ome of oe aig and Probable

Error of Difference Error of Difference

A + .261+.085 —.003 +.060 —1.676 + .686

B — .316 +.057 +.134 +.040 +1.608 +.375

C + .165=.058 —.144 +041 =]

D + .307 +.053 —.019 +.037 — 542 +.375

E +. 1 +.071 +.036 358 +.382

r — .158+.070 + 049 513 +.434

G + 274+.075 —.175 + .054 —1.632 +.422

H + 317 +.062 —.085 + .044 —1.139 +.401

I +1.044 +.105 —.418 +.074 —4.340+

J — .236+.060 = + .873 +.432

Consider first the differences in the mean weight.

Seven are positive and three are negative. All of the

86 THE AMERICAN NATURALIST [Vou. XLVIII

seven positive differences are at least 2.5 times their prob-

able error; four of them are over five times their prob-

able error. The mortality is therefore almost certainly

selective, with a tendency to leave the surviving popula-

tion with seeds distinctly heavier on the average than

those which were planted. On the other hand, there are

the three cases in which the seeds which produced plant-

lets were on the average lighter than those which failed

to germinate. One of these differences is only 2.2 times

its probable error, and so perhaps not statistically trust-

worthy. Of the other two, one is over 5.5 times and the

other nearly 4 times its probable error. There can be

little doubt that in at least one of these cases there is a

tendency for the lighter seeds to show a viability greater

than that of the heavier. In garden beans, too, strong

evidences of differences between strains in this regard

have been pointed out.

The interpretation of the variabilities offers greater

difficulties than does that of the means. More data and

more refined methods of analysis are necessary for a final

solution of the problem. It appears, however, that in

seven of the ten series the variability of the seeds which

survived is less than that of those which failed. This is

true whether absolute variability as measured by the

standard deviation or relative variability as expressed

by the coefficient of variation be used in the comparison.

As far as these data go, therefore, they are in general

accord with those for Phaseolus. In both of these Legu-

minose the mortality which occurs before germination is

not random but differential. But in both cases, and espe-

cially in Pisum where the seeds used are of commercial,

not pedigreed, stock and number as yet only about 10,000,

far wider series of experiments and much refinement of

methods of analysis are necessary to establish fully the

nature and the immediate (physical or chemical) cause of

this selective death rate.

COLD SPRING HARBOR, N, Y.,

July 28, 1913

THE INHERITANCE OF A RECURRING SOMATIC

VARIATION IN VARIEGATED EARS

OF MAIZE!

PROFESSOR R. A. EMERSON

UNIVERSITY OF NEBRASKA

INTRODUCTION

Taz inheritance of variegation has special interest and

importance in genetics. It is with forms of variegation

that the only two certainly known cases of non-Mendelian

inheritance have had to do. I refer to Baur’s experiments

with Pelargonium, in which crosses of green-leaved and

white-leaved forms exhibited somatic segregations in F,

that bred true in later generations, and to Correns’s work

with Mirabilis, which showed green and white leaf color,

to be inherited through the mother only. De Vries’s con-

ception of ‘‘ever-sporting’’ varieties was apparently

founded largely upon the behavior of variegated flowers

in pedigree cultures, from which he reached the conclusion

that the variegated color pattern and the monochromatic

condition arising from it as sports are non-Mendelian in

inheritance. Correns, however, has shown that in Mira-

bilis jalapa the inheritance of these sports is distinctly

Mendelian, and the results of East and Hayesindicate the

Same for Zea mays. In this paper I shall present data

from maize and attempt to show how they can be inter-

preted in strictly Mendelian terms.

Variegation is distinguished from other color patterns

by its incorrigible irregularity. It is perhaps most often

seen in the coloration of flowers and leaves but also occurs

in fruits, seeds, stems, and even roots of various plants.

It is characteristic of the ears of certain varieties of maize

known, at least in the Middle West, as ‘‘calico’’ corn. In

1 The experimental results reported here were presented at the Cleveland

meeting of the American Society of Naturalists, January, 1913. Research

bulletin No. 4 of the Nebraska Agricultural Experiment Station.

88 THE AMERICAN NATURALIST [Vou. XLVIII

these varieties the pericarp of most of the grains has few

to many narrow stripes of dark red, the remaining area

being colorless or showing a sort of washed-out red.

Often broad red stripes appear on some grains, a single

stripe covering from perhaps one tenth to nine tenths of

the grain. Not uncommonly there are entirely colorless

grains (so far as pericarp is concerned) and also solid red

grains scattered over the ear. Much more rarely there

is found a ‘‘freak’’ ear with a large patch of self-red or

nearly self-red grains. Or sometimes an ear is composed

largely of red or almost red grains with a small patch of

striped or nearly colorless grains. In such cases it is not

uncommon for the margin of the red area to cut across a

grain so that one side—always the side toward the red

patch—is red and the other side colorless or striped. Ears

that are colorless throughout, except for a single striped

grain, are not unknown and there are even known ears

that are red except for a single stripedgrain. Very rarely

a plant has one self-red ear and one variegated ear on the

same stalk. It is also conceivable that all the ears of a

plant might thus become red, but of course such a red-

eared plant rising as a bud-sport could not ordinarily be

distinguished from a red-eared plant arising as a seed-

sport.

Variegated ears generally have variegated cobs, the

amount of red in the cob ordinarily varying with the

amount of red on the grains. In some ‘‘freaks’’ a part

of the cob is solid red and the rest variegated. In a few

such cases the red part of the cob corresponds exactly in

position to the freak patch of grains. This is more fre-

quently true when the grains of the freak patch are dark

variegated than when they are self-red. In other ears

there is no change in the cob corresponding to the change

in the grains. The husks of variegated ears are also

rather commonly variegated. In a few freak ears the red

side of the ear is enclosed in reddish husks, the remainder

of the husks being light striped. Red-eared plants aris-

ing as seed-sports always have solid red cobs and usually

solid reddish husks.

No. 566] INHERITANCE IN EARS OF MAIZE 89

The first account, so far as I am aware, of the inherit-

ance of the striking somatic variations so commonly found

in variegated plants was given by de Vries? in his dis-

cussion of ever-sporting varieties. The study was made

in the years from 1892 to 1896 with a variety of Antir-

rhinum with striped flowers. De Vries’s records are re-

produced diagrammatically in Fig. 1.

Pi Striped

plant

l |

Fi Striped Red

plants plants

90% 10%

F: Striped Red

plants plants plants plants

98% 2% 24% 76%

l |

Striped Red

branches ERES

F; Striped Red Striped ed

plants plants plants plants

98% 2% ý 71%

Bs Stri iped Red

plants plants plants plants

95% 5% 0

FIG. 1. DIAGRAM FROM DE VRIES’S RECORDS SHOWING THE INHERITANCE OF

VARIEGATION AND SELF-RED IN THE FLOWERS OF Antirrhinum.

Of these results de Vries says:

From these figures it is manifest that the red and striped types differ

from one another not only in their visible attributes, but also in the

degree of their heredity. The striped individuals repeat their peculiarity

in 90-98 per cent. of their progeny, 2-10 per cent. sporting into the uni-

form red color. On the other hand, the red individuals are constant in

71-84 per cent. of their offspring, while 16-29 per cent. go over to the

striped type. Or in one word: both types are inherited to a high degree,

but the striped type is more strietly inherited than the red one.

De Vries’s results were in some respects very similar

to those of Correns and it is probable that he would have

interpreted them in the same way had he then been famil-

lar with Mendelian phenomena.

2 Vries, Hugo de, ‘‘ Species and Varieties,’’ pp. 309-328 (1905).

90 THE AMERICAN NATURALIST [Vow. XLVIII

Correns? has reported results of a careful study of the

inheritance of the self-green condition appearing as a `

bud-sport on variegated-leaved plants of Mirabilis jalapa,

and also of a self-color appearing in striped-flowered

plants of the same species. His results for self-green

variegation of the leaves are shown diagrammatically

ig. 2. The results are stated in approximate per-

jo I have seen no report in which the detailed

records were given.

Variegated

ah i i

Variegated Green

branch air We

Variegated reen Variegated Green

plants plants plants

100-a* a 25

F: |

>66 <33 = 33

Peart p ETE,

Vgtd. T E Green Green Vgtd. Green Vgtd. Green Green

pl 100 100-a a 25 75 100

Fs S aE

pp ol >66 <33 66 33

no a h h | a | Be

ye ee Gy VG

i ee Mi 5 BA ARA Sas 78 o 100 oa A E Nd a a 28 T 100 100

Fic. 2. CORRENS’S DIAGRAM SHOWING THE INHERITANCE OF VARIEGATION AND

LF-GREEN IN THE LEAVES OF Mirabilis jalapa

The diagram shows that a variegated branch of a varie-

gated plant produces in F, mainly variegated plants, but

occasionally a wholly green plant, while a green branch

from the same plant produces in F, 25 per cent. varie-

gated and 75 per cent. green plants. The F, variegated

plants, however produced, behave in later generations

just like the original variegated parent plant. The F,

green plants, whether produced from green or variegated

branches, are always of two sorts, namely, those that are

homozygous and therefore breed true green, and those

3 Correns, C., Ber. Deutsch, Bot. Gesel., 28: 418-434, 1910. Der Uber-

gang aus dem homozygotischen in einen heterozygotischen Zustand im selben

Indiwiduum bei buntblattrigen und gestreiftbliihenden som ae eae

* Numerals indicate approximate percentages; a— 0-10 per

No. 566] INHERITANCE IN EARS OF MAIZE 91

that are heterozygous and therefore produce progenies of

green and variegated individuals in a ratio of approxi-

mately 3to1. Correns points out that a green branch of

a variegated plant behaves as though it belonged not to a

variegated plant at all, but to a hybrid between a varie-

gated plant and a green one, in which green is dominant,

and that half of the germ cells produced by the green

branch carry a factor for green and the other half a factor

for variegation. Similar results were secured from

branches with self-colored flowers on plants with striped

flowers, except that such branches produce few if any

more self-colored plants than are produced by branches

with striped flowers. Plants with self-colored flowers, no

matter how they arise, behave as they would if they had

occurred in an F, progeny of a cross of striped by self-

colored plants.

RESULTS oF EXPERIMENTS WITH MAIZE

Hartley* in 1902 gave an account of an experiment with

variegated maize. In a comparatively pure white strain,

which occasionally produced a red ear, there was found an

ear similar to some of the‘‘freak’’ ears noted earlier in

this paper. Itis described as being red except for a spot

covering about one fifth of the surface, in which the grains

were white with fine red streaks. The excellent plate ac-

companying the account, however, shows that most of the

“‘red’’ grains had white streaks at the crown and that the

cob was light-colored, not red. From the near-red grains

of this ear there was produced a crop of 84 red ears and

86 pure white ones, while from the variegated grains of

the same ear there came 39 light variegated ears and 36

white ones. Hartley refers to the parent ear as a ‘‘sport

or sudden variation from the type’’ but does not indicate

whether the ‘<type’? in mind was the white variety or the

red ears occasionally produced by it. Both the color of

the grains and cob and the production of about 50 per

cent. of white ears from both the red and the variegated

grains indicate very clearly that the parent ear was a

* Hartley, C. P., Yearbook, U. S. Dept. Agr., 1902: 543-544.

92 THE AMERICAN NATURALIST [Vou. XLVIII

heterozygous, variegated one and that it probably came

from a white seed crossed by a stray grain of pollen from

a variegated-eared plant, just as the occasional red ears

in the white variety were certainly produced by stray pol-

len from red-eared plants.

More recently Hast and Hayes® reported like behavior

of a similarly variegated ear. An ear having on one side

solid red grains and on the other white and very light

variegated grains, similar to some of the ‘‘freak’’ ears

noted earlier in this paper furnished the material for the

test. The ear was produced from a white seed in a field

of otherwise pure white corn and was therefore doubtless

heterozygous for pericarp color and was probably pol-

linated in large part from plants without pericarp color,

so that 50 per cent. white-eared plants were to be expected

in its progeny. The white, the light variegated and the

solid red grains were planted separately. The white and

the variegated seeds alike produced light variegated and

white ears, 15 of the former and 15 of the latter. Thered |

seeds produced 22 whiteearsand 22solidred ears. The

authors’ interpretation of these results is that the white

seed which gave rise to the original colored ear had been

fertilized by pollen from a red-eared plant and that the

F, plant, ‘‘due to produce a red ear varied, somatically so

that one half of the ear was red and one half striped.”

The authors further state:

This variation was transmitted by seeds, but at the same time the

hybrid character of its seeds was unchanged as shown by their segrega-

tion into reds and whites in the next generation and the normal segre-

gation of the hybrid dark reds in a further generation.

In the light of my own observations, it is equally pos-

sible and seems more likely that the white seed from which

the original red-and-variegated ear came was the result

of pollination from a plant with variegated ears, and that

the somatic variation was from variegated grains to solid

red grains rather than from red to variegated. But the

important fact is that a somatic variation was later in-

herited in a strictly Mendelian way.

5 East, E. M., and Hayes, H. K., Bul. Conn. Agr. Expt. Sta., 167: 106-107.

1911. l

No. 566] INHERITANCE IN EARS OF MAIZE 93

In 1909 I obtained results somewhat similar to those re-

ported by East and Hayes. A few ‘‘freak’’ ears were

secured, mainly from local and national corn expositions.

Nothing was learned as to their parentage or pollination.

Obviously, however, the parentage of the red, the varie-

gated, and the white grains of any one ear was the same,

and it is reasonable to suppose that the different sorts of

grains of any one ear were pollinated with approximately

the same kind or the same mixture of pollen. The results,

as shown below, were essentially like those of Hartley and

of East and Hayes.

Number of Plants with

Variegated Ears White Ears

Seeds Planted —— 2

Red Ears |

UIE as os eyes 43 | 0 33

Variegated and white. ......| 0 | 22 29

The results from four other ears were somewhat differ-

ent, probably owing to differences in their pollination.

(See Fig. 3.) They were as follows:

progeny of

Fic. 3. A, “freak” ear of maize; B, progeny of striped seeds; C

self-red seeds.

94 THE AMERICAN NATURALIST [Vou. XLVIII

Number of Plants with

Seeds Planted ;

| Red Ears | Variegated Ears | White Ears

ARR GPa gia OA et ite 128 | 32 | 69

8 | 103 68

Two other ears of similar history, while they gave quite

as striking results as those noted above, probably do not

belong here since none of their immediate progeny were

variegated and no variegated ears have occurred in later

generations. These two ears were made up of red grains

and white grains only. The results were as follows:

Number of Plants

Seeds Planted Red Ears White Ears

BOG oe T A oa ee 77 85

While ores wi es 0 jy ae

The white ears bred true in later generations and the

red ears produced reds and whites in typical Mendelian

fashion. No such somatic variations as these have oc-

curred in my cultures of self-red or white maize, so that I

have been unable to study them further. Somatic varia-

tions in variegated corn, however, are not rare. Unfor-

tunately several of the most pronounced of those occur-

ring in my cultures were open-pollinated and therefore

of little or no use in a careful study. I have therefore

been obliged to make use in large part of the few solid

red and nearly solid red grains scattered over otherwise

more or less evenly variegated ears.

From twenty-three self-pollinated, variegated ears of

plants that were homozygous for pericarp color, grains

with various amounts of red were selected and planted.

The results are summarized as follows:

Number of Plants with

Seeds Planted

Self-red Ears Variegated Kars Non-red Ears

PO ects cee tee enes 8 9 0

Nearly self-red............ 56 : 16 0

More than half red........ 9 34 0

an Py 3 Rane eres 5 22 0

Narrow red stripes......... 33 * 394 0

vee ken 1 22 0

No. 566] INHERITANCE IN EARS OF MAIZE 95

Besides these 23 ears, 20 other selfed ears from homo-

zygous plants contained only narrow-striped seeds from

which there were produced 16 plants with red ears, 280

with variegated ears, and none with white ears. Similarly

21 selfed ears with narrow-striped seeds only, from plants

that were heterozygous for pericarp color, produced 28

plants with red ears, 411 with variegated ears, and 208

with non-red® ears. Variously colored grains from 42

self-pollinated, heterozygous, variegated ears gave the

following results:

Number of Plants with

Seeds Planted

Self-red Ears Variegated Ears | Non-red? Ears

tees OE Ro ee ee 15 1 | 6

easly irad | (6. i, 17 8 | 8

More than one half red..... 46 51 31

Less than one half red...... 8 34 | 21

Narrow red stripes......... 57 767 300

ROU Go rut tea a cabs 0 10 | ON

In the progenies of these 63 self-pollinated ears that

were heterozygous for pericarp color, there were approxi-

mately 2.5 plants with pericarp color to one without it.

All the classes of grains from self-red to non-red yielded

both colored and non-colored ears, thus indicating, as

already shown by East and Hayes, that the somatic varia-

tion in the seeds does not change their hybrid character.

Considering only the plants with pericarp color, in the

progenies of both heterozygous and homozygous varie-

gated ears, 106 progenies in all, marked differences are

Seen in the percentages of self-red ears from seeds of the

different color classes, as follows:

* Some of these ears had what I have termed ‘‘half-red’’ pericarp, i. e.,

pericarp with a reddish color extending part way from the base to the

crown of the seeds. (See Ann. Rpt. Nebr. Agr. Expt. Sta., 24: 62. 1911.)

Half-red differs from self-red and variegated red not only in distribution

but also in almost never developing fully in the heterozygous condition. It

is hypostatie to self- but shows between the red stripes of variegated

Seeds. Since its presence does not mask either self-red or variegated-red

and since it is still aiekeeerskie to both of them, half-red is here in-

cluded with non-red. Variegated ears =a never, in my observation, pro-

duced half- ok grains as somatic variat

T Some of these were half-red. See hoiii 6.)

96 THE AMERICAN NATURALIST [Vou. XLVIII

Number of Plants with Per Cent. Self-red

Seeds Planted ans | Colored

Self-red Ears Variegated Ears ars

EPO oo ee ae ee 23 10 69.7

Nearly. self-red oo a 73 24 75.3

than one half red..... 55 85 39.3

Less than one half red...... 13 56 18.8

Narrow en 2h NA 134 1,852 6.7

INOW TON er 32 3.0

In comparison with the cases reported by Hartley and

by East and Hayes and one of my first cultures from

open-pollinated ears, in all of which red grains produced

no variegated ears and striped grains no red ones, the

striking features of the results from these 106 self-pol-

linated ears are the facts that the wholly red grains

yielded some variegated as well as red ears and that the

striped grains and even the wholly non-red grains yielded

some red as well as variegated ears. The percentages noted

above indicate in a general way that for self-pollinated,

variegated ears, the more red there is in the seed planted

the larger the percentage of red ears in the progeny.

These records, however, do not give a wholly trustworthy

indication of the mode of inheritance of the somatic vari-

ations concerned here. If there is a modification of some

factor in the female gametes, associated with a visible

modification of somatic cells of the pericarp and even at

times of the cob and husks, modifications that do not be-

come visible until long after the gametes are formed, may

there not be a similar modification of the same factor in

the male gametes, though here not associated with any

visible change in somatic cells because of the fact that the

staminate inflorescence dies too soon after the pollen is

shed? If male gametes do carry such modified factors

and if the modification is as irregular in occurrence as the

somatic modifications seen in variegated ears, so that any

part of the tassel, from all to none, may produce gametes

with the modified factor while not showing any visible

somatic modification, it is obvious that the real nature of

the male gametes of any variegated-eared maize plant

can not be foretold. The mere fact that a variegated ear

is self-pollinated, therefore, does not insure that its seeds

are fertilized with pollen of known character.

No. 566] INHERITANCE IN EARS OF MAIZE 97

That the male gametes of variegated-eared maize do

often carry factors for self-red is shown by crosses of

pure non-red strains with pollen from plants with varie-

gated ears. The plants that furnished the pollen for

` these crosses were in some cases the same ones whose self-

pollinated ears were concerned in the records discussed

above. The results of these crosses are summarized here.

Hight non-red ears crossed by plants that were homozy-

gous for pericarp color yielded 17 red-eared, 116 varie-

gated-eared and 8 white-eared® plants. Similarly, 14 ears

of pure non-red strains crossed by pollen from plants

heterozygous for pericarp color yielded 26 red-eared, 192

variegated-eared and 229 white-eared plants. Consider-

ing merely the plants with colored ears, 22 crossed ears

produced 43 red-eared to 308 variegated-eared plants, or

a little over 12 per cent. self-red.

Since the male gametes of variegated-eared corn have

now been shown occasionally to carry a factor for self-

red, it is obvious that only from crosses of variegated-

eared plants with pollen from pure non-colored strains,

can a definite idea of the inheritance of the somatic varia-

tions in pericarp color be gained.® Twelve ears from

homozygous, variegated plants cross-pollinated by non-

red strains might have afforded important evidence, but

for the fact that 7 of them contained only narrow-striped

grains and the other 5 no fully or even nearly self-red

grains. The results are summarized here:

Number of Plants with

Seeds Planted

Self-red Ears Variegated Ears Non-red Ears,

More than one half red..... 5 11 P

ss than one half red...... 0 15 Q

Narrow red stripes......... 2 281 9

Nli hacer nen oz 2 =

Some of the 8 white ears may have been extreme light types of varie-

gation, for in some other cases very light variegated and wholly white ears

have been observed on the same plant. And of course some of them may

have been due to accidental pollination of the parent ear.

® Though the genetic factors for pigment patterns in maize seem to be

distinet from the factors for the pigment concerned in these patterns, no

non-colored maize that I have used has ever given any indication in crosses

of carrying pattern factors.

98 THE AMERICAN NATURALIST [Vou. XLVIII

The principal facts of interest here are the production

of only one red-eared plant to about 140 variegated-eared

ones from narrow-striped seeds, and of about one red-

eared to two variegated-eared plants from seeds with

from one half to perhaps three fourths red.

Of 20 variegated ears, heterozygous for pericarp color,

that were crossed with pollen from pure non-colored

strains, 5 had only narrow-striped grains and 15 had

variously broad-striped grains and even some self-red

ones. The summaries of these crosses are as follows:

Number of Plants with

Seeds Planted

Self-red Ears Variegated Ears | Non-red Ears

Seli-ted ee es 9 0 11

Nearly self-red............ 5 0 2

More than one half red..... 4 2 | 2

th half red...... 3 5 | 9

Narrow red stripes......... a 265 | 301

E ae 0 27 | 20

Here again, just as with homozygous, variegated ears,

the more red there is in the pericarp the more likely are

the female gametes to carry a factor for self-red. While

the number of individuals dealt with are too few to afford

reliable evidence, it is suggestive to note that the ratio of

red-eared to variegated-eared plants, though not the ratio

of red-eared to total plants, is greater in case of parent

ears that are heterozygous than of those that are homozy-

gous for variegated pericarp.

So far nothing has been said of the results in genera-

tions later than the one grown from the selected seeds

(F). Let us now see what results follow when the varie-

gated ears and the red ears produced as explained above

become the parents of second generations (F,) from the

selected seeds. The variegated ears so produced behave

like the original variegated ears from which seeds were

selected and their progenies have, therefore, been included

in the data already presented. There remains only to

present the records of the progenies of red ears.

Data are available from 7 F, red ears obtained from

self-pollinated, homozygous, variegated plants. Five of

No. 566] INHERITANCE IN EARS OF MAIZE 99

these red ears were self-pollinated and two were crossed

with pure white-eared plants. The results in F, and F,

were as follows:

Number of Plants with

Seeds Planted from Self-red | Variegated | Non-red

Ears Ears Ears

F reds from selfed, homo., vgtd. P:’s

EN A Se a Shine oe cea ee eae 119 37 0

Ore SC WN se i Eee cess 46 45 0

F2 reds from selfed F; reds

a ais ek its bie eS ies ek eo 9 2 0

SUF SOUGG Fk 6 E E 16 0 0

F: reds from Fi reds X white

OS ONTO 58 Ve te bc ee eo ae Os we T 26 0 5

EE ONE X WAI EET ee 40 0 37

The above is approximately what would have been ex-

pected, had the F, red ears that arose from self-polli-

nated, homozygous, variegated-eared plants been pro-

duced by a cross between red-eared and variegated-eared

races.

Of the F, reds arising from self-pollinated, heterozy-

gous, variegated-eared plants, nine were selfed and two

were crossed with whites. The results secured in F, and

F, follow:

Seeds Planted from

Fı reds from selfed, hetero., vgtd. Pi’s

OREO ONIN CON oe ok cc occa a 104 23 0

LO A WN a a a Ae 6 7 0

Smee bled ocea a 105 0 38

1 ar X White... ....* eed 12 0 7

F2 reds from selfed F; reds of (a)

sopra! oy o ma earn Pie Poet icy re 59 12 0

RO OER eo ae a ee 23 0 0

From the above it appears that the F, red ears, arising

from self-pollinated, heterozygous, variegated-eared

plants behave in some cases as if they were hybrids be-

tween red-eared and variegated-eared races and in other

cases as if they were hybrids between red-eared and

white-eared races.

Of the four possible sorts of red-eared ‘‘sports’’ from

variegated-eared plants, two remain to be treated. Be-

100 THE AMERICAN NATURALIST [Vou. XLVIII

cause of their similar behavior they will be considered

together here. Of the F, red ears arising from homozy-

gous, variegated-eared plants that had been crossed with

white-eared races, three were self-pollinated and two

crossed with whites. Of the F, red ears arising from

heterozygous, variegated-eared plants that had been

crossed with white-eared races, four were selfed. The

results in F, and F, are:

x oa | Vari -

Seeds Planted from = i Mes! ipa se oe

Fı reds from vgtd. Pi’s X white

P;’s homozygous

Goars BONERS FSCS ees 54 0 16

2 Carn K WHOS. ce wi kev et ee ne oe 34 0 43

Prs heterozygous

ES BOCES i oc otis 5 a eke ng a 102 0 47

F: reds from selfed F, reds

3 Gate sella i eee eek 32 0 10

E a a A E ees E 43 0 0

So far as these results go they indicate that F, reds

arising from crosses between both homozygous and heter-

ozygous, variegated-eared plants and white-eared races

behave as if they were hybrids between red-eared and

white-eared races.

One homozygous, variegated-eared plant was cross-

pollinated by a homozygous red race. From the varie-

gated ear produced, self-red, nearly self-red, and narrow-

striped seeds were planted. All resulted, of course, in

red-eared F, plants, 16 in all. A self-pollinated F, red

ear from a narrow-striped seed gave in F, 24 red-eared

and 11 variegated-eared plants—somewhat fewer reds

than were to have been expected. An F, red ear from a

nearly self-red grain, when cross-pollinated with non-red,

yielded 9 reds and 11 variegated in F,. A third F, red-

eared plant, this one from a self-red grain of the varie-

gated parent ear, bred true red in F,. One ear of this F,

plant was selfed and yielded 14 reds in F,, and another

ear was cross-pollinated by non-red and yielded 29 reds.

There are various other somatic variations rather fre-

quently seen in maize, but they are apparently not in-

No. 566] INHERITANCE IN EARS OF MAIZE 101

herited. There are sometimes found variegated ears

with a large patch of self-red cob but with little or no cor-

responding change in the color of the overlying grains.

I have as yet no evidence that this somatic variation in

cob color is inherited through the seeds of the self-red

part of the cob. Such seeds apparently always produce

ears with variegated grains and variegated cobs, just as

do other seeds of the same parent ear. Of course varie-

gated seeds from a self-red patch of cob occasionally give

rise to a self-red ear, as discussed in detail in this paper,

and such red ears always have self-red cobs, but this is

also true of all self-red ears, whether or not they are pro-

duced by red or by variegated seeds and without respect

to whether the part of the cob underlying these seeds is

self-red, finely variegated, or entirely white.

Another form of somatic variation seen in ears of maize

is the occurrence of patches of considerable size, the

grains of which, though variegated, are much darker in

color than the grains of the rest of the ear. Such patches

of grains are often quite as strikingly distinct in appear-

ance as patches of self-red grains, and are apparently

even more likely to correspond exactly in outline with an

underlying patch of self-red cob than are patches of self-

red grains. Moreover, such dark, variegated grains often

present a rather definite color pattern. The crowns are

often made to appear almost solid red by the widening —

and convergence at the crown of narrow red stripes ex-

tending down toward the base of the grain particularly on

the side opposite the germ. Another type of dark, varie-

gated grains differs from the lighter, variegated grains

of the same ear principally in the greater development of

the somewhat washed-out red apparently underlying the

dark red stripes of the variegation pattern proper. I

have grown numerous progenies from dark and light

Variegated grains of the same ears, but as yet have no

evidence that such somatic variations are inherited. Not-

withstanding this, I have strains of maize breeding true

to a very dark type of variegation, others to a medium

102 THE AMERICAN NATURALIST [Vou. XLVIII

sort of variegation, and still others to exceedingly light

types of variegation. There can be no doubt that some of

these different types of variegation are inherited, but the

mode of inheritance in crosses has not been fully worked

out.

One other form of grain coloration that might be called

an extremely dark type of variegation is to be noted. The

grains are self-red throughout except for a nearly color-

less crown formed by converging light stripes extending

some way down the side of the grain opposite the germ,

almost exactly the reverse of one of the types of dark

variegation described above. Variegations of this sort

behave in inheritance almost exactly like fully self-red

grains, giving a large percentage of red-eared progeny.

And these red ears are apparently always fully self-red,

never showing the pattern of converging light lines seen

in the parent seeds. Many such seeds have been included

in the results recorded earlier in this paper where they

were listed as ‘‘nearly self-red.’’

INTERPRETATION OF RESULTS

Any interpretation of the data presented here must take

account of these facts: (1) that the more red there is in

the pericarp the more frequently do red ears occur in the

progeny, and (2) that such red ears behave just as if they

were F, hybrids between red and variegated or red and

white races. The development of red in the pericarp is

` evidently associated with and perhaps due to a modifica-

tion of some Mendelian factor for pericarp color in the

somatic cells. The zygotic formula of a plant homozy-

gous for variegated pericarp may be designated as VV,

and that of a plant heterozygous for variegated pericarp

as V—. If in any somatic cell VV, from unknown causes,

a V factor were transformed into a factor for self-color,

S, that cell would then have the formula VS. Any peri-

carp cells descended from it would without further modi-

fication be red. If all the pericarp cells of a seed were

thus descended, the seed would be self-red, just as it would

No. 566] INHERITANCE IN EARS OF MAIZE Ț 103

if the plant bearing it were a hybrid between pure red and

variegated races. Moreover, one half of the gametes

arising from such somatic cells would carry V and one

half would carry S, just as if the plant were a hybrid of

red and variegated types. Or, if both V factors were

changed, the grains would be self-red as before, but all

instead of half the gametes would carry S. If, however,

the modification from VV to VS should occur very early

in the life of the plant, or even of the embryo, all the ears

of the plant might thereby become self-red, and one half

of all the gametes both male and female might then carry

S and the other half V as in the ordinary hybrid. Or the

plant might then become a sectorial chimera with one

variegated ear and one red ear, the gametes from the one

side of the plant all carrying V. If the modification

occur much later, say soon after the ear begins to form,

there might then be merely a solid patch of red grains on

an otherwise variegated ear. In this case only those

gametes arising from these smaller masses of tissue would

carry half S and half V. If, however, the modification

occur after the grains begin to form, the latter might be

perhaps three fourths red, or one half red, or merely have

narrow stripes of red, depending upon the amount of peri-

carp directly descended from the modified cell. In this

case it seems reasonable to assume that the larger the

mass of modified tissue the greater the chance that the

‘gametes concerned should carry S. Finally, if in certain

grains the change never occurs, they should show no red

and the gametes formed in connection with them should

all carry V, none S.

Similarly, it may be assumed that in any cell of a heter-

ozygous, variegated-eared plant, V—, the V factor may

as before become an S factor. The effect on pericarp

color would be exactly the same as in a homozygous, vari-

egated plant, and, of the gametes arising from the modi-

fied tissue, one half would carry S as in the other case,

but the other half, instead of carrying V, would carry no

factor and would be represented by —.

104 THE AMERICAN NATURALIST [Vou. XLVIII

If the interpretation suggested here is correct, it is to

be expected that the more red there is in the pericarp of

any seeds, i. e., the larger the mass of tissue descended

from the cell in which the change from V to S took place,

the greater the chance that the female gametes concerned

carried the factor S. With heterozygous, variegated-

eared plants, V—, however, never more than half of the

gametes concerned could carry S even in case of self-red

grains, the other half of the gametes carrying no factor,

—. Of the heterozygous, variegated ears the progenies

of which have been reported here, some were selfed, some

crossed with white, and some open-pollinated. From self-

pollinated ears, self-red and nearly self-red seeds yielded

32 red-eared, 9 variegated-eared, and 14 non-red-eared

plants, or practically 58 per cent. self-red. This excess of

self-red ears may be due, in part at least, to the presence

of the S factor in some of the male gametes concerned, but

the numbers are too small to give very reliable indica-.

tions. From similar ears that instead of being selfed

were crossed with white, so that the results could not have

been influenced by factors present in the male gametes,

self-red and nearly self-red seeds produced 14 plants with

red ears and 13 with non-red ears, or about 52 per cent.

red. While these numbers are very small, the fact that

no variegated ears were produced, but that every ear with

any red color was self-red, is noteworthy. From the

open-pollinated, heterozygous ears included in my cul-

tures self-red seeds gave progenies consisting of 171 red-

eared, 32 variegated-eared, and 102 non-red-eared plants,

or about 56 per cent. red.

In ease of homozygous, variegated-eared plants, VV, all

the gametes associated with seeds that later become self-

red could carry S only if both V factors of the somatic cells

from which the gametes arise were changed to S factors.

Because of the rarity of changes from V to S, unless both

V factors are influenced alike by whatever causes the

change, so that both change simultaneously to S factors,

the chance is slight that more than one will ever change.

No. 566] INHERITANCE IN EARS OF MAIZE 105

In the latter case only about 50 per cent. of the gametes

associated with self-red grains of homozygous, varie-

gated ears could be expected to carry S, just as in the

case of heterozygous ears. None of the open-pollinated

ears whose progenies I have grown were homozygous for

variegated pericarp, and none of the homozygous ears

that had been crossed with white contained any self red or

nearly self-red seeds. The only data, therefore, that bear

upon the point at issue are those obtained from self-pol-

linated, homozygous, variegated ears. The self-red and

nearly self-red seeds of such ears produced 64 red-eared

and only 25 variegated-eared plants, or about 72 per cent.

self-red. This may mean that in some cases both V

factors were changed to S factors, but the results may

just as likely be due to the presence of S in an unusually

large percentage of the male gametes concerned. The

production of the 25 variegated-eared plants, however, is

very good evidence that, in at least a very considerable

number of cases, not more than one of the two V factors

could have been changed to S.

If the change from V to S should happen to occur at such

a time that the grain rudiments became sectorial chimeras

consisting of say one half modified cells and one half un-

modified ones, one half of the pericarp would be expected

to show red color and the other half no color. It would

be expected further that the chances of a particular

gamete’s arising from a modified or from an unmodified

cell would be equal. If then one half of the gametes asso-

ciated with these one-half-red grains arise from cells in

which only one of the V factors has been changed to S,

one fourth of the gametes should carry S and three

fourths should carry V,or one fourth S,onefourthV, and

one half —, depending upon whether the ears concerned

are homozygous or heterozygous for variegated pericarp.

Such grains from homozygous ears should, therefore,

whether selfed or crossed by white, yield about one red

ear to three variegated ones. Similarly, from hetero-

zy gous ears, grains with one half their pericarp red should

106 THE AMERICAN NATURALIST [Vow. XLVIII

yield about one red to two variegated to one white if self-

pollinated and one red to one variegated to two white if

crossed by white. (This is on the assumption that no S

factors are carried by the male gametes.) Let us assume

that by lumping together all the seeds listed in the fore-

going records as ‘‘more than one half red” and as ‘‘less

than one half red’’ the whole lot would average about one

half red, and compare the results with the expectation as

noted above. From grains of these two classes from

homozygous ears both selfed and crossed by white, there

resulted 19 red-eared and 82 variegated-eared plants, or

a ratio of about 1:4.3 instead of 1:3. From heterozy-

gous ears self-pollinated grains of these two. classes

yielded 54 red-eared, 85 variegated-eared, and 52 white-

eared plants, and similar grains crossed by white yielded

7 red-eared, 7 variegated-eared, and 20 white-eared plants,

or ratios of 1.04:1.63:1 and 1:1:2.86 instead of 1:2:1

and 1:1:2, respectively. The observed ratios are cer-

tainly suggestive but must not be given undue importance,

for there is no assurance that the seeds used really aver-

aged one half red and no assurance that some of the male

gametes in the case of the selfed seeds did not carry S.

We must now examine the results secured in genera-

tions later than F, and note whether the hypothesis under

consideration applies equally well to them.

It will be recalled that F, red-eared plants that arose

from homozygous, variegated ears which had been self-

pollinated (see page 99) yielded in F, only red-eared and

variegated-eared progeny. On our assumption the for-

mula of the parent variegated ears was VV, but the red

grains of these ears were VS and the gametes associated

with them therefore either V or S or all S. Female

gametes carrying S would have produced red ears in F,

whether the male gametes carried S or V, and female

gametes with V could not have produced red ears except

when the male gametes uniting with them carried S. The

F, red-eared plants must therefore have been VS or SS,

the former being expected much more frequently than the

No. 566] INHERITANCE IN EARS OF MAIZE 107

latter, owing to the rarity of S in male gametes. Only 7

such red ears were tested and all yielded red and varie-

gated ears in typical Mendelian ratios, showing that all

of them were VS like any F, hybrid between red and

variegated races. Of two F, reds from selfed F,’s, one

again yielded reds and variegates and one apparently

bred true red. Three F, reds, from F, reds crossed by

whites, yielded reds and whites only—typical Mendelian

results throughout.

When F, red-eared plants arose from either homozy-

gous or heterozygous, variegated ears that had been cross-

pollinated by whites they yielded only red-eared and

white-eared, never variegated-eared, offspring (see page

100), just as if they were F, ears of a cross of reds with

whites. By hypothesis the parent variegated-eared plants

were V— and VV, and their red grains S— and SV (or

possibly SS). The gametes associated with such grains

were therefore S and —, and S and V (or possibly all 8).

The male gametes from white races were all —. The F,

plants were therefore S—, V—, and ——, only those with

S— having red ears. The five red-eared F, plants that

were tested produced in F, red-eared and white-eared

plants in Mendelian ratios. Of the F, red-eared plants

one bred true in F and three again segregated into reds

and whites.

When heterozygous, variegated, parent ears were self-

pollinated, the F, red-eared plants behaved in some cases

like hybrids of red with variegated races and in other

cases like hybrids of red with white races (see page 99).

Our assumption is that the variegated-eared parent plants

were V— and their red grains S—. The gametes asso-

ciated with these red grains were of course S and —. ‘The

male gametes of the same plants were doubtless largely

V and —, though a few were probably S. The F, plants

must therefore have been ——, V—, S—, SV or SS. Reds

with SS would be expected only rarely, and of the 11 F,

reds tested none had that formula, else they would have

bred true in F,. Seven of the 11 F, reds evidently were

108 THE AMERICAN NATURALIST [Vou. XLVIII

S—, for they yielded F, progenies consisting of reds and

whites only. Four of the 11 were obviously SV, for they

yielded F,’s of reds and variegates only. Of the latter

F, reds, one bred true in F, and four again segregated

into reds and variegates.

From a self-red seed of a homozygous, variegated ear

that had been cross-pollinated by a pure red race, an F,

red-eared plant was produced and this plant bred true red

in F,. From a nearly self-red seed of the same varie-

imh parent ear, an F, red was produced but yielded

reds and variegates in F, just as did a similar F, ear

from a seed with narrow red stripes (see page 100). The

variegated parent ear was VV and the red and near-red

grains probably VS. The gametes associated with these

grains were V and S. The male gametes were all S.

Therefore the F, reds were in part VS and in part SS.

By way of summary, it is recalled that, in all, 28 F, red-

eared plants were tested by F, progenies. Only one of

these bred true and that one came from a red grain of an

ear that had been cross-pollinated by a pure red race.

Disregarding the three F, red-eared plants thus produced

and the 9 red ears produced from seeds of variegated ears

that had been cross-pollinated by white races and that

therefore could not have bred true, there remain 16 F,

reds, none of which bred true in F,. Had these F, red-

eared plants behaved as did the F, green-leaved plants

produced by green branches of variegated-leaved parents

in Correns’s experiments, approximately 5 of the 16

should have bred true. It will be recalled that Correns

found that such green branches always produced green-

leaved and variegated-leaved plants in the ratio of 3:1,

and that one of the three bred true and the other two

again segregated, just as must have happened if the green

branch had been a part of an F, hybrid of green with

variegated instead of a part of a homozygous variegated

plant.

The difference between Zea and Mirabilis is, however,

not a fundamental one, but is due merely to the circum-

No. 566] INHERITANCE IN EARS OF MAIZE 109

stance that Mirabilis has perfect flowers while Zea is

monecious. In Mirabilis both male and female gametes

of a green branch arise from somatic cells in which the V

factor has changed toa G factor. If a change in only one

V factor is responsible for the production of the green

branch, the somatic cells of such a branch must all be VG

and the results reported by Correns are the only ones to

be expected. With Zea mays, however, all the grains of

one ear of a variegated-eared plant might arise from cells

having VS, so that half of the female gametes would carry

S, while little or no corresponding change might take

place in the staminate inflorescence and therefore no (or

very few) male gametes would carry S. From such an

ear of maize only about one half, instead of three fourths,

of the F, plants should have red ears and none (or very

few), instead of one third, of the F, plants should breed

true.

The occasional green plants (‘‘a’’ per cent.) arising

from variegated branches in Correns’s experiments with

Mirabilis are more nearly comparable to F, red-eared

maize plants than are the green plants arising from green

branches. It is quite conceivable that on a variegated

branch the male gametes might arise from cells that are

VG, while the female gametes arise from cells that are

VV, or the reverse, though this difference between male

and female gametes would hardly be so common an occur-

rence as with maize where the staminate and pistillate in-

florescences are situated so far apart. It is worthy of

note in this connection that of the occasional green plants

produced by selfed seed of variegated plants in Correns’s

experiments with Mirabilis (see diagram, Fig. 2), less

than one third bred true and more than two thirds segre-

gated into green and variegated. (Correns indicates this

merely by the signs < and > in connection with 33 per

cent. and 66 per cent. respectively, in his diagram, and

gives no indication of how much less than 33 per cent.

bred true or how much more than 66 per cent. segregated.)

De Vries’s results with Antirrhinum yield readily to

110 THE AMERICAN NATURALIST [Vou. XLVIII

the same analysis used with Zea and Mirabilis. Selfed

seed from striped-flowered branches gave a small per

cent.—from 2 to 10—of red-flowered plants. Only a few

of the red-flowered plants were tested and these were

found to yield 76 per cent. red to 24 per cent. striped.

Selfed seed from red-flowered branches of striped-flow-

ered plants yielded 71 per cent. red-flowered and 29 per

cent. striped-flowered plants, approximating the 75 per

cent. and 25 per cent. indicated by Correns’s results with

Mirabilis. None of these red-flowered plants bred true,

but only one test, and that of only a few plants, was made.

The results were 84 per cent. red-flowered and 16 per cent.

striped-flowered plants. It seems quite likely that had

de Vries tested more red-flowered plants he would have

` found some of them to breed true.

Correns’s results with striped and red flowers of Mirab-

ilis differed in one important respect from his results

with variegated and green plants of the same species, as

well as from the principal results with Zea reported here

and from de Vries’s results with striped-flowered and red-

flowered forms of Antirrhinum. When _ red-flowered

plants arose from striped-flowered varieties of Mirabilis,

they behaved just as did the green plants that arose from

variegated forms. But selfed seeds from wholly red-

flowered branches of otherwise striped-flowered plants

yielded little if any larger percentages of red-flowered

plants than did selfed seeds from striped-flowered

branches of the same plants. It would seem that in case

of Mirabilis flowers, when the self pattern arises as ‘a

somatic variation from the variegated pattern there is no

corresponding change in the Mendelian factors for these

patterns. In case of seed-sports from variegated-flow-

ered to red-flowered plants, however, the factors for vari-

egation are affected just as in case of green plants arising

from variegated ones and of red-eared maize plants aris-

ing from variegated-eared ones. The apparently non-

inherited somatic variations of maize plants, noted briefly

earlier in this paper, are possibly of the same nature as

No. 566] INHERITANCE IN EARS OF MAIZE 111

the somatic variations in variegated flowers of Mirabilis.

Some of these variations in maize are self-red cob patches

on otherwise variegated cobs, and dark, variegated grains

occurring in patches or scattered over light, variegated

ears.

GENERAL CONSIDERATIONS

The experiments of de Vries, Correns, Hartley, and

East and Hayes, as well as the records reported in this

paper, all indicate that certain somatic variations are in-

herited in strictly Mendelian fashion. All these somatic

variations consist in the appearance of self-colors on

plants that are normally variegated in pattern. The fact

that variegated plants occasionally throw both bud-sports

and seed-sports with self-colors is not, in general, to be

taken as an indication that the variegated plants in ques-

tion are heterozygous. Such behavior seems to be insep-

arably associated with variegation. Correns has pointed

out (loc. cit.) that variegated Mirabilis plants can not be

considered mosaics of green and ‘‘chlorina’’ types due to

heterozygosis, since they do not segregate into chlorina

and green, but into variegated and green. The same rea-

soning applies to variegation in the color of maize ears.

Variegated-eared plants do not throw reds and whites, but

reds and variegates. The conclusion seems irresistible

that self-color occurring as a somatic variation is due to

the change of a Mendelian factor for variegation into a

factor for self-color. If this be granted, the behavior of

these variations in later generations is a mere matter of

simple Mendelian inheritance.

From the title of his paper and the tone of his discus-

Sion, it is clear that Correns regards, as the most signifi-

cant feature of these inherited somatic variations, the

change from a homozygous to a heterozygous condition.

He even refers to them as cases of ‘‘vegetativen Bastar-

dierung”’ or ‘‘autohybridization.’’ To me, however, the

essential feature is the change of one Mendelian factor

to another. The fact that this modification of genetic

factors results in a change from homozygosis to heterozy-

112 THE AMERICAN NATURALIST [Vow. XLVIII

gosis seems wholly incidental. It follows from the circum-

stance that usually only one of the two V factors of so-

matic cells is modified. My own data do not in fact show

that the change always affects only one of the factors at a

time. While the results prove that this is true in a part

of the cases at least, the F, ratios suggest the possibility

of both factors being modified in some cases.

It is of course utterly impossible at the present time to

conceive of the cause or even of the nature of this change

in factors from V to S. We can only conjecture at pres-

ent as to whether the change may possibly be associated

with changing metabolic processes in the maturing plant,

or perhaps be connected in some way with changing ex-

ternal influences, or even be a quality inherent in the V

factor itself. It is perhaps significant that in maize, at

least, the change, whatever its cause, occurs very rarely

early in the life of the plant and apparently becomes in-

creasingly more frequent as the plant matures. Wholly

red ears in variegated-eared plants are extremely rare;

large patches of red grains are somewhat less rare; indi-

vidual red grains occur on most variegated ears; red

stripes on the individual grains are very frequent, in fact

all but universal in some strains, though in other strains

—very light variegated ones—there may be only a few

striped grains on a whole ear, the others being wholly

colorless. As a matter of fact, even the presence of an

ear with red pericarp throughout on a variegated-eared

plant may not be good evidence that the change in factors

occurred before the ear began to form. If the change

took place before the ear was laid down, it would seem

that the cob should always be self-red, since the red-eared

progeny of such modified grains of the variegated parent

plant invariably have red cobs, and cob and pericarp

colors are coupled absolutely in later generations. But

red ears, or nearly red ears, with light variegated instead

of red cobs, have been found to occur as somatic variations

on variegated-eared plants. Such behavior suggests that

sometimes the factor change may occur almost simul-

No. 566] INHERITANCE IN EARS OF MAIZE 113

taneously in the rudiments of every grain so that the

grains become self-red while the cob remains variegated.

We might, of course, account for the appearance of self-

colored grains on a variegated cob on the basis of sepa-

rate factors for cob and pericarp color!® by the assump-

tion that one of these factors may be modified while the

other remains unchanged. But we should then have the

no less difficult problem of accounting for the universal

appearance of red cobs with F, red ears without respect

to whether the parent grains stood on red or variegated

cops.

Forced to its logical limit, our conception of the V fac:

tor is that of a sort of temporary inhibitor, an inhibitor

that sooner or later loses its power to inhibit color devel-

opment, a power that once lost is ordinarily never re-

gained. Ofcourse it may be that there is present in varie-

gated maize merely a dominant factor for self-color, S, that

is temporarily inactive, but that sooner or later becomes

permanently active. Even if this be true, S as an active

factor and 9 as an inactive factor are certainly as distinct

in inheritance as they are in development and therefore

deserve to be designated separately. And since in one

case there results self-color and in the other variegation,

the factors may as well be called S and V as anything else.

It is of course also conceivable that the S factor may re-

peatedly arise de novo, though this seems very unlikely.

Whatever our conception of the nature of the factors

for variegation and for self-color in maize ears, these

factors are certainly as distinct in inheritance as any two

factors could well be. Moreover, there is abundant evi-

dence, which can not be given here, that they are strictly

allelomorphie, as indeed they must necessarily be if one

arises by modification of the other—this on the assump-

tion that the factors are definitely localized in certain

10 Evidence that there are distinct factors for cob and pericarp color was

Presented in a previous paper on coupling and allelomorphism in maize.

Ann. Rpt. Nebr. Agr. Expt. Sta., 24: 59-90. 1911

11 This problem is discussed in another paper on the simultaneous modifi-

cation of distinct Mendelian factors. AMER. NAT., 47: 633-636. 1913.

114 THE AMERICAN NATURALIST [Vov. XLVIII

chromosomes. Furthermore, these factors are to be re-

garded as pattern factors. Though they must influence

the development of the pigment in order to produce a pat-

tern at all, they are now known to be distinct in inherit-

ance from the factors for pigment—a fact that I have

been able to show by use of a race of maize with a peculiar

brown pericarp in addition to races with red pericarp.

SuMMARY

A somatic variation in maize is shown to be inherited in

simple Mendelian fashion. The variation has to do with

the development of a dark red pigment (or in one stock

a brown pigment) in the pericarp of the grains, often

associated with the development of an apparently similar

pigment in the cob and husks.

Plants in which this pigment has a variegated pattern

may show any amount of red pericarp, including wholly

self-red ears, large or small patches of self-red grains,

scattered self-red grains, grains with a single stripe of

red covering from perhaps nine tenths to one tenth of the

surface, grains with several prominent stripes and those

with a single minute streak, ears with most of the grains

prominently striped and ears that are non-colored except

for a single partly colored grain, and probably also plants

with wholly self-red and others with wholly colorless ears.

It is shown that the amount of pigment developed in the

pericarp of variegated seeds bears a definite relation to

the development of color in the progeny of such seeds.

This relation is not such that seeds showing say nine

tenths, one half, or one tenth red will produce or even tend

to produce plants whose ears as a whole or whose indi-

vidual grains are, respectively, nine tenths, one half, or

one tenth red. Experimental results indicate rather that

the more color in the pericarp of the seeds planted the

more likely are they to produce plants with wholly self-

red ears, and, correspondingly, the less likely to yield

plants with variegated ears.

Self-red ears thus produced are shown to behave in in-

No. 566] INHERITANCE IN EARS OF MAIZE 115

heritance just as if they were hybrids between self-red

and variegated races or between self-red and non-red

races, the behavior in any given case depending upon

whether the parent variegated ears were homozygous or

heterozygous for variegated pericarp and whether they

were self-pollinated or crossed with white.

It is suggested that these results may be interpreted by

the assumption that a genetic factor for variegation, V,

is changed to a self-color factor, S, in a somatic cell. All

pericarp cells directly descended from this modified cell

will, it is assumed, develop color, and of the gametes aris-

ing from such modified cells one half will carry the S

factor and one half the V factor if only one of the two V

factors of the somatic cells is changed, or all such gametes

will carry S if both V factors are changed.

The V factor is thought of as a sort of temporary, re-

cessive inhibitor that sooner or later permanently loses

its power to inhibit color development, becoming thereby

an S factor. Or it may be that the dominant factor, S,

is temporarily inactive, but sooner or later becomes per-

manently active. Again, the S factor may repeatedly

arise de novo. The cause of any such change in factors

is beyond intelligent discussion at present.

The results of Correns with Mirabilis and of de Vries

with Antirrhinum are shown to be subject to the same

analysis as that used to interpret the results secured with

maize,

~ AL

W N YN 3

{fe CELTE fiú

EP malt

l. "uye

Wihaareiitnn ih

‘adoD 4abwnwo sninpsoydypg” 10 NOILVUOLSUWY

- —, Pagik < y =

-2 Sinua r Dep aps 7 fm eye

ees a Sy ae h Dea

AN AA zi

a Se A $ :

Wii “Yi ijt pt yHyhn Wik rafy

WIH PI f R ge

if, / H Y f XS g `

IIs Mee SP SELE Y

| tol Wie yes SN

1 s n. ITN

` IIN

MEA x N

(ue ; a rA N :

\* aa ae `

\\ Me

\ yar

MA i `

soy

Zz i ay ee oe

Y % r 4

RESTORATION OF EDAPHOSAURUS CRUCIGER

COPE

Proressor E. C. CASE

UNIVERSITY OF MICHIGAN

In the year 1882 Cope described from the Permian beds

of Texas, an imperfect reptilian skull which he called

Edaphosaurus pogonias. Two years later he described

for the first time, the wonderful vertebra with elongate

spines bearing lateral projections on the sides. These

vertebr he assigned to the same genus as the skull but

later they were removed to a separate genus as he con-

sidered that the two specimens represented different

forms of reptilian life. The vertebre with long spines

and cross pieces were placed in the genus Naosaurus—

‘‘Ship-lizard,’’? a name suggested by the fancied resem-

blance of the spines with their lateral projections to the

masts and yard-arms of a full-rigged ship.

From the time of the original description until 1907 the

two genera were regarded as distinct but in that year

Case’ suggested that the two genera should be united and

that the skull described as Edaphosaurus by Cope be-

longed with the vertebral column and limb bones de-

Scribed under the name Naosaurus. The similar condi-

tion of elongate spines, but without cross pieces, on the

vertebre of the carnivorous genus Dimetrodon very nat-

urally led to the belief that the two forms Edaphosaurus

and Dimetrodon were similar in other parts of the body

and Naosaurus merely exhibited something of the extrav-

agance in spines, rugosities, tubercles, etc., which is such

a common feature in the most highly specialized members

of any group which is approaching the final stages of its

family or generic life. The close relationship of the two

genera was so probable that it was accepted by all paleon-

* Publication 55, Carnegie Institution of Washington.

117

118 THE AMERICAN NATURALIST [Vou. XLVIII

tologists and even Case was very reticent in his sugges-

tion that they were much farther apart than was usually

thought. Following the generally conceived idea of Nao-

saurus a composite mount was prepared in the American

Museum of Natural History in New York in which the

skull and limb bones of a Dimetrodon were associated

with the vertebral column of a Naosaurus. This restora-

tion was published by Dr. Osborn in the Bulletin of the

American Museum and a model of the creature in the

flesh was prepared under his direction by Mr. Chas.

Knight. Case in his ‘‘Revision of the Pelycosauria of

North America’’ republished this restoration by Osborn

but at the same time published an alternative restoration

in which the skull described as Edaphosaurus was asso-

ciated with the vertebral column of Naosawrus and the

two genera were united under the former name, as it had

priority.

The composite restoration prepared at the American

Museum has gained wide circulation in the text books but

later discoveries have shown that it was unfortunate. In

the summer of 1911 Dr. F. v. Huene, of Tübingen, while

a guest of the joint expedition from the universities of

Chicago and Michigan to the Permo-Carboniferous beds

of New Mexico, discovered the remains of a skeleton of

Edaphosaurus in which both the skull and a portion of

the vertebral column were preserved. As the vertebre

bore the typical cross-pieces of the genus Naosaurus the

identity of the two genera was established but new evi-

dence was speedily coming; Case in the summer of 1912

discovered in the Permo-Carboniferous beds of Archer

County, Texas, the nearly perfect vertebral column of an

Edaphosaurus (Naosaurus) cruciger Cope with the limb

bones, and a crushed skull, identical with the skull origin-

ally described as Edaphosaurus.

From this skeleton, now preserved in the museum of

the University of Michigan, the author has prepared the

restoration shown in Fig. 1. The only conjectural parts

are the size of the feet and the length of the tail; the re-

No. 566] EDAPHOSAURUS CRUCIGER 119

mainder is based upon careful measurements from a

single specimen. So far from being a carnivorous, rap-

torial animal similar to Dimetrodon, Edaphosaurus was

harmless, molluscivorous or insectivorous with possibly

some ability to masticate vegetable matter. The edges

of the jaws were lined with sharp conical teeth and upon

the palate and the dentary bones were strong plates sup-

porting numerous blunt, conical teeth. The head in all

specimens recovered seems rather small for the size of

the body and in this is peculiar in the Permo-Carbonif-

erous reptilian fauna, in which the reverse is the rule.

The shape of the head in the restoration is taken from the

nearly perfect and undistorted skull in the museum of the

University of Chicago. The elevated dorsal spines begin

with the third vertebre and speedily reach a considerable

height. The lateral projections are elongate at the base

of the spine but above the middle are reduced to mere

nodules irregularly arranged. The author is not in ac-

cord with the suggestion made by Jaekel and Abel that

the spines were separate, and can see no reason for the

suggestion made by the former that the spines were mov-

able. The strongly interlocking zygapophyses render such

an idea impossible to any one familiar with the skeleton.

Nor does the author believe that the spines were of any

use to the creature as offensive or defensive weapons;

rather, as he has frequently expressed himself, he believes

that they were in the nature of excessive growths which

may have had their inception and impetus in some useful

function, but grew beyond that use as the animal became

more specialized. The union of the spines into a thin

dorsal fin is far more probable and the idea is supported

by the presence of rugosities and the channels of small

nutrient vessels such as would lie beneath a thick dermal

covering. The anterior and posterior faces of the bases

of the spines have sharp, low ridges which give place to

Shallow grooves farther up the spine; only near the top

are the spines similar on all sides. Moreover in the liv-

ing genus Basiliscus, which has elevated dorsal spines,

120 THE AMERICAN NATURALIST [Vou. XLVIII

and in the genera of the chameleons in which the same

thing occurs, for example, Chameleo cristatus Stutch., the

spines are united into a thin dorsal crest by the integu-

ment and are further united by a thin membrane carry-

ing scattered muscle fibers. The outline of the dorsal

fin shown in the restoration is suggested by all the speci-

mens in which the spines have been preserved. The sharp

recurvature of the spines in the lumbar region is less

pronounced in the specimen from which the restoration

was drawn than in some other and it is possible that in

other species there was even more of an overhang of the

posterior end. The spines are abruptly shortened in the

pelvic region and rapidly decrease on the tail. The length

of the tail is not known but in all probability was elongate

rather than short and stumpy.

The limbs were short and heavy with the forearm and

foreleg shorter than the proximal segment of the limb, a

condition which is quite common in slow moving forms or

those of aquatic or palustrial habit, and just the reverse

of the condition found in the active, raptorial Dimetro-

don. The bones of the feet have not been found in posi-

tion, but in the great Brier Creek Bone-bed in Archer

County, Texas, excavated by an expedition from the Uni-

versity of Michigan in the summer of 1913, numerous

large foot bones of a character different from those of

Dimetrodon or the cotylosaur Diadectes were found as-

sociated with the spines of Edaphosaurus and with large

claws. It is believed that the foot of that animal was of

goodly size and armed with sharp claws well fitted for

digging in the soft earth or vegetation, tearing open rot-

ten logs and overturning rocks in search of food.

It has been noted by all collectors in the Texas beds that

isolated vertebre of Edaphosaurus are among the most

common fossils found but that any portion of an asso-

ciated skeleton is extremely rare. This has led to the

suggestion that the remains of the animals were trans-

ported for some distance after death, probably by rivers

from a higher land.

No. 566] EDAPHOSAURUS CRUCIGER 121

Edaphosaurus was a highly specialized creature, slug-

gish in movement and entirely harmless, living upon mol-

luses, insects and perhaps vegetation. It probably lived

in the woods or near swamps at some distance from the

lowlands upon which were deposited the deltas which

make up the Wichita and Clear Fork formations.

In conclusion the author wishes to express his thanks to

Dr. Ruthven, of the University of Michigan, for many

valuable suggestions in arranging the pose and propor-

tions of the restoration, and to Mr. Irwin Christman, of

the American Museum, for the painstaking care with

which his suggestions have been followed in making the

drawing.?

2 A full account of the known specimens of Edaphosaurus and Naosaurus

and a complete synonymy of the two genera will be found in Publications

55 and 181 of the Carnegie Institution of Washington.

SHORTER ARTICLES AND DISCUSSION

HUMIDITY—A NEGLECTED FACTOR IN ENVIRON-

MENTAL WORK

AN admittedly rough but probably fair estimate of the relative

interest which has been taken in the relation of the various

environmental factors to insects, at least, may be made from the

fact that Bachmetjew in his admirable compilation? of the work

along these lines devotes, in round numbers, four hundred pages

to temperature, one hudred and fifty to food and chemicals,

seventy to light, forty-five to humidity, fifteen to electricity and

magnetism and thirty to mechanical and other factors. Why is

it that temperature is given about a third more attention than all

the other factors put together? Is it true that it is nearly ten

times as interesting or important as humidity ?

A partial answer to the first question undoubtedly is that tem-

perature is easily controlled as well as measured, whereas humid-

ity, for example, is not easily controlled and the means of

measuring humidity in small containers are untrustworthy and

expensive. Furthermore, work with temperature gives results.

The unfortunate part is that these results have usually been as-

cribed wholly to temperature.

In the course of some work at the Carnegie Station for Ex-

perimental Evolution I found that I could change to a surprising

extent the markings on the larve of a moth (Isia isabella) by @

varying the temperature at which they fed and moulted. How-

ever, such changes were much more definite when the tempera-

ture was kept constant and humidity varied. I did not have the

necessary apparatus for getting accurate control of either factor,

but I feel confident that temperature had little or no direct in-

fluence. It was acting through its influence upon humidity.

It would seem unnecessary to urge upon experimenters such a

fundamental principle in the logic of cause and effect, but the

fact is that with only two or three exceptions none of the more

than a hundred papers having to do with the effect of tempera-

ture upon insects tell us anything about the effect of temperature

1 ‘‘ Experimentelle Entomologische Studien vom physikalisch-chemischen

Standpunkt aus.’’ Zweiter Band. Sophia, 1907.

122

No.566] SHORTER ARTICLES AND DISCUSSION 123

per se. A few state that the atmosphere was ‘‘moist’’ or ‘‘dry,’’

but even then how moist or how dry is not usually mentioned

unless it is believed to be saturated or absolutely free from moist-

ure. It is clearly incumbent upon the one who makes such a criti-

cism to show, either by his own work or in a review of that of

others, that humidity is a factor of such importance that the

criticism is worth the making—especially since the point is so

self-evident and has been made in the past. The following notes

are an attempt to justify the preceding.

The experiments of many workers show that when lepidop-

terous pup are subjected to abnormal temperature part, at

least, of the adults which emerge differ from the normal. The

observations have usually been made on color changes, and

Fisher? especially has shown that warm conditions (36° to 41° C.)

produce the same or similar effects as do cold conditions (0° to

10° C.), also that hot conditions (42° to 46° C.) produce effects

which are similar to those produced by freezing (—20° to 0° C.).

Fisher apparently had no means of successfully controlling the

humidity but Tower? claims to have had this in his ‘‘ Investiga-

tion of Evolution in Chrysomelid Beetles of the Genus Leptino-

tarsa’’ and he obtained similar results, stating them as follows:

The result produced by either a higher or a lower temperature is the

development of a greater amount of pigmentation and a consequent me-

lanie tendency in variations. This stimulus in both directions to inereased

pigmentation reaches a maximum between 5° and 7° C. deviation from

normal. Beyond these, as the temperature further deviates, there is

a rapid fall in melanism, first to the normal, and then to a condition

below normal, until a marked albinie tendency is found; and this de-

crease in pigmentation continues until the zero point is reached, be-

yond which no pigment whatever is produced. The zero point is

reached much sooner, however, in high-temperature experiments than

in low.

Tower then gives the results of experiments in which all the

environmental conditions, except humidity, are ‘‘normal.’’

Normal humidity for Leptinotarsa decemlineata is taken as rang-

ing from 43 per cent. to saturation with an average of 74 per cent.

The humidity in various experiments ranges from 10 per cent. to

Saturation. The lowest natural humidity of which I have seen

a record is 5 per cent. It occurred in Death Valley, California,

*See Archiv fiir Rassen- und pense ST: -Biologie, 1907, IV, pp. 761-

793, for Fisher’s statement concerning criticisms of his co nelusi

x Carnegie Institution of Washington, Pon aala No. 48, 1906.

124 THE AMERICAN NATURALIST [Vow XLVIII

where the monthly means for May to September inclusive varied

from 20 per cent. to 27 per cent. The annual mean at Cairo,

Egypt, is 56 per cent. and at Ghardaia (Algerian Sahara) is

50 per cent. at 7 A.M. and 26 per cent. at 1 p.m. The humidity

at Buitenzorg, Java, during the height of the rainy season fluc-

tuates between 70 per cent. and 97 per cent. during the day.

Naturally, when dew is being deposited the humidity is practi-

cally 100 per cent. It will be seen then that even Tower’s ex-

treme averages (see below) are not beyond the range of

possibility in nature, although they are as great as it is possible

to use in experimental work, since at an average of 34 per cent.

humidity only 0.4 per cent. of the larve reached the adult stage

and atmosphere can not be kept supersaturated.

The beetles were seriated according to an arbitrary scale in

which ‘‘20 equals total melanism and 0 total albinism.’’ It is

difficult to suggest a better method of measuring the extent of

melanism than this, although we could wish for diagrams to aid

us in grasping just what the scale means. I have tabulated the

experiments and interpolated the normal data.

Relative Humidity Per Cent. of Melanism

Average Range Morjaiiiy Mode Range

100 100-100 90 4 9

95 82-100 30 7 3-11

84 55-100 15 12 7-16

74 43-100 ? 9 5-13

66 33-100 35 ll 6-

60 30-100 80 5 3-11

50 25-83 92 3 i

34 10-55 99.6 2 1—4.

It will be seen that mortality increases rapidly as the humidity

departs from normal but this can not account for the change

in color since the range of melanism is doubled and in three of

the experiments even the mode falls below the normal range. AS

stated by the author:

The results of experiments with deviations of humidity are almost

exactly the same as those which were obtained from experiments with

deviations of temperature. Such deviations from the normal either to-

ward an increase or a decrease, produce up to a maximum increased

pigmentation and a consequent melanie tendency, but beyond this the

effect is reversed, pigmentation is retarded, and the tendency toward

albinism becomes more and more pronounced as the deviation from the

normal becomes greater.

No. 566] bre car eon CRUCIGER™ 125

Alari on

The point which concerns aa cm discussion is that not

only does humidity have a definite regularly acting influence, but

that its results are similar to those of temperature and, as with

temperature, plus and minus variations of certain intensities

bring about similar effects. If, as has usually happened, the hu-

midity is not controlled in experimental work on the effect of

temperature, how can it be said that the observed results are the

effect of changes in temperature ?

Tower made certain experiments in which both temperature

and humidity are abnormal, normal average temperature being

taken as 22.2° ©. Unfortunately, proof reading or something of

the sort was faulty when it came to publication. Experiment

26 would be the most valuable for our present purpose, but the

table includes records of relative humidity 35 and 39 per cent.

above normal, 7. e., relative humidities of 109 and 113 per cent.,

respectively, if, as in the other experiments, 74 per cent. is

‘normal’? humidity. These are clearly impossible. The text

figure illustrating this experiment does not help us since hu-

midities are not given and furthermore the temperatures in the

figure are rather consistently one degree different from those

given in the table. Since there are two errors in text-figure 15,

which illustrates the experiments with humidity as the only vari-

able, it is likely that the figure is the thing that is at fault here.

Several other similar discrepancies could be pointed out (as, for

example, the temperatures in experiment 24, which concerns the

combination effect of humidity and temperature) but it is prob-

able that the author’s notebook records are correct and the tem-

perature discrepancies in the published report are so slight that

we may accept his conclusion. It is

that when temperature and moisture are the variables in a given en-

vironmental complex, the trend of general color modification is con-

trolled by moisture (relative humidity), excepting in conditions where

the temperature deviation is so excessive that the ordinary physiological

and developmental processes are greatly inhibited. In experiments

approximating natural environmental complexes, however, moisture is

the dominant factor in influencing coloration.

Even if there were no other reasons for urging the necessity

of taking humidity into account, I feel that Tower’s work would

be ample justification. Before taking up those reasons let us

notice several cases where, on account of the striking results of

the experiments, we must regret our lack of information as to the

real cause or the relation of the several causes.

126 THE AMERICAN NATURALIST [Vou. XLVIII

This same work of Tower is one of them. The effects just noted

were merely ontogenetic. However, he made other experiments

in which the effect seemed to be passed on by heredity. The fac-

tors in the various experiments with L. decemlineata were 35°,

45 per cent. and low atmosphere pressure (p. 287) ; ‘‘hot, dry”?

(p. 288); ‘‘hot, dry and low pressure’’ (p. 288); and ‘‘hot,

moist’’ (p. 291), probably 31.2°, 94 per cent. Those with L.

muliteniata were 30° and saturation (p. 292 and p. 293) ; and

the one with L. undecemlineata was ‘‘10 C. above the average and

a relative humidity of 40 per cent.’’ The work is of such im-

portance because of its pioneer character that it would be un-

gracious to complain too strongly, but the fact is that it is

impossible to tell from the data given whether the effects are

caused by humidity or by temperature or by a combination of

the two. Bateson’s idea that there are no effects to be explained

need not concern us here.

There is a long series of interesting papers starting in 1895

by Fischer. As has already been mentioned, he finds that certain

high temperature grades produce effects which are similar to

those produced by certain low temperature grades. The eon-

ditions of humidity are rarely mentioned, not to say considered.

However, he occasionally confesses that they are important, as

when he tells us* that it is necessary to have the warm air dry an

the cold air moist in order to get similar forms of Vanessa by the

application of moderate cold and moderate heat. I suspect hu-

midity largely enters into the other experiments also for in one

with high temperature,> which gave the same results as certain

low temperatures and presumably high humidity he says the hu-

midity was high.

Like Tower’s experiments with beetles these concern color

alone. Pictet® and Federley,” especially, have considered the

effect of environmental factors upon the form of lepidopterous

scales. Federley calls his work ‘‘Temperatur-experimente’’ and

Pictet ‘‘Influence de 1’Humidité’’ but neither enables us to dif-

ferentiate the effects of the two factors, although both obtained

striking results. Kominsky* modified to a considerable extent

4 Algemeine Zeitschrift fiir Entomologie, VIII, p. 274, 1903.

5 Illustrierte Zeitschrift fiir Entomologie, IV, p. 134, 1899.

6 Mémoires de la Société de Physique et d’Histoire Naturelle de Genève,

XXXV, Fase. 1, 1905.

7 Festschrift fiir Palmen, No. 16, Helsingfors, 1905.

8 Zool. Jahrbiicher. Abt. fiir Allg. Zool. und Physiologie, pp. 321-338,

1911.

No.566] SHORTER ARTICLES AND DISCUSSION 127

not only the color and form of scales but also the form of an-

tennæ, legs and other body parts of Lepidoptera. He exposed

the pupe to 42.5° C., humidity not given; 38° to 39° C. and 42°

to 43° C., relative humidity 80; 8° C., “high humidity’’; 0° C.,

“‘very high humidity’’; — 7.5 to 5° C., relative humidity 80-90

and 50; and —11° C., humidity not given. For the most part

the humidity was high and probably had much to do with the

results, but we can not be certain.

All the experiments just considered were made upon pupe.

It should be remembered that only about one fourth of the weight

of lepidopterous pupe consists of solids, and that the only way

they can replace fluids lost by evaporation is by chemical changes

in these solids. It is probable that they do so to some extent,

although this has not been accurately determined. It is known

that under normal conditions pupe lose in weight and the per-

centage of solids increases. Naturally, a change in the humidity

of the surrounding air would modify this physiological process

and it is difficult to believe that it has not quite as much effect as

changes in temperature, the humidity remaining the same. It

is easy to see that, if the air is made more absorptive or less ab-

sorptive either by the temperature changes themselves or by other

means, and then the physiological activities are slowed or quick-

ened by temperature changes, the effects will be much greater

and might easily pass as due entirely to the temperature changes.

The species which have wet and dry season forms in regions

where the temperature is fairly constant throughout the year, as

well as the tendency for the animals of moist regions to be mel-

anic and of arid regions to be light colored, speak for the impor-

tant influence of humidity. But there is another point in

distribution to be considered. The study of distribution was

long, and still is, largely an effort to get the ranges of animals

and plants to fit isotherms. When yearly averages do not work,

winter minima or summer maxima or accumulated temperatures

are tried. The success which often attends these efforts shows

that man is very ingenious and also that temperature is really

one of the controlling factors, but it does not show that it is the

only factor or, in fact, that it has any direct influence.

The areas of grassland and forest in North America cut across

isotherms as though they were merely political boundaries but

Transeau® has shown that if we plot the ratio of temperature to

°? AMER, Nat, XXXIX, pp. 875-889, 1905.

128 THE AMERICAN NATURALIST [Vow XLVIII

humidity we get a very close correspondence between distribution

and climatice factors. Schimper’? has brought together a great

deal of evidence which indicates that, as far as plants are con-

cerned, even the major divisions of the world’s surface into arctic,

temperate and tropical are fundamentally a question of the de-

mand for and supply of water.

Furthermore, if recent climatic changes have an effect upon

the origin of new characters and the distribution of the organisms

possessing certain characters, humidity is deserving of more

attention than temperature, since practically the only evidence

we have of such changes concerns humidity.

It should not be forgotten that even aquatic organisms are

subject to what amounts to changes in humidity. Peat bog plants

take on many characteristics of a desert flora, although their

roots are covered with water. It is water, however, which is

not easily available, because of the chemicals which it carries.

It is water which is physiologically dry.

Finally, the great amount of work which has been done upon

artificial parthenogenesis and related subjects is, in a way, a

study of the influence of environmental factors. The obvious

factors concerned have usually been various chemicals but at

foundation humidity, in a broad sense, the addition or withdrawal

of water by osmosis seems to be a factor of prime importance.

Frank E. Lutz

AMERICAN MUSEUM OF NATURAL History

10 ‘‘Plant Geography upon a Physiological Basis,’’ translated by W. R.

Fischer. Oxford, 1903.

VOL. XLVIII, NO. 567”

fup

MARCH, 1914

THE

AMERICAN

NAIURALISTI

A MONTHLY JOURNAL

Devoted to the Advancement of the Biological Sciences with

Special Reference to the Factors of Evolution

CONTENTS

age

The Effect of Extent of Distribution on Speciation. Asa C. CHANDLER - 129

- - i621

Biology of the Thysanoptera. Dr. A. FRANKLIN SHULL -

Shorter Articles and Correspondence: The Endemic Mammals of the British

Isles. ProfessorT.D.A.COCKERELL - - y

- Notes and Literature: Swingle on Variation in F; Citrus — and the

185

Theory of Zygotaxis. Dr. ORLAND E. WHITE

THE SCIENCE PRESS

LANCASTER, PA. GARRISON, R. ¥.

NEW YORE: SUB-STATION 84

Two Important Biological Journals

The Journal of Genetics

Edited by

W. BATESON and R. C. PUNNETT

A periodical for the publication of records of original research in her redity, variation, and allied topice.

The JOURNAL also contains, from time to time, articles + eoniaasis zing > existing state of knowledge in the

varions branches of genetics. e numbers are fully illustrated, and, aa the —— -matter demands it,

colored plates are fears employed. The JOURNAL is issued in parts’ ae terial a mulates, and a volume

(appeari ng, 80 tar as possible, anonkit A consists of four such parts. Bitederpélen price, $7.50 a volume;

single copies, $2.5

Biometrika

Edited by KARL PEARSON

A journal for the statistical study of biological problems, appearing fate times a year. A volume

se about 500 pages, with plates and tables, is issued annually. e X beginsin 1914. The

nine preceding volumes form a veritable encyclopaedia of biological sons Ble a nike fields of painos eo"

craniometry, medicine, heredity, etc. Subscription price, $7.50 a volume ; single copies, $2.5

Prices on back numbers of these journals vary, and postage from London is charged on back volume

and seats copies.

Published in America for the Cambridge University Press of England by

The University of Chicago Press

Chicago, Illinois

University Control

By J. MoKren Carrett, Professor of Psychology in Columbia University

Together with a serics of Two Hundred and Ninety-nine Unsigned Letters by Leading Men of Scien?

holding Academic Positions and Articles by Joseru Jastrow, Gores T. Lapp, Joun J. STEVENSON, >

Camaron, J. McKean CATTELL, Gmorom M. STRATTON, STEWART PATON, Joun Jay CHAPMAN, JAMES

Munnoz rr sao Gouin Boronu

These ASEM and examples are taken from Professor Cattell’s informed and thorough discussion oie

subject of pope control, a subject upon which À he has had much to say of late, finding occasion for ¢#

eriticism of existing “an conditions, and standing as the peyare of an academic a and a tea

ing profession upon which a man may enter without forfeiting his self-r —The Dial.

Sentences and paragraphs that serge the expert, highly-trained mind, the suggestions that coms to te

fresh a d tell us that a new day is abo ut to dawn in educatioual writing —The Boston Evening Transcript

SCIENCE AND EDUCATION

A series of volumes for the promotion of scientific research and educational prog” Ait

VOLUME I. The Foundations of Science. By H. Porveané. Containing the authorized English translati

by Gmoros Bruce Hausrep of “Science and Hypothesis,” “The Value of Science,” and “Scien

and Metho -00 net. E.

VOLUME II. Medical Research and Education. By Ricaazp M. Prance, Wruuam H. Werom, W- o

Howser., P. Maru, LewrLLys F. Banker, CaarLms S. Minor, W. B. T r?

Councttman, Tuzosatp Burra, G. N, Srawarr, C. M. Jackson, E. P. Lron, James B. HARRICE

M. Dopson, C. R. Barden, W. Orntzs, S. J. Meurzer, Jauns Ewrwa, W. W. Keen, Haner H.

som, CanmwrTIAN A. Herter, and Hanny P. Bowprros, Price, $3.00 net.

VOLUME III. University Control. Now Ready. Pages 2+484. Price, $8.00 net.

GARRISON, N. Y. THE SCIENCE PRESS LANCASTER, PA:

SUB-STATION 84, NEW YORK CITY

THE

AMERICAN NATURALIST

Vou. XLVIII March, 1914 No. 567

THE EFFECT OF EXTENT OF DISTRIBUTION

ON SPECIATION

ASA C. CHANDLER

CONTENTS

hac paea UE CLE Te ee et ae

eneral Statement of Hypothesis.............eccseeecececceees 130

2. Tests of Theory by Comparison of Families............--++++ee0e 131

gle IR ee POET ME hoe, Oe" appr ran? 131

ON ida: vos wes sts dpa pa dad el Ghh E sek 133

3. Test of Theory by Comparison of Faunas of AreaS.............+++ 135

OBAE ogie caes ia epee eee 135

es oe eek as iiss T N 141

peel ak Airy Wii os i ose S «ee oN vide Vi ee ween Ones 143

ME i od ig ed nu dcp a a 145

Amphipoda (Marine Gammaridea)....... 147

4. Considerations Preliminary to Theoretical Explanation 148

sg sine SSE ees a A 148

Pee MN i CSU A Cee a a aa, 149

ifferentiation of Genera vs. Differentiation of Bpecies. .< 5 «seas 151

5. Theoretical Explanation of Hypothesis............ ....... 154

I. ao aa a A ... 156

en A O a a E 158

Waite engaged in some research work on the geo-

graphie distribution of mammals under the supervision

of Professor H. D. Reed at Cornell University in the fall

of 1910 and the spring of 1911, certain conceptions re-

garding the relation between extent of distribution and

the generic and specific modifications of mammals were

brought to light. Due to the valuable and helpful criti-

cism of Professors C. A. Kofoid and J. C. Merriam, and

Dr. J. Grinnell, and of other members of the University

of California, and to the advice and aid of Professor

C. A. Kofoid, the rather vague ideas then formed have

been worked over and crystallized into their form as

Presented in this paper. _

129

130 THE AMERICAN NATURALIST [VoL. XLVIII

In the past, much of the work that has been done on

zoogeography has dealt with a study of the facts of dis-

tribution, both present and past, as they stand, together

with a study of the factors influencing distribution and

speculations regarding the explanation of some of the

interesting and apparently anomalous facts thus brought

to light. In all of this work, the distribution of animals

has been considered almost entirely as the effect of cer-

tain biological and geological causes. The present paper

is intended to show that the distribution of animals is not

only the effect of other causes, but is in itself the cause

of other effects, and that extent of distribution has a

direct influence on the modification and speciation of the

group concerned.

To find out how far-reaching and how potent is this

effect, much further study is necessary, not only of the

distribution of various groups, but of their classification

and systematic relationships as well.

In brief, the effect of extent of distribution on groups

of different systematic rank may be stated as follows:

As the range of a group of animals, be it genus, family, or

order, is extended, the species increase out of proportion

to the genera, the genera out of proportion to the families,

and the families out of proportion to the orders. In

other words, if we assume that in a distributional area

of certain extent, there are three genera and six species,

in a distributional area of twice that size, there will not

be six genera and twelve species, but more probably only

four or five genera, and twelve species; i. e., if in the first

case the index of modification (a term here used to indi-

cate the average number of species per genus) be two, in

the second case it will be greater than two.

As new distributional areas are added, other factors

remaining equal, there is a constant increase in number

of species and subspecies, going hand in hand with a

diminishing rate of increase in genera, the result being a

constantly larger index of modification as the area in-

habited by a group of animals is extended.

No. 567] EFFECT OF DISTRIBUTION ON SPECIATION 131

It should be remarked that a unit of area in this con-

nection should be considered a distributional unit, not a

geographical unit. In other words, while the addition of

one hundred square miles might or might not involve a

change in the life of a region, the addition of a new ‘‘life

zone,’’ ‘‘fauna,’’ or association’? (see p. 155) would

inevitably involve a biotic change, and therefore the addi-

tion of one or several of any of these distributional areas

should be considered as an addition of a unit, comparable

to another unit of similar kind.

Two possible ways of testing this hypothesis present

themselves. We may compare the faunas of distribu-

tional areas of dissimilar size, or we may compare the

specific and generic differentiation found within families

occupying areas of different extent. The former method

we should expect to work out with a fair degree of

accuracy, but the latter involves so many modifying cir-

cumstances that even if sufficient data were at hand, it

would be difficult to prove anything by it. In the first

place there is the difficulty of comparing, in a distribu-

tional sense, the areas occupied by different families,

Since, as pointed out above, the geographic areas do not

necessarily coincide at all with distributional areas; in

the second place, while it is justifiable to compare the

speciation of a family in one region with the speciation of

the same family in another region, it is of doubtful value

to compare the speciation of one family with that of

another in the same or different regions, unless the other

factors controlling their speciation be comparable or

nearly so. In view of this there are few families which

could be advantageously compared with each other as to

Speciation in relation to extent of distribution, yet in the

families which do seem to lend themselves to such a com-

parison, the evidence all points towards the correctness

of the law here proposed.

The bats seem as favorable for such an interfamily

comparison as any group of mammals that could be

Selected, and the table (Table I) of their distribution by

132 THE AMERICAN NATURALIST [Vou. XLVII |

TABLE I

DISTRIBUTION AND SPECIATION OF FAMILIES OF CHIROPTERA

Data Derived from Sclater and Sclater (1899)

> | Ss | Index

Family Distribution | Gen | Sp. of Mod

V pa Se Pair eee E Aa T «oo bis va ee ues | 17 | 190 11.18

Emballonuride..... Warm parts of both hemispheres. ... . 15 79 5.27

Pletccddn. Shik ees Old Wordi eosl kk ES | 18 | 110 6.11

Rhinolophide...... Oe WOME Se ek ie bees eaves agers Us ea G 10.16

Nycteride......... Warm pate of Old World.......... Bee. p 7.50

Pil lipatormsdie. . . i NeDeranacel. 4 enar a FOE REE 8 | 2.25

families is significant. One family, the Vespertilionide,

is cosmopolitan, inhabiting every zoologic region an

every life zone, and it has 11.18 species per genus, the

highest of any family of bats. The Phyllostomidæ, on the

other hand, has the narrowest range, occupying only the

warm zones of one zoologic region, namely, the neotropic,

and has in 36 genera only 81 species, giving 2.25 as the

TABLE II

DISTRIBUTION AND SPECIATION OF FAMILIES OF INSECTIVORA

Data Derived from Sclater and Sclater (1899)

Family Distribution Gen. | Sp. of Mod.

Sorickde. oo os. ose. Palearctic, rte poche and

Nearctic regions, all zones........ 11 125 | 11.36

Erinaceidæ........ Palearctic, Ethiopian, ad Oriental

i cece eal EES eee eee 2 16 8.00

TA. 3. Ss ss Palearctic Nearctic regions, tem-

perate zo gag Oy Be neues re it 25 2.27

Wunaltde. . ooo oes tal ber mena gesa, Luwir 2 15 7.50

DOPE Ethiopian region, warm zones....... 3 17 5.66

Potamogalide..... C rear pots neg Sadana

ICR SONNE tc no 2 3 1.50

Galeopithecide..... Maley only, y, forests, tropical zones. 1 2 2.00

hrysochloride....|South Africa. icase ikonia 1 7 7.00

entetide: . E ES A T E E E EE N TL n 3.00

lenod cehs end Bara a 1 | 2 2.00

index of modification. The other figures in this table are

significant, but the indices of modification in the families

Rhinolophide and Nycteride are abnormally large, and

will probably be reduced by subsequent subdivision of

genera, or discovery of new forms.

able II shows the generic and specifie differentiation

No. 567] EFFECT OF DISTRIBUTION ON SPECIATION 123

of the various families of Insectivores, but as some of the

families have not been as intensively studied as others,

and as the conditions affecting their distribution and

speciation are so different in different families, we could

hardly expect accurate results, and yet the table clearly

shows a tendency for the families having wider ranges

to have a higher index of modification, the almost cosmo-

politan shrews, for instance, having 11.36 species per

genus, and the families with restricted range (Galeopi-

thecide, Solenodontide, Centetide and Potamogalide),

having only 1 to 3 species per genus. The Talpide and

Chrysochloride do not seem to conform in their speciation

to what should be expected.

When the specific and generic subdivisions of all the

families of mammals have been worked out more per-

fectly, and their ranges in a distributional sense, i. e.,

through life zones, faunas, and associations, are more

accurately known, some interesting facts concérning the

relation between their indices of modification, and the

extent of their ranges, might be brought out.

It is interesting to note that there is a considerable

number of conspicuous examples of wide-ranging genera

which are remarkably poor in species. Among carnivo-

rous mammals there are many such cases, these animals

seeming to be adaptable to an almost unlimited range of

environmental conditions without modification, or, in

other words, their germ plasm is not stimulated to change

by altered conditions of climate or environment. The.

tiger, for instance, is equally at home in the bleak frozen

steppes of Siberia, or in the hot humid jungles of India.

The genus Cynaelurus is widely distributed over the

Ethiopian and Oriental regions, and yet it contains but a

Single species, with several geographic races. Among

birds there are a number of similar examples, the most

striking case, perhaps, being Pandion, a cosmopolitan

genus with but a single species. The same peculiar condi-

tion occurs among lower animals, as for instance in the

Dinoflagellate genus Diplopsalis, which is cosmopolitan

134 THE AMERICAN NATURALIST [Vou. XLVIII

TABLE IIIA

SPECIATION OF MAMMALS IN VARIOUS DISTRIBUTIONAL AREAS IN CALIFORNIA

Data from Grinnell (19134A), (1908), Grinnell and Swarth (1913)

Boreal and Upper Transition Zones

San Jac. Mts. San Bern. Mts. Si R

cow | BUNS | Stas | Soma

Gen. Sp. Gen. | Sp. Gen. Sp.

LAE TS A A O EN E 1 1 1 | 1 2 4

DOVOD 6 60s lees 1 1

ORCI yo ois oe ov 1 1 1 1 1 3

Antilocapridez......

Rodents...» TA 8 10 12 21 57

a ec: 4 4 5 5 6 22

in eee

Aplodontide....... 1 1

Lit g 1 FAT VE 2 3 3 5 7 17

Geomyidez......... i 1 1 1 z 5

Saren AEE $ 1 1 2

n ee ee 1 2

Erethisontida eee 1 1

Ochotonide........ 1 3

ES UE 2 4

Carnivora......... 6 6 2 (7) 2 (8) 14 21

WOH... Coie hes 2 2 (2) ) 2 3

Canidae 6595's: 1 1 1 (2) 1 (2) 3 6

Mustelide......... 3 3 1 T 10

Procyonidæ........ (1) (1) 1 1

Uim: syne ces (1) (2) 1 1

Insectivora........ 2 2 2 2 4 11

POM se 1 1 1 1 2 q

pe n E P 1 1 aS 1 2 4

Cheiroptera........ 2 3 2 3 4 T

Phyllostomidæ.....

Vespertilionide..... 2 3 2 3 4 7

Molosside.........

COUR es acca 18 20 17 (22) | 20 (26) 45 100

Indices of modifica-

ON kes ces peers 1.11 1.17 (1.81) 2.22

in warm and temperate seas, and yet is composed of not

more than two species. No adequate explanation of these

exceptional cases has been offered, and it is probable that

their speciation, or lack of it, is due to conditions of their

existence or constitution which we do not understand, or

do not recognize.

No. 567] EFFECT OF DISTRIBUTION ON SPECIATION 135

To test the law by comparison of faunas of areas of

different extent, a series of tabular comparisons of the

faunas of various regions of different size and character

was made. In all of these tabulations, care has been taken

in the choice of areas for comparison to make them of un-

equal size from a distributional point of view, and to

make them reasonably comparable. An arctic and a

tropical region, for example, are not considered reason-

ably comparable as regards number of genera and

species, nor is a region on the outskirts of the range of a

group considered comparable with a region near its

center of distribution.

Table III shows a comparison of the mammals of vari-

ous parts of California. The regions compared are as

follows: (A) the boreal and transition zones of (a) the

San Jacinto Mountain range, (b) the San Bernardino

Mountain range, and (c) the entire Sierra range, includ-

ing the Warner and Shasta Mountains to the north, and

the San Bernardinos and San Jacintos to the south; (B)

a comparison of all the zones of (a) the San Jacinto

Mountains with the immediately adjoining country, (b)

the Sierra range as defined above, and including their

foothills, and (c) the entire state. |

A careful study of Table III brings out a number of

interesting and significant facts, and bears out the law

here proposed with unexpected accuracy, barring one

seeming exception which, as we shall see later, can not

truly be considered as such.

Let us compare first the three areas in which only the

two uppermost life zones are involved, and from which

the species invading only the lower Transition zone have

also been excluded. First, a word as to the areas com-

pared. The Boreal and Transition zones of the Sierras

take in over one half of all the representation of these

zones within the whole state. These zones of the San

Bernardino and San Jacinto mountain masses are, as

compared with the entire range, very small indeed, and

comprise almost as small areas as could justifiably be

136

THE AMERICAN NATURALIST

TABLE IIIB

(Data as above)

(Data as in Table IIIA)

[Vou. XLVIII

All Zones

San Jac. Mts. Si R Californi

on (2,500 Sq. M.) (60,000 Sq. M.) (158,000 Sq. M.)

Gen. | Sp. Gen. Sp. Gen, Sp.

Ungulata. -...0...- 2 2 3 7 4 10

Bovis... 6 eek 1 1 1 2 1 y

se NCEE AE 1 1 2 5 2 7

Antilocapride...... 1 1

Moecenue. ...is 25% 16 41 28 110 31 203

Betvide: anon 5 7 z 26 7 41

storim. so o 1 2

Aplodontide....... 1 1 1 2

vee eee ROE ae 5 14 10 33 11 64

eomyidæ......... 1 4 1 9 1 19

Héteromyidæ...... 3 12 4 24 4 48

TA TT OE T 1 2 1 5

Erethizontidæ...... 1 1 1 1

Ochotonide........ 1 3 1 3

Lepore. <i E 2 > 11 3 18

Carnivora. ...-.:.. 9 10 15 29 17 51

Fade. a aso 2 2 2 3 2 6

Cepia. ia sk. 3 4 3 9 3 17

Mustelide......... 3 3 i 13 9 22

Procyonide........ 1 1 2 3 2 4

DV Or ee, 1 1 1 2

Insectivora........ 3 5 12 6 20

WE ras oa. 2 2 3 ' 8 4 14

Tee. -<.3c 1 1 2 4 2 6

Cheiroptera........ 4 7 7 12 11 26

Phyllostomide..... to 1

Vespertilionide..... 4 T 6 11 8 21

Molossidæ......... 1 1 2 4

TOL. es Sees 34 63 58 170 68 310

Indices of modifi-

cation 1.85 2.93 4.56

considered to be individual faunal units. The San Jacin-

tos are somewhat smaller than the San Bernardinos, but

the difference is almost inconsiderable when compared

with the Sierras. Before examining the table, let us see

No. 567] EFFECT OF DISTRIBUTION ON SPECIATION 137

what conditions in number of genera and species would

be expected in these three areas. The San Bernardinos,

being almost as small a faunal unit as should be sepa-

rately considered, we should expect to approach a mini-

mum index of modification, i. e., a minimum number of

species. per genus, approaching one as a limit. On the

San Jacintos, these being smaller than the San Bernardi-

nos, we should expect fewer types according to the law

suggested by Grinnell and Swarth (1913), that the num-

ber of persistent types in a disconnected area varies

directly with the size of the area. On the entire Sierra

range we should expect, due to the greatly increased

territory, a considerable increase in genera, but a very

much greater increase in species. Looking now at Table

II, we find that with the single exception of the car-

nivores on the San Bernardino Mountains, not one dis-

crepancy exists. The Ungulates, Insectivores and bats

are represented by the same numbers of genera and

species on both of the small areas, and all of them show

a marked increase in genera and species on the larger

area, in every case with an increase in the index of

modification.

The rodents, which show a larger degree of differentia-

tion than any of the other groups, show a very interesting

advance in the index of modification as the area is ex-

tended. The carnivores, as stated above, show a seeming

discrepancy, inasmuch as there are six genera and six

Species existing on the San Jacintos, and only two genera

and two species on the San Bernardinos, whereas, if they

conformed with our laws of distribution, we should expect

at least six, and possibly seven or eight, species to be

found there. On page 35 of Grinnell’s ‘‘Biota of the San

Bernardino Mountains’? (1908) we find reference to a

number of carnivores now rare or extinct on the San

Bernardinos, which undoubtedly have been exterminated

by man within the last fifty years. Counting these forms,

which it seems to me we are justified in doing, the table

bears out the law without a single exception, not only for

138 THE AMERICAN NATURALIST [VoL. XLVIII

the total of mammalian forms, but the totals for each

order and for each family.

In comparing the three areas in which all the life zones

are involved, the truth of the effect of extended distribu-

tion on speciation is still more forcibly impressed upon

us. In this case we are comparing areas which are suc-

cessively larger in size, the San Jacintos, with their foot-

hills and low passes involving the fauna of an area of

about 2,500 square miles, the Sierras, about 60,000 square

miles, and the whole state of California about 158,000

square miles. The following table, derived from Table

III, is very significant in showing the diminishing in-

Genera | Species Index of Modification

Group |

Ungulates.....| 2 3 4 2 7 10 1.00 | 2.33 | 2.50

Rodents. ..... 16 28 31 41 110 203 2.56 | 3.93 | 6.45

Carnivores... . 9 15 17 10 29 51 1.11 | 1.93 43.00

Insectivores. .. 3 5 6 3 12 20 1.00 | 2.40 3.33

i 4 7 11 T 12 26 1.75 1.71 : 2.36

TO. es 34 58 68 63 170 310 1.85 | 2.93 | 4.56

crease of genera, and the constantly increasing addition

of species as the area is enlarged.

By comparing the upper zones of the San Jacintos

with the San Jacintos as a whole, and the upper zones of

the Sierras with the Sierras as a whole (see Table III),

we find that increasing the life zones has in a lesser

degree the same effect as increasing the geographic area

regardless of zones; in other words, adding life zones

tends to have the same effect on speciation as adding

faunas and associations without life zones. The follow-

ing table (derived from Table III) illustrates this:

San Jac. San Jac. Si Sierras

Mammals (Upper Zones) | (All Zones) (Upper Benes) (All Zones)

Genera Rl ERE UNM rie 18 34 45 58

Species 20 63 100 170

Index of mod : 1.11 1.85 2,22 2.93

Another rough test of the hypothesis was made in a

comparison of the mammalian faunas of some of our

No. 567] EFFECT OF DISTRIBUTION ON SPECIATION 139

large continental islands and zoologic regions, the results

being shown in Table IV. The data used in this table are

TABLE IV

SPECIATION OF MAMMALS IN VARIOUS CONTINENTAL ISLANDS AND ZOOLOGIC

REGIONS

Data from Sclater and Sclater (1899)

( Africa ó ‘eee 2.017 000 Guinea Madagascar

11,770,00 947, 312,000 228,000 M.

Group Sq. M.) ata Sq. M.) ea M.) ( Peg

Sp. Gen. Sp. Gen. Sp. Gen. |Sp.| Gen.| Sp. Gen.

Ungulates...| 155 35 1 1

Rodents..... 196 41 69 8 18} 5 13 T

vores.. 59 22 9 7

Insectivores, 73 20 9

o PN Sa 101 19 83 26 39; 16|) 21 12

Lemurs..... 8 3 36 11

Primates T2 6

Hyraces..... | 14 1

Elephants... 1 1

Edentates 6 9

pi | 144 36 36) 14

Monotremes 5 3 3 2

Totals. 2... | 685 | 128 301 73 | 169 59 | 96| 37| 100 47

Index of R |

ification...) 5.35 4.12 | 2.86 2.59 2.13

by no means up to date, being taken from the summaries

in Sclater and Sclater (1899), but the subsequent additions

to the faunas of the places concerned, and the splitting

up of genera and species, have probably been approxi-

mately proportionate in each of the five areas, and there-

fore the figures used are sufficiently accurate to be signifi-

cant. Comparing Africa, the Australian region, Australia,

New Guinea and Madagascar, which rank in size in the

order given, we find that the indices of modification of

their mammalian faunas are as follows: Africa 5.35,

Australian region 4.12, Australia 2.86, New Guinea

2.59, and Madagascar 2.13. Certainly these figures are

Significant.

Comparing the mammalian faunas of the various

islands of the Philippine Archipelago (Table V), we find

that there is even here some corroboration of our law of

140 THE AMERICAN NATURALIST . [Vou. XLVIII

TABLE V

SPECIATION OF MAMMALS IN ISLANDS OF THE PHILIPPINE ARCHIPELAGO

Data from Hollister (1912)

Island ` | Sq. Miles Sp. Gen. Index of Mod.

Reais oe aks cee | 40,969 72 40 1.80

UATE PVs Vy AG ne sa el 36,292 61 32 1.90

“EE rey Oo |` 5,031 16 13 1.23

TaS oe D E | 4,881 14 13 1.07

Founy oo acs 4,611 10 8 1.25

ee REO aa a Se She E | 4,027 21 18 1.16

Mindat., o oe. | 851 17 11 1.54

A N EUN S es 2,722 9 8 1.12

ir Sa wis bowie Ge koe | 1,762 8 7 1.14

ots a | 441 13 3 1.00

Maan us Rae Sak | 1,236 5 4 1.25

speciation. Considering the large element of chance in

the animal population of a group of islands of such small

size as those of the Philippines, where the various islets

are at a varying distance from each other, and their

faunas have originated from different sources, the rela-

tion between their size and the differentiation of their

forms is remarkably regular. In Table V, where the main

islands have been listed in order of size, with their num-

bers of genera and species of mammals, the deer have

been excluded entirely, since their generic and specific

differentiation is in too chaotic a state to be used. The

most striking fact brought out by the table is the lead

which the two large islands, Luzon and Mindanao, show,

not only in total number of forms, but in index of modifi-

cation as well. With the possible exception of Mindoro

and Palawan, practically none of the smaller islands is

supporting as large a variety of mammalian forms as

could be expected of it, a fact which might be explained

in a number of ways.

In all of the tabafations i given, the marine mammals

have been ‘entirely excluded since the factors affecting

their distribution and speciation are so different from

those of terrestrial mammals. In the majority of cases

marine mammalian families have a paucity both of genera

and species, a circumstance brought about by a number

of factors. Generally speaking, large, wide-ranging

}

No. 567] EFFECT OF DISTRIBUTION ON SPECIATION 141

forms, or forms which are poor in numbers of individ-

uals, are poor in genera and species, possibly due to the

comparative uniformity of their environment, which is

usually coincident. Most marine mammals are of these

kinds, and their paucity of types is emphasized by the

comparative uniformity of their environment, even in the

most FAITEN groups. From a distributional point of

view, i. e., taking into account life zones, faunas and

SA a cosmopolitan, oceanic, surface group of

animals does not range through as great a variety of

ecologic niches and environmental and climatic condi-

tions as does a cosmopolitan terrestrial group.

In order to determine whether the principles of distri-

bution and differentiation here set forth would apply to

birds as well as to mammals, a number of series of com-

parisons was made as with mammals, and with exactly

comparable results.

TABLE VI

SPECIATION OF BIRDS IN VARIOUS CALIFORNIA AREAS

Data from Grinnell (1913B), (1908), Willett (1912)

i California

oe Pan bee hae tgs Sonan Pamei a (158,000 54 ie)

- p

Gen Sp. Gen Sp. ‘Gen, Sp.

Passeres........... 62 82 79 114 87 197

a a E E 16 20 19 23 20 38

a A kk. 3 3 7 8 15

ceipitres. ........ 5 5 10 14 12 17

Columbe.......... 1 1 2 2 3 3

Moo fi 52 1 1 3 3 6 11

balie o. 3 3 4 4 9 10

OS a 1 1 5 6 6 8

Wades... 2 2 7 7 8 11

Anses... 2 2 5 5 11 11

Other water birds 1 1 12 14 16 26

coun Re 97 121 153 199 186 | 347

Index c ofr f mod... a. Ae 1.25 1.30 i 18

Table VI gives a comparison of genera and species

of resident birds of (a) the San Bernardino Mountain

region, (b) Southern California, and (c) California as

a whole. Almost without exception, in each individual

group of birds there is a reduction in the index of modi-

142 THE AMERICAN NATURALIST [Vou. XLVIII

fication as the area is restricted from California to the

Pacific Coast region of Southern California, and finally

to the San Bernardino region. The totals reflect the trend

in each group. While in the largest area the number of

genera is considerably less than double what it is in the

smallest, the number of species is more nearly tripled.

The Southern California area is intermediate.

TABLE VII

SPECIATION OF RESIDENT BIRDS IN AUSTRALIA AND TASMANIA

Data from North (1901-1909)

Australia Tasmania

ans (2,947,000 Sq. M.) (26,000 Sq. M.)

Sp Gen Fam Sp. Gen Fam

PRSROrOR: . 5. 3 os. 304 119 26 53 42 15

Wie Oe Pea eas eset cate e. 29 18 6 7 7 3

DE fee bee P A T eh 9 2 2 1 1 1

BOM UIIOR foe eo ie ke ees 27 17 2 11 9 2

AS T AT E E E se eee 57 14 3 11 9 3

ORME od ie ees se ee ss we 426 170 39 83 68 24

Index of generic mod Sah 4.35 2.83

Index of specific mod............... 2.30 1.22 | sie

Table VII shows a comparison of the families, genera,

and species of resident birds of Australia and Tasmania,

from North (1901-1909). Here again, in addition to a

very marked diminution of the total number of types in

Tasmania as compared with Australia, each group shows

a considerable decrease in the ratio of genera to families,

namely, from 4.35 in Australia to 2.83 in Tasmania, and

of species, to genera going from 2.30 in Australia to 1.22

in Tasmania.

Table VIII is a similar comparison of (a) the resident

birds of Ireland, from Hartert (1912), (b) the resident

birds of all the British Isles, from Hartert (1912), (c) all

the species of the Palaearctic region, the great majority

of which are resident in one part or another, from Dresser

(1902), (d) all the species of Japan, many of which are

not resident, from Ogawa (1908), and (e) all the species

of Kamtschatka, where the majority are resident, from

No. 567] EFFECT OF DISTRIBUTION ON SPECIATION 148

TABLE VIII

SPECIATION OF BIRDS IN VARIOUS PALEARCTIC REGIONS

Data from Hartert (1912), Dresser (1902), Ogawa (1908) and

Stejneger (1885)

British Palearctic Kam-

Ireland | “Isles. | Region | (1477o | tschatka

Sq. M.) (120,900 | (19,150,000 | Sq. M.) (105,000

Group z Sq. M.) | 5q. M.) Sq. M.)

Sp. | Gen. | Sp. | Gen. | Sp. | Gen. | Sp. | Gen. | Sp. | Gen.

A S eee oe. 57 | 35 85 | 42 | 610; 116 | 180 | 64 | 55 | 38

APS es T 4 4 T 7 1 | 34 16 8 5

van MO O si 2 5 4 34 1i | 14 4 3

a E ES p 4 4 12 7 66: 31| 23) H | 15 7

CAMUIN T Sc si cc cs 4 2 4 2 2 0: 23 6 0 0

E eR re a) 4) 8) Ti 7p ed ALPS BS

ET S T ee oe 10 9 15 1l 97; 32} 45) 21 25 17

sa, PEN i ieee ie AR 4 4 5 5 13 | 27 9 1 1

UU TT A se i 1 1 1 1 31; 1 23 12 0

PION ek: 8 8 16 | 12 24 89] 21; 28; W

Other water birds...... 23 | 14 | 80| 15] 129) 35 32 | 18

Total aR e ee Wed wea aloe ae 124 | 87 | 188 | 113 |1,251) 310 | 491 | 204 | 183 | 112

index of mod... i- 1.42 1.66 4.00 2.40 1.63

Stejneger (1885). The increase in index of modification

from Ireland to the British Isles, and then to the entire

Palaearctic region, is almost exactly what should be ex-

pected. The greater number of both genera and species

in Japan as compared with Kamtschatka reflects the

greater variety of ecologic niches in a warm country as

compared with a cold one of comparable size. A com-

parison of the resident species of Japan with the resident

species of the British Isles would be of very great inter-

est, but such a list of Japanese birds is not available. The

very striking similarity between the speciation of birds in

Kamtschatka, and that in the British Isles, both in num-

ber of genera and of species, is very remarkable. The

interesting manner in which the balance of nature is pre-

Served is shown by the large representation of raptorial

birds to parallel the abundance of shore birds and Anseres.

That reptiles and amphibians are influenced in their

Speciation by their distribution is indicated by Table IX,

which shows a comparison of the genera and species of

amphibians, lizards, and snakes, in three of the geo-

graphic areas defined by Cope (1898).

144 THE AMERICAN NATURALIST [Vou. XLVII

TABLE IX

SPECIATION OF AMPHIBIA AND REPTILIA IN NORTH AMERICAN AREAS

Data from Cope (1889), (1898)

Lower California District) Western Sub-region | Medicolumbian oh

(12,000 Sq. M.) (500,000 Sq. M.) (4,500,000 Sq. M.

Index of Index of Index of

Sp.| Gen. Mod. Sp. a Sp.| Gen. Mod.

od | Mod.

Amphibia... .. ee 8 1.25 23| 10 | 2.30 |130| 28 4.64

Lacertilia. .... 17| 13 1.30 |28| 13 2.15 |143| 31 4.61

Ophidia....... 16| 12 1.33 |20| 9 2.22 |191| 45 | 4.24

The ‘‘Lower California district’’ consists of only the

tip of Lower California; the ‘‘ Western subregion’’ em-

braces the Pacific slope of North America from Northern

Mexico, east of the Sierras, to Oregon, where it crosses

the Sierras to the Rocky Mountains, including northern

Idaho, eastern Montana, and most of British Columbia.

The ‘‘ Medicolumbian region’’ includes northern and cen-

tral Mexico, and most of the United States and Canada

north to a line drawn diagonally from New England to

Alaska, interdigitating on its border with the ‘‘ Holarctic

region.”’

The almost exactly parallel increase in the indices of

modification in the three groups of cold-blooded verte-

brates considered, as the area is extended, is quite remark-

able. All three groups average from 1.25 to 1.33 species

per genus in the smallest area, from 2.15 to 2.30 in the

intermediate area, and from 4.24 to 4.64 in the largest

area.

As suggested by Professor Kofoid, a factor influencing

speciation in such diverse vertebrates as mammals, birds,

reptiles, and amphibians, should be very widely appli-

eable to speciation in the entire animal kingdom.

A series of statistics relating to various orders of in-

sects and other invertebrates has been compiled to ascer-

tain whether in these groups as well as in vertebrates, the

number of species increases out of proportion to the

genera, as the size of the area, in a distributional sense,

is enlarged.

No. 567] EFFECT OF DISTRIBUTION ON SPECIATION 145

TABLE X

SPECIATION OF ELATERIDZ IN VARIOUS AREAS OF UNEQUAL SIZE

Data from Schwarz (1906)

Region Sq. Miles Sp. Gen. Index of Mod.

“esis aay «eee 11,770,000 574 55 10.43

pep car Fuk care pe 228,000 245 36 6.80

nn Pe TSS NS oe 1,760,000 438 53 8.26

Peay cis Geneon 296,700 150 40 3.75

MUMIA ours 184,000 177 41 4.31

hemi ea ee ae 0,000 125 3.37

CMM Ss 6s si iors eae 25,333 96 28 3.42

eee 2,947,000 386 42 9.19

ETE ee 312,000 20 3.05

New Zealand.......... 104,750 137 24 5.70

T TETE 26, 7 1.86

Table X was compiled to show the number of genera

and species of beetles of the family Elateride in various

continents and islands, the regions chosen for comparison

being well defined areas of unequal size

A careful inspection of this table shows that with only

two exceptions the indices of modification are directly

proportional to the size of the areas. Borneo and New

Guinea, however, not only show a smaller index of modi-

fication than should be expected of them, but are poor in

total number of types. Nevertheless, when we reflect that

these two islands are not nearly so thoroughly known to

science as are the other areas considered in the table, no

great significance can be attached to their seeming paucity

of known types.

Table XI shows the number of genera and species of

Limnophilide, a family of Trichoptera, in eastern North

America (east of the Rockies) as compared with North

America as a whole. It will be noticed that while in the

TABLE XI

SPECIATION OF LIMNOPHILIDÆ (TRICHOPTERA) IN NORTH AMERICA

Data from Ulmer (1907)

Region | Sq. Miles | Sp. | Gen. | Index of Mod.

North America............ | 8,000,000 | 98 | 27 | 3.63

Eastern North America. . 5,000,000 45 20 2.25

146 THE AMERICAN NATURALIST [Vou. XLVIII

larger area the number of species is more than double

what it is in the smaller area, the increase in genera is

only about one third, increasing the index of modification

from 2.25 to 3,63.

~ Table XII shows practically the same thing in the case

of the hawk moths of the family Sphingide.

i TABLE XII

` SPECIATION OF SPHINGIDÆ IN AMERICAN AND AFRICAN AREAS

Data from de Rothschild and Jordan (1907)

Area Sq. Miles Sp. | Gen. | Index of Mod.

Wek hd 8 ie a 76,000 61-| 20 3.05

Mexico kai Sai Dittik: 64s 975,200 | 122-| 34 3.58

Bouth Aois... o cs bie 7,000,000 | 197 | 35 5.62

Mex., Cent. roy anes. Amis: a S 7,975,200 | 237 | 40 - |. 5.92

Mex., Cent. Am., S. Am., and W. I.. 8,051,200 | 262 | 41 6.39

Bonbon. ési Si ews iv ies 965 7 5 1.40

MO ef on heed ee 8,000 9 20 1.95

By es ce ee ees Ge Ca iS eee 11,772,000 166 48 3.45

Afrion and Mad: ine neti Re eek vee 12,000,000 195 53 3.67

Afric: deny and Bourbon: os isini 12,000,965 197 53 ae 3.71

te this case iwo series of piaiations were made, one

showing the number of genera and species in various

Neotropical areas, and combinations of these areas, the

other showing a similar tabulation for various Ethiopian

areas, with similar combinations. It will be observed that

the speciation in the West Indies is very large for the size

of the area involved, but when we consider the abundant

opportunity that has been given for isolation to operate,

this is not surprising. The index of modification is quite

low. Mexico and Central America have a larger specia-

tion, compared with South America, than would normally

be expected, the reason being that Central America is the

American center of distribution. The index of modifica-

tion, however, reflects the smaller size of the area, being

considerably lower than that for South America. : The in-

crease in index of modification from 5.62 to 6.39, as areas

are successively added to South America, is significant.

Looking now at the Ethiopian regions, we find that there

is the same disproportionate increase of species over

No. 567] EFFECT OF DISTRIBUTION ON-SPECIATION 147

genera in successively larger areas, the index of modifica-

tion increasing from 1.40 in the small island of Bourbon

to 1.95 in Madagascar, and 3.45 in Africa. Combining

Africa and Madagascar, this is increased to 3.67, and with

the island of Bourbon, to 3.71.

Table XIII is one of especial sriterest; since it digas

TABLE XIII

SPECIATION OF MARINE GAMMARIDEA (AMPHIPODA) IN VARIOUS SEAS

Data from Stebbing (1906)

Area Sp. Gen. -| Index of Mod.

M Mediterranean en ris ee a a ae 147 | 2.19

Arptic Godam i ai aa a eee a 311 | 140 2.22

ec Aono (JOUER. ky. 5c. i. hee ics ree 475| 176.|. 2.70

me Anau Ocak ia Gl AAS 44 | 1.47

Aroue and N. Allie oso os ces Be canes 588 |- 191 | 3.07

Arctic, N. Atlantic, so? S Atlante ok coe eek 645 | 207 3.11

Arctic, N. eae S. Atlantic pe Med. Bee st 735 | 214 | 3.43

Whee Gy 6a see 1,383 |.313 | 4.22

with a marine instead of a terrestrial group. It embodies

the results of a compilation of the marine genera and

species of Amphipoda of the suborder Gammaridea in a

number of the oceans and seas of the world. Since it is

primarily a cold-loving group, the largest numbers are

found in the cold seas, the Arctic and North Atlantic being

the home of considerably over half of the known marine

species. It is very likely that when the Antarctic regions

have been studied as thoroughly as the northern regions,

the number of species from that part of the world will be

very considerably increased. At the time of Stebbing’s

work on Amphipoda, our knowledge of Antarctic ome

contiguous areas was very meager.

The steady inerease of the index of modification from

the smaller to the larger seas is striking. The Mediter-

ranean Sea, although it is the most thoroughly known of

all, has the lowest index of modification, namely 2.19, the

FOS Ocean -comes next with 2.22, and then the North

Atlantic with 2.70. The small number of species from the

South Atlantic and Antarctic regions has already been

148 THE AMERICAN NATURALIST [Vou. XLVIII

mentioned, and its low index of modification may be at-

tributed to the same sort of imperfect knowledge as in the

ease of Borneo and New Guinea in Table X. The con-

stant growth of the number of species per genus from

2.22 to 3.43 as the various seas and oceans are added to-

gether, exactly parallels the results obtained in a similar

way fora terrestrial group in Table XII. The comparison

of the speciation of the largest area for which it was

worked out, with the speciation of the entire group, many

species and genera of which inhabit fresh water, is inter-

esting, jumping as it does from 3.43 to 4.22. From the

facts brought to light by this table it can hardly be doubted

that practically the same influence is brought to bear on

the speciation of marine as on terrestrial organisms by

the extent of their distribution. ©

The theoretical explanation here proposed for this phe-

nomenon involves a number of complex problems relating

to evolution and speciation, including isolation, effect of

time, causes of specific and generic modification, etc., each

of which will be dealt with in the following pages as they

seem to influence the law here proposed.

Let us first consider the factor of isolation in relation

to the production of new forms. As excellently stated by

Cook (1909), isolation can not be considered as a cause or

factor in evolution, since changes in the characters of

species are not dependent upon the subdivision of species

to form additional species. To quote from him:

The separation of species into two or more parts allows the parts

to become different, but there is every reason to believe that evolutionary

changes of the same kind would take place if the species were not

divided. That the isolated groups become different, does not indicate

that isolation assists in the process of change. It gives the contrary

indication that changes are restricted by isolation. If isolation did not

confine the new characters to the group in which they arise, the groups

would remain alike, instead of becoming different. . . . Isolation is

the shears that splits the species, not the loom that weaves it.

Therefore, while isolation can not be considered a factor

in evolution, it is an important factor in speciation.

Species vary in many directions or orthogenetically pro-

No. 567] EFFECT OF DISTRIBUTION ON SPECIATION 149

gress in a definite direction, but the trend of variation or

progression may be different in one locality, and tend

towards a different result, from that of another locality.

Whether the evolution, usually in more or less divergent

directions, of segregated groups of individuals be looked

upon (1) as the accumulation of numerous slight varia-

tions which have a different average character in any two

portions of a species, as originally explained by Darwin

(1859, Chap. 4) or (2) purely as the result of natural selec-

tion, as argued by Wallace (1858), or (3) as the result of

a change in the average character of two portions due to

the uneven occurrence of mutations in the two portions,

a conclusion reached by Dewar and Finn (1909, p. 380), or

(4) as the result of orthogenetic evolutionary tendencies

inherent in the species and influenced by the environment,

as Eimer suggested (1897, Chap. 1), does not concern us

here,—the general tendency appears to be that two iso-

lated portions of a species as a general rule trend in

different directions, and diverge farther and farther as

long as they are isolated.

It is assumed that the greater the length of time given

for the influence of isolation to be felt, the farther apart

are the two originally identical divisions likely to trend,

however the dissimilar evolution be interpreted. As

stated by Tower (1906), in speaking of the method of

evolution of the Chrysomelid genus Leptinotarsa,

We can interpret the conditions found by any of the current

hypotheses; but explaining a condition by an hypothesis is not the

same as that the conditions found are evidence in support of an

hypothesis, although it is often so used.

The existence of distinct variations, subspecies, and

ultimately species and genera, in isolated areas is a too

frequently observed phenomenon to be looked upon as

anything else than a self-evident truth, but that this should

necessarily be considered as supporting any particular

theory of evolution can not be argued.

The profound results of prolonged isolation may be

observed in the fauna of some of our long-separated con-

150 THE AMERICAN NATURALIST [Vou XLVIII

tinental islands, such as Madagascar, Australia and New

Zealand. Decreasing degrees of isolation may be observed

in our West Indian islands, where some generic differenti-

ation has occurred; in the Santa Barbara islands, where

there has been a differentiation of species; and the de-

tached mountain ranges of Southern California, where the

upper life zones are at present in an isolated condition,

but have been so only long enough to develop a few new

subspecies, and to lose many of the types of the mother

range, in accordance with the law proposed by Grinnell

and Swarth (1913) that ‘‘the smaller the disconnected

area of a given zone, or distributional area of any other

rank, the fewer the types which are persistent therein.’’

From this it is apparent that the time element, in con-

junction with isolation, may have a very decided effect on

the number of genera and species in a family, but since,

from a geologic point of view, animals appear to have

reached a new equilibrium very quickly after a geographic

change, the time element may have little effect on the num-

bers of genera and species relative to each other in any

given area. In other words, as fast as new genera are

produced in a given area, the species within the genera

will tend to be produced in the same ratio, thus leaving

the index of modification unaffected.

As an example of the effect of time and isolation let us

take a hypothetical case. Let us assume that a certain

island became divided into two islands of unequal size,

and that after a short period of segregation, just long

enough for the fauna to readjust itself to the smaller

areas and reach a new equilibrium, we had say six species

in three genera on the larger island, and three of the same

species in two of the genera on the smaller one. After

a long period of isolation we should have approximately

the same number of genera and species on the two islands,

but they would have diverged to generic differentiation.

In other words, the effect of time in conjunction with iso-

lation is to increase the number of genera and species in

the family, while the index of modification undergoes little

change.

No. 567] EFFECT OF DISTRIBUTION ON SPECIATION 161

This leads us to a consideration of the factors involved

in the differentiation of genera as contrasted with the

differentiation of species. In general.it may be said that

extrinsic modifications, i. e., those which are in some way

connected with changes in temperature, humidity, char-

acter of flora, food, and other environmental conditions,

and which usually affect such characters as color, size,

length of hair, etc., lead to differentiation of species and

subspecies primarily. On the other hand, intrinsic modi-

fications, i. e., those which are related directly or indi-

rectly to a change in the habits or mode of life. of the

animal or the occupation of a new niche in nature, usually,

if not always, lead to generic or family differentiation,

since it is evident that changes fitting an animal to live

arboreally instead of terrestrially, for instance, are of

such a nature, that if they are perpetuated and carried to

perfection, will not stop at specific difference but will

become of generic importance.

It might be argued that there are no modifications which

might not, if carried far enough, ultimately lead to generic

differentiation. This is possible, but very improbable,

because the modifications here alluded to as ‘‘extrinsic’’

are of such a nature that in the varying climatic condi-

tions there are likely to be intermediate forms which make

the division of the more widely separated ones into genera

impracticable. In the case of our ‘‘intrinsic’’ modifica-

tions, intermediate forms are not so likely to exist when

once the incipient changes leading to an altered mode of

life have reached a fair degree of perfection.

As a concrete example of what is meant by extrinsic

and intrinsic modifications, let us take the squirrels of a

given region, say eastern North America. There are four

genera to be distinguished,—Sciurius, Tamias, Sciuro-

pterus and Arctomys. The genus Sciurus contains

Strictly arboreal, mostly nut-eating, omnivorous forms.

Tamias includes forms which are terrestrial, diurnal,

dwelling in natural or artificial holes and crevices, and

with a device for carrying food in their cheeks. Sciuro- .

152 THE AMERICAN NATURALIST [Vou. XLVIII

pterus is an arboreal type which is nocturnal, and has de-

veloped characters which enable it more easily to travel

from tree to tree. Arctomys is the most highly modified

form, and has departed most widely in its habits; it is

entirely terrestrial, seeks shelter in artificial burrows,

eats grass, and hibernates.

Were we to study the characters separating these gen-

era, we should find that they are all characters which

enable the animal best to occupy the ecologic niche it fills.

If now we select any one of these genera and examine its

species, we perceive that the differences we find are not

such as could clearly be related to differences in mode of

life or habits, but rather such differences as are induced

by the circumstances mentioned above, such differences

being size, color, length of feet and tail, texture of fur,

etc.—t. e., extrinsic variations.

An intorcating example of both extrinsic and intrinsic

modifications in an incipient stage may be found in the

song-sparrows of western United States. Let us compare

the form of the humid northwest coast belt, Melospiza

melodia morphna, with the form of the arid Arizona des-

erts, M. m. fallax. The differences to be observed in color

and size are very noticeable, and would undoubtedly lead

to their separation into two distinct species were it not for

the complete chain of intermediate forms. But even if the

chain of intermediate forms were not complete, and after

a period of segregation the numerous intergrading sub-

species became broken up into a few well-marked species,

nevertheless, unless a change in mode of life of the bird

were involved, however far the extremes of color and size

might tend, they could not be given generic distinction

because of the intermediate forms, inhabiting semi-arid or

semi-humid regions, which would be almost certain to

exist. It hapens, however, that Melospiza melodia mor-

phna, and M. m. fallax, do differ considerably in mode of

life, the former being a beach comber, the latter a nomad

of the desert. It would be expected, therefore, that if

these two subspecies were isolated, the modifications re-

No. 567] EFFECT OF DISTRIBUTION ON SPECIATION 153

lated to their difference in mode of life, already shown

in an incipient manner, would soon lead to their generic

differentiation.

It is not argued that under a given set of ecologic con-

ditions, only one type could be produced, nor that accord-

ing to the idea of some zoologists, as set forth and refuted

by Grinnell and Swarth (1913), should individuals of one

geographic race be transplanted into the region of a dif-

ferent geographic race, the first race would assume within

a few generations all the characters of the second race.

Whether the changes due to the influence of the environ-

ment be looked upon as the results of natural selection

and adaptation, or merely as the results of a stimulus to

the germ plasm, the new type would not necessarily be

always the same, this, however, depending upon the num-

ber of potential responses in the type, and, as excellently

shown by Ruthven (1909) in his study of evolution in the

genus Thamnophis, upon the modifications previously

undergone by the type we are dealing with.

It is very evident that there are many variations in

animals which seem to fall into neither the extrinsic nor

intrinsic category, but which are neutral and vary inde-

pendently of climate or habits, and may be inherited phy-

logenetic tendencies. It is very largely due to these

neutral variations, frequently to be ascribed to ortho-

genetic evolution, tending in different directions in dif-

ferent places, and given an opportunity to diverge by iso-

lation, that different species may be produced to occupy

regions of similar climatic and environmental conditions,

and different genera may be found occupying the same

ecologic niches.

To choose an example in the same family quoted before,

we may cite the case of Tamias in eastern North America,

and Eutamias in western North America. In this case the

characters separating the genera are not clearly related

to their mode of life, the chief difference being the loss of

one small premolar in Tamias, and its retention in Euta-

mias. The extent of divergence of these neutral varia-

154 THE AMERICAN NATURALIST [Vou. XLVIII

tions depends on the duration of geographic segregation,

and may therefore be of specific, generic, family, or ordi-

nal rank.

To sum up, specific modifications may be of three kinds:

(1) extrinsic modifications, induced by changes of climate

and environmental conditions; (2) neutral modifications,

due to a different trend of evolution in segregated regions;

(3) incipient generic modifications. On the other hand,

generic modification may be either intrinsic modifications,

concomitant with changes in mode of life or habits of the

animal, or neutral modifications as above, given generic

value by a longer period of segregation.

Having dwelt for some length on these preliminary con-

siderations, let us now apply them to the case in hand and

see how they affect differentiation into species and genera

through extension of range.

. It is a well-known biological fact that different types of

a group of animals, at least of higher animals, are found

associated with different environments; nearly related

species do not, as a rule, live comfortably together in the

same environment, and nearly related genera do not

occupy the same ecologic niche in a given zoogeographical

area. This does not seem to hold true for animals of

lower organization, as conclusively shown by Kofoid

(1907). It is common for a group of animals, unless hin-

dered by an impassable barrier or unfavorable environ-

mental conditions, not only to continually extend its range

into new territory, but also to attempt to live in as many

different niches in nature as possible within a given area.

Such attempts to invade new ecologic niches are frequently

concomitant with heritable modifications better fitting

them to occupy their new situation, though it is difficult to

say whether these modifications are causes or results of

the change in mode of life. However this may be looked

upon, the tendency to occupy new niches in nature is fre-

quently accompanied by intrinsic modifications, and there-

fore by generic differentiation.

From this we may safely assume that in a given area

No. 567] EFFECT OF DISTRIBUTION ON SPECIATION 155

a family of animals, by adaptive evolution, will approach

a maximum of generic differentiation which can be sup-

ported in that area. In other words, every suitable eco-

logic niche which is represented in the region considered

will be invaded by the family, and even in a small area

there is likely to be a considerable generic differentiation,

especially if isolation has had any opportunity to operate -

within the area, in breaking up the genera and species.

Let us assume that in one unit of area a certain family,

Sciuridae for example, was represented by three genera,

each with three species. Second, let us assume that this

family kept spreading into additional units of area. With

each new unit, the chance of new suitable ecologic niches

being represented would decrease, and therefore the

chance of new genera being represented would decrease,

since if a genus were fitted for its niche in nature under

certain conditions of climate and environment, it would

in the majority of cases not be likely to undergo any

radical changes in the occupation of the same niche

under somewhat altered conditions of climate and environ-

ment; i. e., the stimulus for intrinsic modification would

be lacking.

On the other hand, with each additional unit of area, the

chances of the combined conditions of temperature, hu-

midity, and environment being different, would remain the

same. In other words, the chances of the three dimen-

sions influencing the life of a region, i. e., ‘‘life zone”

(controlled by temperature), ‘‘fauna’’ (controlled by hu-

midity), and ‘‘association’’ (controlled by the effect of

the other two plus a number of other environmental con-

ditions), intersecting at the same point would be almost

equally improbable with each succeeding unit of area.

Since it is changes in “‘life zone,’’ ‘‘fauna,’’ or ‘‘associa-

tion’’ which produce extrinsic changes, and therefore lead

to differentiation of species and subspecies primarily, the

increment of species would average nearly the same for

each succeeding unit of area, other factors remaining

equal. It should also be taken into consideration that

156 THE AMERICAN NATURALIST (VoL. XLVIII

with the invasion of new zoogeographic areas, contact

with allied forms is frequently experienced, and oppor-

tunity is thus afforded for cross breeding and hybridiza-

tion, the result of which upon the germ plasm appears to

be as influential in the production of new forms as is the

shock of new environmental conditions. The constant

increase in species and subspecies accompanying invasion

of new territory, going hand in hand with a diminishing

increase in genera, results in the constantly larger index

of modification as the area inhabited by a group is

extended.

SuMMARY

1. Extent of distribution has a direct influence on the

speciation of the group concerned in this way, that as the

range of a group of animals is extended, the species in-

crease out of proportion to the genera, the genera out of

- proportion to the famliies, and the families out of pro-

portion to the orders.

2. Comparison of different families having unequal geo-

graphic ranges is usually inaccurate due to the great dif-

ferences in the other factors controlling their speciation.

Those families which do lend themselves to such a com-

parison show decidedly the effect of extent of distribu-

tion, e. g., the bats and some of the insectivores, the fami-

lies of widest distribution having the largest indices of

modification. A number of exceptions exist in the form of

certain wide ranging genera which have a paucity of

species. We have no adequate explanation for this

phenomenon.

3. Comparison of the faunas of areas of different size

gives very accurate results. A number of tabulations show

as a whole an invariable increase in the index of modifica-

tion as the distributional area is extended by the addition

of either life zones, faunas, or associations. Such tabu-

lar comparisons were made for all the classes of ter-

restrial vertebrates, for several families of insects, and

for the marine Amphipoda of the suborder Gammaridea.

Allowing for explicable exceptions, the increase in number

No. 567] EFFECT OF DISTRIBUTION ON SPECIATION 157

of lower systematic groups out of proportion to the in-

crease of higher systematic groups as the area considered

is enlarged is a remarkably constant and wide-spread

phenomenon.

4, The theoretical explanation here proposed for this

phenomenon involves a number of complex problems

relating to evolution and speciation, including isolation,

the time element, and causes of specific and generic

modification.

5. Isolation is an important factor in speciation, since

the separation of species into two or more parts allows

the parts to become different. The degree of divergence

of the segregated parts is largely dependent upon the

duration of segregation.

6. Time, in conjunction with isolation and evolution,

tends to increase the number of genera and species in a

family, but the index of modification, i. e., the average

number of species per genus, remains approximately the

Same in a given area.

7. Three types of modifications in animals may be

named :—first, ‘‘extrinsie’’ modifications, which are in-

duced by climate and other environmental conditions, and

which lead to differentiation of species and subspecies

primarily; second, ‘‘intrinsic’’ modifications, which are

concomitant with a change in habits or mode of life of the

animal, due to the occupation of a new ecologic niche, and

which usually lead to generic or family differentiation;

and third, neutral modifications, which are merely the

result of the natural tendency of all animals to vary and

to be subject to more or less orthogenetic evolution,—

modifications which can not be correlated with environ-

mental conditions, nor with a change in mode of life of

the animal, but which may be influenced largely by in-

herited tendencies. Such modifications are responsible for

the production, through isolation, of different species to

live under the same climatic and environmental conditions,

and of different genera to occupy the same ecologic niche.

8. Specific modifications may be of three kinds: (1) ex-

158 THE AMERICAN NATURALIST [Vou. XLVIII

trinsic modifications, (2) neutral variations in segregated

regions, (3) incipient generic modifications. Generic modi-

fications may be (1) intrinsic modifications, or (2) neutral

varations, given generic value by a longer period of

segregation.

9. Since different. types of a group of animals are

usually found associated with different environmental con-

ditions or different ecologic niches, and since it is common

for animals, if unhindered, not only to extend their range

continually into new territory, but also to occupy new

ecologic niches, and since these tendencies lead to specific

and generic differsitingions: respectively, any given area

will have a differentiation of species proportionate to its

variety of environmental conditions, and of genera pro-

portionate to its variety of suitable ecologic niches.

10. Since, as the area of distribution is extended, the

chance of new conditions of climate and environment being

represented remains approximately the same, the increase

in number of species is nearly proportional to the increase

in the area of distribution, but since the chance of new

ecologic niches being represented in most cases constantly

decreases, the increase in genera proceeds at an ever-

diminishing rate. This, going hand in hand with the

nearly constant increase in species or subspecies, results

in a constantly increasing index of modification.

LITERATURE CITED

Cook, O. F.

1908. Evolution without Isolation. AM. NAT., 42, 727-731.

Cope, E. D.

1889. The Batrachia of North America. Smithsonian Inst. Nation. Mus.

Bull., 34, 1-525, 86 pls., 119 figs. in text.

1898, The Caiectilhada. Lizards and fechas of. North America. Smith-

sonian Inst. Nation. Mus. Rep., 1898, 155-1270, 36 pls., 347 figs.

arwin, Ch.

; 1859.. The pe of Species, 2d:ed., 1869. New York, D. Appleton & Co.,

, 440, 1 pl.

Dewa, D; BP Finn, F.

The Makitig of Species. London, John Lane Co., xix, 400, 15 pls.

No.567] EFFECT OF DISTRIBUTION.ON SPECIATION 169

Dresser, H. E.

+ 1902.7 A Manual of Palearctic Birds, London, Vols, 1 and 2, 922, 2 pls.

Eimer, Th. G. H.

1897. Die Entstehung der Arten auf Grund von -Vererbung erworbung

Eigenschaften nach den Gesetzen organischen Wachsens, II.

Leipzig, W. Engelmann, xvi, 513, 235 figs. in text.

Grinnell, J.

1908. The san of the San Pree Mountains. Univ. Calif. Publ,

Zool., 5, 1-170, pls.

-19134.' A saei List Fs the Mammals of California, Proc. Calif.

Acad. Sci., 4th series, 3, 265-390, pls. 15-16.

1913B. A Distributional List of the Birds of California. Mss.

Grinnell, J., and Swarth, H. $S.

n Account of the Birds and Mammals of the San Jacinto Area

of Southern California, with Remarks upon t the Behavior of

Geographic’ Races on the Margins of their Habitats. Univ.

Calif. Publ. Zool., 10, 197-, pls. 6-10, figs. 1-3 (in press).

Hartert, E., Jourdain AB geek R., Ticehurst, N. F., and Witherby, H. F.

1912. A Handlist of British Birds. London, " -Witherby & Co., XII, 237.

Hollister, N.

1912. A List of the Mammals of the Argir Islands, Exclusive of the

Cetacea. Philipp. J. Bci., D, 7,

Kofoid, C. A.

1901. The Limitations of Isolation in the Origin of Species. Science,

25, 500-506.

North, A, J.

1901-1909. Nests and Eggs of Birds Found Breeding in Australia =

Tasmania. Austr. Mus., Sydney, Sp. Catalogue I, Vol. 1,

366, pls. A 1-8, B 1-7, ‘text figs.; Vol. 2, vii, 380, pls. A Hoe

B 8-13.

Ogawa, M.

1908, A Handlist of the Birds of Japan. Annot. Zool. Jap., Tokyo, 6,

337-420.

de Rothschild, W., and Jordan, K.

1907, Lepidoptera, Fam. Sphingidæ. Genera Insectorum (ed. by Wyts-

n, P.), 57, 157, 8 col. pls.

— A. y

1909. A Contribution to the Theory of Orthogenesis. AM. NAT., 43,

09.

1906.” Coleoptera, Fam. Elateride. Genera Insectorum (ed. by Wytsman,

1906. Amphipoda, I. Gammaridea. Das Tierreich (ed. by F. E.

eral Pein R. Friedländers Son, 21 xxxix, 806, numerous

figs. in t

160 THE AMERICAN NATURALIST [Vou. XLVIII

Stejneger, L.

1885. Results of Ornithological Explorations in the Commander Islands

and in Kamtschatka. Smithsonian Inst. Nation. Mus, Bull., 29,

382, 8

Tower, W. L.

1906. An Investigation of Evolution in Chrysomelid Beetles of the Genus

Leptinotarsa. Washington, Carn. Inst. Publ., 48, x, 321, 30 pls.,

31 figs. in text.

Ulmer, G.

1907. Trichoptera. Genera Insectorum (ed. by Wytsman, P.), 60, 259,

13 col. pls., 28 black pls.

Wallace, A. R.

1858. On the Tendency of Varieties to Depart ee from the

Original Type. J. Proc. Linn. Soc., London, 2.

Willett, G.

1912. Birds of the Pacifice Slope of Southern California. Pac, Coast

Avifauna, 7, 1-122.

BIOLOGY OF THE THYSANOPTERA!

DR. A. FRANKLIN SHULL

UNIVERSITY OF MICHIGAN

I. FACTORS GOVERNING LOCAL DISTRIBUTION

INTRODUCTION

Tue Thysanoptera, commonly called thrips, are only

beginning to be known, in this country, by systematic

entomologists. The systematic knowledge is mostly con-

tained in the monograph of Hinds (1902), a more recent

synopsis by Moulton (1911), and a few other papers deal-

ing with new species and with relationships, prominent

among which is the work of Jones (1912). Biologically

the group is still less known. A considerable number of

papers have been issued from experiment stations, de-

scribing the life history (egg, larval, pupal and adult

Stages) and habits of thrips of economic importance. Be-

sides these the principal recent work of a biological

nature is a paper of my own (Shull, 1911), on the ecology,

method of locomotion, mode of reproduction, and dissemi-

nation. The life cycle of mast species is still largely un-

own.

The first section of this paper is an attempt to carry

into further detail the study of the ecology of the Thy-

Sanoptera. The first ecological scheme, so far as I am

aware, worked out for the Thysanoptera was that of Jor-

dan (1888), who divided thrips into three classes: first,

the flower-dwellers; second, the leaf-dwellers ; and third,

all other thrips (for example, those living on fungi, under

wet leaves, under bark of trees, on roots, on lichens, ete.).

The inadequacy of this classification, and the difficulty of

applying schemes of ecology adapted to other groups of

sects, was pointed out in my earlier paper, where I pro-

1Contributions from the Zoological Laboratory of the University of

> i No. 142 (Biological Station Series, Zoological Publication,

161

162 THE AMERICAN NATURALIST [Vou. XLVIIL

posed a new scheme, based on my observations in the

field. In this scheme, Thysanoptera were divided into

two groups: (1) interstitial species, those living in closely

concealed situations, as among the florets of composite

flowers, or in clusters of young leaves; and (2) super-

ficial species, those living on exposed surfaces, for ex-

ample, the surface of leaves. The interstitial species

were further divided into an anthophilous division

(flower-dwellers) and a phleophilous division (those

living under bark seales on trees). The superficial spe-

cies were either poephilous (on grass) or phyllophilous

(on leaves of plants other than grasses). The distinction

between poephilous and phyllophilous seemed warranted,

since grass-dwellers were found on many different

grasses, but rarely on other kinds of leaves.

Such a classification undoubtedly describes the facts,

but does not explain why the habitats named are the ones

chosen(?). The factors determining habitat were be-

lieved by me at that time to be character of food, and pro-

tection afforded. In some species one of these factors

predominated, in other species the other factor, while

others may have been influenced largely by both. In the

light of recent ecological studies, however, the explana-

tion of local distribution in terms of such general environ-

mental factors seems inadequate. Largely owing to the

work of Shelford (1911) upon the tiger-beetles, much

emphasis is now being placed upon the ecological impor-

tance of physiological factors. With a view to relating the

distribution of Thysanoptera to the physiology (more

specificially, behavior) of the various species, and thus

explaining that distribution in more definite terms, the

experiments and observations recorded in this paper

were made,

This work was done largely at the University of Michi-

gan Biological Station, at Douglas Lake, Michigan, sup-

plemented by observations at Ann Arbor, Michigan, in

Ohio and elsewhere.

No. 567] BIOLOGY OF THE THYSANOPTERA 163

Facts To BE EXPLAINED

The following are some of the facts of habits and dis-

tribution of the more abundant species for which physio-

logical explanations were sought. Some of these facts

are stated in my former paper, some of them doubtless

the common property of all thysanopterists; others, so

far as I know, have never been recorded.

Euthrips tritici is found almost exclusively in situa-

tions where it is concealed, as among the florets of com-

posite flowers, in clusters of young leaves, or in almost

any close crevice where the tissues are not too hard or

tough to be pierced. It appears to make little difference

what species of plant is inhabited, provided a concealed

situation is available. In the paper cited above (Shull,

1911) I gave a list of seventy species of plant on which

Euthrips tritici was taken, and I have since collected it on

a number of plants not included in that list. But with

rare exceptions, it has been found in crevices where it was

not readily visible. In related plants, it is always more

abundant in those affording concealed situations. Thus,

in white clover (Trifolium repens) and in red clover

(T. pratense), this thrips is usually abundant; while on

the related yellow, and white, sweet clovers (Melilotus

officinalis and M. alba, respectively), growing along with

the red and white clovers, Euthrips tritici is usually rare

or wanting. The flowers of Melilotus are widely sepa-

rated from one another on the stem, and do not afford

concealment (Shull, 1911).

If, while Euthrips is in one of these crevices, it is dis-

turbed, as by gently rubbing or pressing the flower, it

quickly comes out of its retreat and crawls rapidly away,

or takes to flight. The larve show the same behavior as

“ee “eae in this regard, except, of course, that they do

Anaphothrips striatus is found usually on grasses of

various kinds, rarely on leaves of other plants. The spe-

cies of grass seems to make little difference. Some indi-

viduals are found in perfectly exposed situations, as on

164 THE AMERICAN NATURALIST [Vou. XLVIII

the upper side of grass blades, others more or less con-

cealed in the rolled up young leaves (Shull, 1911). I have

found, however, that among the adults, those in exposed

situations are almost exclusively females, while those in

the rolled young leaves are either males or females. (For

the first time on record, the males of this species, as will

be shown in the second part of this paper, have been found

in considerable numbers.) The larve, according to my

observations, may be either exposed or concealed; the ex-

posed ones are predominantly the older larve.

In one of the grasses (Spartina michauxiana) on which

Anaphothrips was found in abundance at Douglas Lake,

Michigan, the leaves bear on the upper surface a set of

fine, but prominent, ridges running parallel to the axis of

the leaf. Adult females and larve of Anaphothrips on

the exposed parts of these leaves were always lodged be-

tween the tops of these ridges, and almost invariably

with their heads toward the base of the leaf. If disturbed,

they began to crawl along the crest of one of these ridges

toward the base of the leaf. It was possible to force them

to turn in the opposite direction, but if allowed to do so

they soon turned again toward the base of the leaf, often

continuing until they were among the rolled young leaves

in the center of the top of the plant.

Anthothrips verbasci is found exclusively on one spe-

cies of plant, the common mullein (Verbascum thapsus).

Furthermore, it is rare that a specimen of mullein, of con-

siderable size, is found free from the mullein thrips. Most

of the thrips are found among the florets or seed pods of

the spike. Less commonly they are to be seen on exposed

surfaces, as on the leaves or stem lower on the plant; but

these exposed individuals are mostly adults. The larve

are usually hidden on the flower spike unless that situa-

tion is crowded by a large number of larve; and the larvé

that are occasionally found exposed are mostly nearly

fully grown.

Anthothrips niger was not abundant enough during my

stay at Douglas Lake that many observations of its

No. 567] BIOLOGY OF THE THYSANOPTERA 165

habitat and behavior could be made. One fact, however,

is of interest in connection with an experiment to be de-

scribed. While the adults live mostly on flowers, some-

times concealed, sometimes more or less exposed, the

larve were always found concealed; moreover, it was

with difficulty that the larve could be driven from their

retreat by pressing the flowers. Frequently such vigor-

ous squeezing was necessary to dislodge them that the

larve emerging were injured; and a flower so treated was

often found later to contain numerous dead larve. In

this respect, the behavior of this species is in considerable

contrast to that, for example, of Euthrips tritici.

The habitats and behavior described above can be ‘‘ex-

plained’’ in large measure if we say, as I at first proposed

(1911), that certain species seek protection, or that cer-

tain other species have specific food requirements. Thus,

it might be said that Euthrips tritici seeks safety in

crevices, and flees danger when disturbed; that Anapho-

thrips striatus ‘‘prefers’’ grass for food, that it requires

as much protection as its commissarial activities permit,

and that its habitat and behavior are such as best fulfill

these requirements. Anthothrips verbasci might be said

to be limited to one article of diet, while protection is a

minor matter.

This explanation might be acceptable as far as it goes,

were it not that no species is immune to attack. I have

seen larve of Anthothrips verbasci frequently captured

by various bugs. Heads of mullein where thrips are

found nearly always bear bugs of the family Capside,

and observations convince me that they prey almost

wholly on the larve of the mullein thrips. The degree to

which they check the thrips was tested experimentally as

follows: Two mullein spikes of approximately equal size

and equally infected with thrips were selected. The

predatory bugs were removed from one of them, after

Which the spike was enclosed in a thin muslin bag. Two

weeks later the bag was removed. The enclosed spike

re a large number of full-grown larve, a few had

166 THE AMERICAN NATURALIST [Vou. XLVIII

pupated, and many were crawling on the inside of the

bag. The spike that was exposed, on the other hand, bore

but little over half the number of larve that were on the

protected one, none were quite full grown, and none had

pupated. Since nothing in the climatic conditions (heavy

rains, for example) could have caused this difference,

it is to be inferred that predatory bugs had devoured the

larger larve in considerable numbers.

Yet Anthothrips verbasci, according to my earlier ex-

planation, ‘‘chooses’’ its habitat almost exclusively in re-

lation to food, protection being a minor consideration.

Can we not explain habitat and behavior in these in-

sects in some way not implying choice, especially choice

between conflicting preferences? May we not assume that

certain elements of behavior are what they are without

reference to their usefulness? If we grant the possibility

of an affirmative answer to these questions, the experi-

ments about to be described will have significance.

EXPERIMENTS ON BEHAVIOR.

The following experiments were designed to show the

reaction of the commoner species of Thysanoptera to

what seemed to me the most probable external agents

affecting their distribution and behavior, namely, light,

contact and gravity. Inasmuch as I was not primarily

interested in how a given reaction was brought about, but

only in its end result, the experiments were rather crude.

Refinements were unnecessary, and their omission en-

abled me to use much greater numbers of individuals than

would otherwise have been possible. From ten to forty

repetitions of each test were usually made. The experi-

ments are described by species, only representative ex-

periments being given.

Euthrips tritici

Light. Exp. 1—Adults of this species were placed in

a glass tube about three feet long and one inch in diam-

eter, closed at the ends with corks. One end of the tube

No. 567] BIOLOGY OF THE THYSANOPTERA 167

was turned toward a small window, while the room was

rather dimly lighted. All the thrips crawled rapidly

toward the window. When the position of the tube was

reversed, the thrips reversed their crawling, again going

toward the window. The reaction was definite and in-

variable.

Exp. 3—A close-fitting sleeve of black building paper

was slipped over one half of the glass tube used in experi-

ment 1. The thrips were collected at the exposed end by

turning that end for a few minutes toward the window.

The covered end of the tube was then turned toward the

window. The thrips crawled rapidly toward the light,

until they reached the shadow of the sleeve. Here they

crawled about, apparently aimlessly, for half an hour an

inch or two within the sleeve or just outside it.

Contact. Exp. 1—When, in the light experiments, the

tube was reversed in position as soon as the thrips

reached one end, the insects immediately turned toward

the opposite end. But if the tube was allowed to rest for

some time, the thrips became settled quietly between the

glass and the sloping surface of the cork. The tube could

then be carefully reversed, and most of the thrips re-

mained lodged between cork and glass for many minutes,

some of them for hours. The positive reaction to contact

counteracted the positive reaction to light.

“ep. 25.—A larva of this species was placed on a glass

plate, upon which rested a microscope slide. When the

larva in its crawling reached the slide, it came to rest in

the angle formed by the glass plate and the edge of the

slide. It remained there many minutes until disturbed.

Gravity. Esp. 17—An adult female was placed in a

glass tube which was enclosed in a black sleeve to exclude

light, and the tube placed in a vertical position. The posi-

tion of the thrips was marked with a wax pencil before

putting on the sleeve. The sleeve was then removed mo-

mentarily at frequent intervals, and the position and

direction of crawling of the insect noted. Most fre-

quently it was found lower than the previous position,

168 THE AMERICAN NATURALIST [Vow. XLVIII

and crawling downward. This was not always the case,

however.

Of other specimens tried, some showed positive geo-

tropism more definitely, some less definitely than the one

described. None showed a negative reaction in the ma-

jority of cases.

Anaphothrips striatus

Light. Exps. 5 and 7.—Adults of this species were

shaken out on a sheet of white paper near a window, and

the course of their crawling was plotted as accurately as

possible in my notes. Some individuals were decidedly

negative to light, crawling directly away from the window

every time they were tried, regardless of the direction in

which they happened to be headed when they touched the

paper. Others were indifferent to light, crawling in vari-

ous directions. Most of the males used were decidedly

negative to light, females usually indifferent.

Exp. 10—Females taken from the exposed portions of

leaves of Spartina michauxiana, and tested as above,

were found in nearly every case to be indifferent to light.

Females from the curled young leaves of the same plants

were as a rule negative to light.

Exp. 6.—Larve were usually found indifferent to light,

regardless of whether they came from exposed or con-

cealed situations.

Exp. 15.—A single larva taken from the exposed part

of a leaf, when placed in a glass tube one end of which

was directed toward the window, crawled steadily toward

the window. When the position of the tube was reversed,

the larva at once reversed its direction. The tube was

then placed in a black sleeve to exclude the light, and kept

there for an hour. When it was removed, the larva

showed for some minutes a decidedly negative reaction to

light. Later, however, its behavior became indefinite,

and soon became markedly positive. Darkness had ap-

parently temporarily reversed its reaction.

Contact. Exp. 22.—A female of this species which was

No. 567] BIOLOGY OF THE THYSANOPTERA 169

negative to light was placed on a sheet of blotting paper.

A small square of glass was placed over her, and sup-

ported at one edge, so that in crawling away from the

window the thrips approached the edge of the glass which

was in contact with the paper. She soon became lightly

wedged between the glass and the blotter, and came to

rest. Blotter, thrips and glass were then carefully

turned through 180 degrees so that the negative reaction

to light would have led the thrips out of its crevice; but

she remained there for a long time. Positive reaction to

contact overcame the negative reaction to light.

Another female, indifferent to light, was placed under

a similar glass. In her random crawling she became

wedged between the blotter and glass, and, notwithstand-

ing that the blotter was occasionally turned in the mean-

time, remained there several hours, until I lifted the glass.

Another female, not negative to light, was placed under

a similar glass square. She crawled from under it, but

happened to crawl against the edge of the microscope

slide that supported the glass cover. She settled quickly

into the right angle formed by the slide and the blotter,

and remained there a long time.

Gravity. Exp. 21—A female which was indifferent to

light was placed in a glass tube, and the tube set in a

vertical position. The thrips immediately began to crawl

downward. The tube was reversed, and the thrips im-

mediately reversed its direction. A sleeve was placed

over the tube to exclude the light, and frequently removed

temporarily to observe the position of the thrips. In

every case she was found crawling downward.

When the tube was held in an oblique position, the re-

sult was the same; the thrips crawled down the slope. If

she was already crawling down, a slope of 5 to 10 degrees

was found to be sufficient to keep her going in the same

direction. But to reverse the direction of crawling, it

was necessary to create a slope of about 45 degrees in the

opposite direction. The same positive geotropism was

Shown when the thrips was placed on an inclined sheet of

170 THE AMERICAN NATURALIST [Vou. XLVIII

paper; but being here at liberty to fly, she soon inter-

rupted the experiment.

Numerous other females were tried, and all showed

positive geotropism, some more promptly than others,

but all perfectly definitely. A single male tested showed

no definite reaction to gravity. A larva, nearly full

grown, subjected to the same tests, showed as definite a

positive reaction to gravity as did any of the females.

With the possible exception of the males, therefore,

Anaphothrips striatus is decidedly positive to gravity.

Anthothrips verbasci

Light. Exp. 4.—Adults of this species, shaken out on

a paper near a window, crawled in various directions.

None of them showed any definite reaction to light.

Numerous larve, none of them over three fourths

grown, crawled directly away from the window in every

instance.

Exp. 12.—In this experiment adults from concealed

places in mullein spikes were compared with those from

exposed situations. They were shaken out on a sheet of

paper near a window, and the direction of crawling noted.

In every case, those from concealed situations showed a

fairly definite negative reaction to light. Of those from

exposed situations, two were plainly negative, the re-

maining ten indifferent to light.

Exp. 11—Larve taken from concealment in a mullein

spike were tested, on a sheet of paper, for their reaction

to light. Those of the smaller sizes crawled directly away

from the window. Those nearly full grown, while on the

whole negative, crawled in a more or less devious path

away from the window. One reddish larva, which from

its color and size must have been nearly ready to pupate,

was especially indefinite in its reaction to light.

Contact. Exp. 18.—Larve of various sizes, which were

found to be negative to light, were placed on a blotter

under a square of glass supported at one edge, as de-

seribed for Anaphothrips striatus. When, in crawling

No. 567] BIOLOGY OF THE THYSANOPTERA 171

away from the window, they became wedged lightly be-

tween glass and blotter, and came to rest, the blotter with

all on it was turned through 180 degrees. The larve

turned their bodies so that their heads were directed away

from the window, but did not crawl away. The positive

reaction to contact overcame the negative response to

light.

An adult tested in the same manner as the larve above

described did not come to rest under the glass square.

But happening to crawl against the microscope slide

which supported the glass, the thrips came to rest in the

right angle formed by the blotter and the edge of the

slide, and remained there a long time.

Gravity. Exps. 13 and 20.—Adults and larve were

put, one at a time, into a glass tube, which was set in a

vertical position, and covered with a black sleeve to ex-

clude light. Some were examined at frequent intervals,

others were left half an hour without examination. In

every case the thrips were found almost precisely where

they were put at the beginning of the experiment. This

species is therefore indifferent to gravity.

Anthothrips niger

Light. Exp. 2—The red larve of this species were

shaken out on a paper near a window, as described in

other experiments. In every case the larva crawled away

from the window for a few seconds at first, then slowly

turned toward the window, and continued indefinitely

toward the light. Once while the larva was crawling

toward the light, I tapped the paper vigorously with a

pencil, so that the thrips was lifted slightly from the

paper and let drop; it immediately reversed its direction,

crawling from the window, but in a few seconds turned

again toward the light. The paper was jarred frequently,

but always with the same result. To show whether the

jarring made the response to light negative, or merely

reversed whatever the larva was doing at the instant, the

tapping was repeated at intervals of one or two seconds.

172 ‘THE AMERICAN NATURALIST [Vou. XLVIII

At the first tap, the larva, which had been crawling

toward the window, immediately turned away from the

light. Before it resumed its positive response to the light,

the paper was tapped again; the negative response con-

tinued. In this way the larva could be kept crawling away

from the light indefinitely. Disturbance makes the reac-

tion of the larva to light temporarily negative; otherwise

it 1s positive.

SUMMARY OF EXPERIMENTS

Euthrips tritici, when disturbed, is positively photo-

tropic in both larval and adult stages. It is positively

stereotropic, and the stereotropism is stronger than pko-

totropism, at least under certain circumstances. Some

individuals appear to be on the whole positively geo-

tropic; others are indifferent.

Anaphothrips striatus Adult males are usually nega-

tively phototropic. Females taken from exposed situa-

tions are usually indifferent to light, those from concealed

situations usually negative. The larve are usually in-

different to light, regardless of the kind of place from

which they are taken; a single larva that was positive was

made negative by keeping it in the dark. Adults are posi-

tively stereotropic. The females and larve are positively

geotropic.

Anthothrips verbasci—Adults taken from concealed

situations are usually negatively phototropic, those from

exposed places tend to be indifferent to light. The larve

are all negatively phototropic, except the full-grown ones,

which may be indifferent. The larve are plainly posi-

tively stereotropic, the adults less plainly so, or not at all.

Neither adult nor larva responds to gravity.

INTERPRETATION OF THE EXPERIMENTS IN THEIR RELATION

TO DISTRIBUTION AND BEHAVIOR oF THRIPS IN NATURE

With the evidence from these experiments before us,

may we not interpret the observed distribution and be-

havior of the Thysanoptera in nature somewhat as fol-

lows? Instead of explaining the fact that Euthrips tritici

No. 567] BIOLOGY OF THE THYSANOPTERA 173

always lives in concealed situations as due to a demand

for protection, we may assume that it is due to the strong

positive stereotropism of this species—aided in some

cases by positive geotropism, where the flower inhabited

is upright, but notwithstanding positive geotropism

where the flower is inverted. The rapid escape by crawl-

ing or flight when disturbed is not due to the fact that

this is the best way of avoiding danger, but to the posi-

tive reaction to light. Other species avoid danger by

going deeper into crevices, because they are negatively

responsive to light.

Anaphothrips striatus lives on grasses doubtless be-

cause it can not live on any other food, or because the

reproductive processes are not stimulated by any other

host plant. But their distribution and behavior on the

grasses may be explained largely in terms of their reac-

tions to the three agents tested in the experiments. The

males usually live in concealed situations on the plants

(curled-up leaves) because they are mostly negatively

phototropic, and crawl down the leaves until they reach

these concealed situations. Females may live either in

exposed or in concealed places, for some of them are

negative to light, others indifferent. The larvæ are either

exposed or concealed, because they are indifferent to

light. The eggs from which they hatch are probably laid

by negatively phototropic females in the young curled

leaves, and the leaves unfold as the larvæ develop; this

explains why the exposed larvæ are much larger, on the

average, than are those concealed in the young leaves.

Perhaps the relation of cause and effect as here stated is

reversed, at least for some cases. Concealment—caused

in one way or another—may lead to negative phototro-

pism, as in the larva which was made temporarily nega-

tively phototropic by being kept in the dark.

The adults are lodged between the ridges on the upper

side of the leaves of the grass Spartina, not for the sake

of protection, it seems to me, but because they are posi-

tively stereotropic. Doubtless between the ridges is the

174 THE AMERICAN NATURALIST [Vou. XLVI

place where they can best suck the juices of the plant, but

there is no need to assume that they deliberately choose

this location in order to get their food most easily. Both

adults and larve rest on these leaves with their heads

directed toward the base of the leaf, and crawl toward

the base of the leaf if disturbed, not because protection is

most quickly to be found among the curled leaves at the

center of the plant, but because the thrips are positively

geotropic.

Anthothrips verbasci—The larve of this species live

hidden among the flowers of the mullein spike, not be-

cause they must get their food there, for they can get it

from any part of the plant; nor do they hide there, it

seems to me, to secure protection. They remain in these

crevices because, excepting the largest larve, they are

positively stereotropic and negatively phototropic. The

adults are sometimes exposed, sometimes concealed, prob-

ably because in the former case they are usually indiffer-

ent to light, in the latter case negatively phototropic. (Or

may they be made negative or indifferent according as

they live—for one reason or another—concealed or ex-

posed?) ,

Thus, while Anthothrips verbasci is limited to one food

plant, and the food requirements are therefore probably

exceedingly important, yet the distribution and behavior

of the insects on this plant may be explained without ap-

pealing to anything like ‘‘choice’’ in other matters.

Regarding Anthothrips niger, I wish to call attention

to but one fact. The difficulty with which the larve are

driven forth from a flower in which they live appears to

be due, not to a persistent attempt at concealment, but to

the fact that on being disturbed they are temporarily

negatively phototropic; if the disturbance is continued,

the negative response continues.

The only argument which, it appears to me, could be

advanced in favor of assuming ‘that the Thysanoptera

choose their locations, instead of adopting simple re-

sponse to external stimuli as the correct explanation of

No. 567] BIOLOGY OF THE THYSANOPTERA 175

distribution, is the possibility that they have learned that

certain modes of behavior are best suited (for example)

to continued safety.

The reply to such an argument is first, that most of my

studies on behavior have been made in an unsettled re-

gion, where the enemies of thrips incident to civilization

are practically wanting, and where even the natural

enemies are not abundant. It could hardly be assumed

that every individual would learn to avoid its enemies in

the course of its short lifetime, yet certain species seem

to be invariable in their response to certain agents.

Furthermore, many of the larve tested in the experi-

ments could have been but a few days old. It is incred-

ible that their reactions should have been, as in fact they

were, as definite and invariable as those of older larve,

if these responses were dependent on experience.

It seems to me, therefore, that the only satisfactory ex-

planation of outdoor behavior and distribution of the

Thysanoptera lies in the assumption that they are in

large measure the result of responses to simple stimuli,

and do not imply any degree of choice.

ORIGIN AND ÅDAPTIVENESS or RESPONSES TO EXTERNAL

STIMULI

The origin of such responses in Thysanoptera as have

been described above is not, I believe, discoverable. Pur-

poseful they most probably are not, as I have shown, if

by purpose we mean conscious direction of actions to

some end. But adaptive they no doubt are in many cases.

Perhaps they are all adaptive, but I confess that my

powers of analysis are not keen enough to prove such a

view correct. That Euthrips tritici is positively photo-

tropic when disturbed is no doubt the cause of frequent

escapes from danger. One may even believe the negative

Phototropism of larvæ of Anthothrips verbasci to be

adaptive, because they are much more sluggish than is

Euthrips tritici, and could not escape quickly even if they

should emerge into the light. They are probably safest,

176 THE AMERICAN NATURALIST [Vou. XLVIII

therefore, if, when disturbed, they retire into still deeper

crevices. But I am unable to discover the adaptiveness

of the response of the larve of Anthothrips niger to light

—at first negative, on being disturbed, but soon becoming

positive. Nor can I understand why the males of Ana-

phothrips striatus are more definitely negative to light

than are the females or larve. These reactions seem to

me to be useless.

We need not demand that all of these responses be

adaptive, any more than that they be purposeful. Re-

sponses have arisen, no one knows how. They have been

preserved, and we can but speculate as to the method of

their preservation. Natural selection may be respon-

sible for the preservation of the useful, and it may have

eliminated responses that were harmful. But other re-

sponses of no value whatever, but likewise harmless, may

have been allowed to persist, without help or hindrance

from selection.

(To be continued.)

SHORTER ARTICLES AND CORRESPONDENCE

THE ENDEMIC MAMMALS OF THE BRITISH ISLANDS

WHEN, in 1891, I was collecting information to be used by

Dr. A. R. Wallace in preparing the second edition of his ‘‘ Island

Life,” I found much skepticism among naturalists concerning

the alleged endemic or precinctive elements of the British fauna.

Dr. Wallace was able to give lists of supposed precinctive species

and varieties belonging to several groups, but for the mammals

he was obliged to state, ‘‘it is the opinion of the best authorities

that we possess neither a distinct species nor distinguishable

variety.” -We little imagined that about twenty years later the

British Museum would issue a work describing ten species and

twenty subspecies of mammals peculiar to the British Islands;

twenty-one of these being actually undescribed at the time I

made my enquiries, and the rest then reposing quietly in the

Synonymy. Still less did we imagine that such a revision, when

made, would be the work of an American, coming over from the

United States National Museum to show Europeans the neglected

wonders of their own fauna! The Catalogue of the Mammals of

Western Europe, by Mr. G. S. Miller, published last year by the

British Museum, is certainly one of the most remarkable zoolog-

ical works ever produced, and is well worthy of the attention of

all naturalists, whether specially interested in the Mammalia

or not. While so many students of genetics are giving us the

results of their experiments in breeding mammals, it is worth

while to turn also to the results of nature’s long-time breeding

experiments, so clearly set forth by Mr. Miller in the volume cited.

What, after all, is the connection between the phenomena seen by

the breeder and the facts of mammalian evolution? Do species

and subspecies differ by ‘‘units,’’ and do the variations observed

in captivity correspond in any way to the recorded specifie and

subspecifie differences?

A complete analysis of Mr. Miller’s volume can not be made at

the present time, but I have extracted the list, given below, of

the forms supposed to be confined to the British Islands, giving

their distribution and principal. distinctive characters. I have

added to Mr. Miller’s list three quite recently described animals.

On examining the list, it appears that a few of the species must

belong to the older fauna of the country, not wholly exterminated

by the glacial ice and periods of partial submergence. Such are

177

178 THE AMERICAN NATURALIST [Vou XLVIII

Mustela hibernica of Ireland and Microtus orcadensis of the

Orkney Islands. It is at least suggestive, in this connection, that

so many of the Scottish islands yield animals differing from

those of the mainland. In the majority of cases, however, the

peculiar British mammals are closely related to those of the con-

tinent, and might well be of very recent origin. There is a

decided tendency to darker colors, such as has been noted also

among British moths. In spite of this tendency, however, some

forms are lighter than their relatives, the most conspicuous case

being the light-tailed British squirrel. In several cases the differ-

ence noted has in part to do with particular phases; thus the

squirrel has no dark phase, and the ermine does not turn so white

in winter. The British red grouse, it will be remembered, is

peculiar in lacking a white winter phase. Some of these differ-

ences may be due to the direct effect of the mild and moist

British climate, and would perhaps disappear in the descendants

of British animals taken elsewhere. The experiments on birds by

Beebe are very suggestive in this connection. In other cases, the

distinctions are such as might readily result from changes in one

or two ‘‘units,’’ such as are observed in experimental breeding.

When we have a variable type, subject to losses and new combi-

nations of unit characters, it is perhaps to be expected that

different groups of individuals, isolated from one another, will

after a time produce different homozygous combinations. That

is to say, the result comes from a long series of ‘‘accidents,”’

which will probably not be duplicated in two different places. In

this way mere isolation may be an adequate cause of modification,

providing always that through variation degrees of hetero-

zygosity have arisen.

In the common house mouse, Mus musculus, Hagedoorn’ has

isolated and figured a great number of color varieties, for nearly

all of which he has constructed zygotic formule. Little? has

also described and figured a similar series of varieties, appar-

ently in ignorance of Hagedoorn’s paper, which he does not cite.

He gives zygotic formule for thirty-two different varieties, but

not all of them are visibly different. Albino varieties, resulting

from the dropping out of a particular determiner, may be pro-

duced, corresponding in other respects to each of the thirty-two

colored forms, although they all look alike, and will only show

their true characters on crossing. Several of the varieties show

1 Zeit. f. ind. Abst. Ver., 1912.

2 ‘í Experimental Studies of the Inheritance of Color in Mice,’’ 1913.

No. 567]. SHORTER ARTICLES AND CORRESPONDENCE 179

noteworthy fluctuating variability, due to differences in ex-

pression,

Mus musculus, then, is very conspicuously variable in color;

yet Miller’s book records only one subspecies, that of the Medi-

terranean region and the Azores, which is less dusky and more

yellowish, with the under parts buffy grayish. It possibly agrees

with Little’s ‘‘dilute black agouti’’ variety. On the other hand,

M. musculus has a recognized subspecies in Mexico, where it

must have developed since the species was introduced by man.

The mice of St. Kilda and the Faroe Islands, although given as

distinct species, are derivatives of Mus musculus, differing in

other points than color. In connection with subspecifie differ-

ences in size, Sumner’s experiments with different temperatures

should be noted, since they prove that differences of temperature

might lead to readily measurable differences in dimensions,

wholly unconnected with losses of determiners or new zygotic

combinations. Whether or not diverse conditions of this sort

would ultimately affect the germ plasm, their effects would be

patent long before and quite independently of any such modifi-

cation. On the whole, the poverty of Mus musculus in subspecies

would suggest that the variations observed by breeders are not,

as a rule, the stuff that new subspecies are made of. Against this

argument may well be adduced the fact that M. musculus is an

urban animal, constantly traveling about, so that incipient races

do not remain isolated. Here the closely related rats, Epimys,

are worth considering. For Europe Miller can only recognize

the Norway, Black and Alexandrian rats, all widespread, prac-

tically cosmopolitan. Yet in the Malay Archipelago, where

Epimys is distributed over myriads of islands, large and small,

the species are innumerable. One can almost take a map and

indicate where new species of Epimys are to be found, namely, on

those islands still’ unexplored. Years ago, when the writer was

actively engaged in studying the British Mollusca and Lepi-

doptera, the question of endemic forms was constantly in mind;

but in those days we failed to discriminate properly between the

different classes of ‘‘varieties.’? We made the mistake of looking

for well-marked “sports” or aberrations, rather than for con-

stant but only slightly distinguished local races. There was a

practical reason for this, in the fact that by searching the litera-

ture we could ascertain whether a well-marked variation had

been reported from the continent; whereas the determination of

subspecifie types analogous to those described by Miller among

180 THE AMERICAN NATURALIST [Vou XLVIII

mammals required long series from different parts of Europe,

and these we did not possess, and could not readily obtain, Miller,

following the custom of mammalogists, lays great stress on sub-

species, but almost ignores individual variations, except such as

are expressed by the statistical data regarding size. By reading

the synonymy, one can see that many such variations have re-

ceived names, and I can not doubt that the time will come when

these names will be generally used. In this case, it will be

extremely desirable to use the same adjectival name for analogous

varieties of different species, and beyond the limits of subspecies

it ought not to be held that a name once used in a genus can not

be employed again. It may be true that most or all of the ‘‘indi-

vidual’’ varieties can be expressed by zygotic formule, but one

can not remember all these formule, nor use them in speech with

any comfort. Moreover, they have to do with the germinal con-

stitution rather than the patent characters. Little provides all

his varieties with polynomial English appellations, but would not

Latin varietal names be better? Following his theory con-

cerning the pigments, some of the varieties receive names

which do not suggest the animals at all; thus ‘‘brown-eyed

yellow,’’ according to the apparently excellent colored plate,

is light orange-ferruginous, while ‘‘sooty-yellow’’ is dark gray

with yellowish under parts. Morgan? describes a wild variety of

M. musculus from Colorado, which he calls ‘‘mauve,’’ but from

the detailed account it is rather ‘‘fauve,’’ namely, fulvous or

yellowish brown. It must be similar to the Old World subspecies

azoricus, or possibly that subspecies introduced? If we had

standard scientific names for the different forms, we should try

to compare our specimens with the types or descriptions of those

names, and it would not be left to authors to use such miscellane-

ous descriptive terms as might occur to them. For Mus musculus,

possibly Little’s apparently excellent colored plates might be

made the standards for a series of names. Thus his Fig. 9

(pl. 3) is the animal named niger as long ago as 1801; Fig. 10,

the dilute black, would naturally take the name nigrescens.

Fig. 12 is probably albicans of Billberg, 1827.

MAMMALS PECULIAR TO THE BRITISH ISLANDS

Insectivora

Sorex araneus castaneus (Jenyns 1838). Great Britain. Not so dark as true

ANEUS.

8 Ann. N. Y. Acad. Sci., XXI, p. 106.

No. 567] SHORTER ARTICLES AND CORRESPONDENCE 181

Sorex granti (Barrett-Hamilton and Hinton 1913). Inner Hebrides. Dif-

“7 from araneus by the contrast between bright-colored flanks and

upper parts; teeth also different.

N ick fodiens bicolor (Shaw 1791). Great Britain. Under parts usually

washed with wood-brown instead of buffy whitish; skull smaller.

Chiroptera

Rhinolophus ferrwm-equinum insulanus Barrett-Hamilton 1910. Central and

. Engl Wing shorter.

Rhinolophus hipposideros minutus (Montagu 1808). England and Ireland.

Wing shorter.

Carnivora

Mustela erminea stabilis (Barrett-Hamilton 1904). Mainland of Great

Britain, er large, with large teeth; color somewhat different,

little darker above. Change to white in winter less complete and asl

n in continental forms.

Mustela erminea ricine (Miller 1907). Islands of Islay and Jura, Scot-

land. Smaller than stabilis; proportions of skull different.

Mustela hibernica (Thomas and Barrett-Hamilton 1895). Ireland and Isle

of Man. Quite distinct; recognized by combination of black-tipped,

heavily penciled tail with entirely dark ear and upper lip. Superficially

like certain North American forms.

Felis sylvestris grampia (Miller 1907). Scotland; formerly throughout

Great Britain. Darker, with more pronounced black markings.

Rodentia

Lepus europeus occidentalis de Winton 1898. England, Scotland and Isle

Lepus timidus scoticus (Hilzheimer 1906). Highlands of Scotland.

Lepus hibernicus Bell 1837. Ireland. Distinguished by the strongly russet

aae ra partial or complete absence of white winter coat.

Scot

Rootomye dini Barrett-Hamilton and Hinton 1913. Island of Mull,

Hebrides

Evotomys glareolus t britannicus (Miller 1900). Great Britain. Smaller;

color less inten

Evotomys Shiimmenkals Barrett-Hamilton 1903. Skomer Island, off coast of

Wales. Color above unusually light and bright; skull peculiar.

Microtus agrestis exsul Miller 1908. North and South Uist, Hebrides. Re-

sembles true agrestis of Scandinavia; teeth peculiar, a character usually

present which elsewhere in the species occurs as a rather rare anomaly.

Microtus agrestis macgillivrait Barrett-Hamilton and Hinton 1913. Island

of Islay, Hebrides.

Microtus agrestis hirtus (Bellamy 1839). England and South Scotland.

Smaller than typical agrestis; upper pe noticeably tinged with rus-

Set, and venter washed with wood-bro

Microtus agrestis neglectus (Jenyns oes Highlands of Scotland. Not

So small as hirtus; upper parts darker.

182 THE AMERICAN NATURALIST [Vou. XLVIII

Microtus orcadensis Millais 1904. South Orkney Islands. Related to M.

sarnius of Guernsey and the Pleistocene M. corneri of South England.

Distinguished by its large pn and dark color.

Microtus — (Millais 1905). Sanday Island, N. Orkney group.

Alli o orcadensis, but skull differing; upper parts much lighter.

Microtus Pree westre Miller 1908. Westray Island, N. Orkney group.

Not so pale as in typical form; teeth differing a little.

Arvicola re (L. 1758). Typical Pagans England and South

Scotla Large; color moderately dar

Arvicola einiiibie ater (Macgillivray TEN = reta Miller 1910. Scot-

land, except southward. Darker, melanism frequent. The name was

changed on account of Hypudæus terrestris var. ater Billberg 1827, but

the change is perhaps needless, as Billberg’s animal was not a sub-

Tae and has not been treated as a species or subspecies under

pines ayer (de Winton 1895). Lewis and Barra islands, Hebri-

des. Large, with small ears; color dark.

Apodemus hirtensis (Barrett-Hamilton 1899). a of St. Kilda, Near

hebridensis, but skull larger and color dar

Apodemus fridariensis (Kinnear 1906). Fair isle, Shetland group. Large;

skull peculiar; colors also somewhat peculia

p reaR flavicollis wintoni (Barrett- Hamilton 1900), England. Under

parts with duller color, pectoral spot more diffus

Mus muralis Barrett-Hamilton 1899. Island of St. Kilda. Like M. musculus

but feet and tail less slender; skull peculiar.

Mus feroensis (Clarke 1904). Faroe Islands. Larger than musculus and

muralis; hind foot very robust; tail ape ed.

Sciurus vulgaris leucourus Kerr 1792. eat Britain and Ireland. Small;

tail drab, fading in summer to cream "buf. No dark phas

Ungulata

Cervus elaphus scoticus Lönnberg 1906. Great Britain. Color darker and

less gray than in the related Norwegian form

Capreolus ko thotti Lönnberg. 1916. Great Britain. Darker, face

darker than body.

I thought it of interest to compare the above British list with

a similar one for the Spanish peninsula (Spain and Portugal).

The latter area is continuous northward with France, but the

Pyrenees constitute a barrier. The Iberian peninsula differs so

much in its recent geological history from Britain, and is at the

same time so much more southern, that we should expect to find

the faunal elements very different. This expectation is realized,

yet the difference in numbers between the two lists is not very

great, and the number of Iberian forms treated as distinct

species is exactly the same (12) as that for the British Islands.

This suprising result is evidently due to the numerous sm

No. 567] SHORTER ARTICLES AND CORRESPONDENCE

183

islands of the British group, such islands being wanting around

the coasts of Spain.

MAMMALS PECULIAR TO THE SPANISH (IBERIAN) PENINSULA

Ins

Talpa cst "(Gabe

Galem pyrenaicus pa putes

Granth j.

Sorex araneus granarius Miller.

Neomys anomalus Cabr.

Crocidura mimula cantabra Sensi

Crocidura russula cintre Mille

Erinaceus europeus adinera B.-

Chiroptera

rianensis (Graells.).

Martes biog mediterranea

(Also Balearic Is

Mustela putorius aureolus (B.-Ham.).

Mungos ose! i (Gray).

Genetta genetta (L.), typical subsp.

Felis akak tartessia (Miller).

Lynx pardellus Miller.

(B.-Ham.).

Mite nivalis iberica (B.-Ham.).

Rodentia

Lepus granatensis Rosenb.

(Also Balearic Is.).

Lepus granatensis gallæcius Miler.

UNIVERSITY OF COLORADO

Eliomys lusitanicus Memen

Glis glis pyrenaicus C

Microtus piisa rogianus ke cd

Microtus asturia

Arvicola sa thes Miller, typical

subsp.

Pitymys lusitanicus (Gerbe).

Pitymys marie (Major).

Pitymys pelandonius Miller.

Pitymys depressus Miller.

Pitymys ibericus (Gerbe), typical

subsp.

Pitymys ibericus centralis Miller.

Mus spicilegus hispanicus Miller.

Sciurus vulgaris numantius Miller.

Sciurus vulgaris infuscatus (Cabr.).

Sciurus vulgaris segure Miller.

Sciurus vulgaris beticus (Cabr.).

Ungulata

Sus scrofa castilianus Thomas,

Sus scrofa beticus Thomas.

Cervus elaphus hispanicus a

Capreolus capreolus can

Capra pyrenaica lusitanica (Paina):

apra pyrenaica victorie Cabr.

Capra pyrenaica hispanica (Schimp.).

Rupicapra parva (Cabr.).

T. D. A. COCKERELL

LITERATURE CITED

Bateson, W. Mendel’s Principles of Heredity Cambridge (England)

tdia rsity Press. 1909.

Castle, W. E.

396 pp.

Heredity of Coat papet in Guinea-pigs and Rabbits.

Publ. Carnegie Inst. of Wash. No. 2

T L. La loi de Mendel et M a la EEE chez les

ouris. 4me note.

Darhishire, A, D.

Arch, Zool. exp. et gén. Not

Notes on the Results of Crossing gk Waltzing

Mice with European Albino Races.

et Revue, 1905.

Biometrika, Vol. 2, p. 101, 1902.

184 ' THE AMERICAN NATURALIST [Vou. XLVIII

Doncaster, L. On the psa of Coat Colour in Rats. Proc. Camb.

Phil. Soc., Vol. 12, pt. 4, p. 215, 1905.

Durham, F. M. A Pilimi Account of s Inheritance of Coat Colors

in Mice. Rept. Evol. C’t’ee. Roy. Soc., IV, 1908.

Hagedoorn, re L. The Genetic Factors in ae Devan of the House-

- mouse which Influence the Coat Color, with Notes on Such Factors in

the aie of Other Rodents. Zeit. fiir indukt. Abst. u. Vererb.,

Bd. 6, pp. 97-136, 1912.

poi E and Castle, W. E. Selection and Crossbreeding in Relation to

Tikin of Coat-pigments and Coat-patterns in Rats and Mice.

rad Carnegie Inst. of Was 1907.

Morgan, T. H. Recent Wicpetinante on he Inheritance of Coat Color in

Mice. Am. Nart., Vol. 43, pp. 494-510, 1909.

NOTES AND LITERATURE

SWINGLE! ON VARIATION IN F, CITRUS HYBRIDS

AND THE THEORY OF ZYGOTAXIS

SWINGLE in two recent papers has published some very inter-

esting observations on Citrus species and their F, hybrids. On

the basis of these observations, the somewhat startling statement

is made that current theories of heredity and variation give no

adequate explanation of variability in F, hybrid generations

from ‘‘pure bred” parent strains. Swingle assumes this vari-

ability to be so great that qualitative differences in chromosomes

can not account for it. As the chromosomes in the F, hybrid

remain unfused until synapsis, there is said to be no opportunity

for quantitative exchange of hereditary substance, so that this

variation can not be accounted for on this basis. Hence,

if proof can be given to show that in certain specific cases, pairs of

gametes of identical hereditary composition? give rise to very diverse

organisms, the way has been opened for a general reinvestigation of the

validity of our modern theories of heredity.

The term ‘‘pure bred’’ as used by Swingle implies that cer-

tain Citrus species reproduce themselves in a relatively faithful

manner from seed, there being no overlapping of distinguishing

Specific characters and very little variation of these characters

intraspecifically. C. aurantium and C. trifoliata are examples

of such widely separated species. The former has been grown

from seed in Florida for two hundred years, and though varia-

tions have appeared, they are said to differ but little from the

general type of C. aurantium, and in no way to approximate

that of C. trifoliata.

On the basis of evidence of this kind, Swingle believes the

various Citrus species (C. aurantium, C. trifoliata, C. medica

limonum, ete.) breed true in nearly all their characters and

especially in those which differentiate them from one another.

Hence, for genetic studies, the germ cells of these species are

t Swingle, W, T., ‘Variation in First Generation Hybrids (Imperfect

minance): Its Possible Explanation through Zygotaxis,’’? IV° Conf. In-

ternat. de Genetique, Paris, 1911, pp. 381-394; ‘‘Some New Citrus Fruits,”

Amer, Breed. Mag., 4: 83-95, 1913.

? The italics are my own.

185

186 THE AMERICAN NATURALIST (Vor. XLVIII

assumed, in respect to these differential characters, to be pure;

or, expressed in more technical language, each species is for the

characters under observation, genotypically homozygous. This

assumption is based on wholly inadequate evidence, as will be

shown later.

Citrus trifoliata crossed with other Citrus species (C. auran-

tium, ete.) gave F, hybrid families showing a large degree of

variability, even when the seeds from a single cross having

identical male and female parents were grown. This variability

expressed itself in foliage, habit of growth, and fruit, and was

especially noticeable in the latter, the fruits of the F, individuals

showing differences in color, size, texture, shape, number of seeds,

and flavor. For example, from a single cross of C. trifoliata X C.

aurantium, the 11 resulting hybrid seeds gave rise to F, plants

(citranges) differing in foliage, habit of growth, and very strik-

ingly in fruit. The fruit of one of these citranges, the ‘‘ Morton,”’

was smooth, round, very large, and orange-colored; those of the

**Colman’’ were rather flattened, globose, pubescent, yellow, al-

most seedless, and lacked the disagreeable oil common to the

others; while those of still another type, the ‘‘ Willits,” were

often monstrously fingered. The ‘‘Phelps’’ was bitter, while the

**Saunders’’ almost lacked this quality. The ‘‘Rustic’’ often has

double fruits with many seeds, and a habit of growth more like

its aurantium parent.

When varieties of the lemon were crossed with C. trifoliata,

still greater differences in the F, generation (citremons) resulted.

These consisted largely of ‘‘abnormal’’ foliage developments.

Hypophylls, though absent in the common Citrus species are ex-

tremely characteristic of C. trifoliata. About 20 per cent. of the

lemon-trifoliata hybrids developed an intensified form of this

character, and this proportion occurred in each case in crosses

involving three different varieties of lemon. The tangerine

orange X grape fruit (tangelo) in the F, generation was almost

as variable as the citrange families. F, hybrids between the

West Indian lime and the kumquat (limequat) were strikingly

different in such characters as aroma, flavor, acidity of pulp and

thickness of skin.

Although much stress has been laid on the differences in these

F, hybrids, there were numerous similarities. For example, all

the Citrus hybrids involving C. trifoliata in their parentage have

compound, semi-evergreen leaves, increased hardiness and fruits

No. 567] SHORTER ARTICLES AND CORRESPONDENCE 187

with abundant bitterish, acid. juice: Two of the citranges (Cal

man and Cunningham) have the pubescent fruit character of

C. trifoliata, while the others are smooth-skinned.

The author’s data led him to formulate in substance the follow-

ing conclusions, which I have grouped and stated in my own

language.

1. Citrus species are but slightly variable in the characters

which differentiate them, and, in the sense that no overlapping

takes place, may be said to breed true, their germ cells being

genetically pure for these differential characters.

2. Individual plants of the F, hybrid generations between these

species are strikingly variable, although all are, in a given cross,

the zygotic product of pairs of gametes of ‘‘identical hereditary

composition.’’

3. Modern theories of heredity can not account for this varia-

tion. ;

These are not the conclusions, however, in which all present-

day geneticists would concur. In the first place, few ‘‘modern’”’

geneticists would take Swingle’s view concerning the ‘“‘pure

breeding’’ ability of the various Citrus species, nor even of C.

aurantium. Webber, in the Encyclopedia of American Horti-

culture, notes that 70 varieties of the common sweet orange are

grown within our borders, and although a few varieties are

fairly constant, the majority of these do not breed true from seed.

Practically the same idea has been gained by certain prominent

taxonomists of the genus Citrus. De Candolle specifically calls

attention to the remarkable variability of the whole group; and

Professor Hume of Florida remarks on the same fact in certain

Experiment Station publications. As to the variability among

the individuals in the special strains used by Swingle in his breed-

ing work, no data are given, so that it can not be affirmed that

inbred progeny from them would have been duplicates as far as

hereditary characters are concerned. Citrus plants naturally

cross fertilize, and from this cause alone no dependence can be

placed on their ability to produce progeny, which are exact dupli-

cates of themselves when inbred; in fact, the inference is that

they would not. “Hence, as far as intraspecific constancy of

hereditary characters is concerned, Swingle’s statement can not

be accepted until more exact information is produced.

Swingle says no interspecific gradations occur between these

various species, especially C. trifoliata and C. aurantium. Grant-

188 THE AMERICAN NATURALIST [Vou. XLVIII

ing this, the two species have clearcut differences in leaves (ever-

green or deciduous, unifoliolate or compound), in resistance to

cold (difference in ability to withstand certain degrees of tem-

perature) and in numerous fruit characters (presence or absence

of pubescence, quality of juice, quantity of seed, size of fruit,

ete.).

From the standpoint of modern theories of heredity as regards

variation in F, hybrid generations, it matters little whether so-

called species intergrade or whether their differences are clear-cut

and all variation is intraspecific. In either case, if crosses were

made, variation among the F, individuals from a single family

might or might not occur. In either case, no violence to modern

theories of heredity would result and no new problems would

arise. But if two species that differ from each other in part or

all of their characters, but breed true intra-specifically (geno-

typically homozygous) are crossed, and F, variation results, then

modern theories of heredity would be compelled to change front

and invoke the aid of new hypotheses. Swingle’s data, assuming

that intraspecific variation in Citrus species occurs, does not

present a problem of this kind at all. C. aurantium and C. tri-

foliata each possess distinctive characters, but convincing data are

not at hand to warrant any belief in the homozygosity of these

differential characters or of even those the two species may have

in common. The evidence directly, and one might almost say

conclusively, opposes such a conclusion. If these species are not

homozygous in all of their characters, then one can not affirm, in

the light of modern theories, that all the gametes produced by a

particular group of individuals called a species are identical in

hereditary composition, nor even that the gametes of one indi-

vidual of such a species are identical as to hereditary potenti-

alities. At the risk of wasting valuable space by repeating what

is extremely common knowledge to genetice students, let us assume,

for the purpose of argument, that C. aurantium and C. trifoliata

are homozygous in all their respective characters except one. In

the former, the character A is heterozygous and peculiar to this

species, Likewise, in C. trifoliata, B is heterozygous and differ-

ential. All the remaining characters of the two species may be

symbolized, respectively, by the formule XX and YY. When

XX Aabb (C. aurantium) is crossed with YYaaBd (C. trifoliata),

the resulting progeny would appear in the approximate propor-

tion of 1XYAaBb:1XYAabb:1XYaaBb:1XYaabb, providing

No, 567] NOTES AND LITERATURE 189

A and B are single factor characters. In the majority of char-

acters, the F, hybrids would be intermediate or possess those of

either one or the other parent, since all the F, individuals would

be alike as far as any hereditary quality symbolized by XY is con-

cerned, providing the plants were all grown under the same en-

vironmental conditions. But these F, individuals would not be

alike as regards the inheritance of the characters A and B. Ex-

perimental evidence from crosses of this kind show us that four

different F, forms may result, the distinctions between them aris-

ing from the presence or absence, through inheritance, of the

characters A and B. Dominance is assumed to be absent in this

illustration.

Swingle’s Citrus hybrids, though involving greater complexity

because a large number of parental characters instead of two are

probably heterozygous, are of the same general type as those of

the illustration and lend themselves to the same interpretation.

Owing to the absence of sufficient exact experimental data, one

can not speak of unit characters and factors in these hybrids, but

one may say without violence to modern theories of heredity that

one or both of the parents involved in the crosses which produced

the Colman and the Cunningham were heterozygous in the factors

or factor for pubescence, that various size factors were hetero-

zygous and that one parent was homozygous for absence and one

for presence of the factors for hardiness, compound leaves and

evergreen foliage.

F, variation in Citrus hybrids then, in the light of the data at

hand, apparently results from differences in the gametie compo-

sition of the heterozygous parents.

Swingle calls attention to other cases of variation in F, hy-

brids from two pure stocks which support his contention that this

phenomenon of F, variation is very general, though usually

obscured through variation due to heterozygous parent stock.

Collins and Kempton? crossed a race of corn breeding true to

waxy endosperm with one constant for horny endosperm. Horny

endosperm was dominant in F, and the F, generation segregated

ìn the expected ratio of 1 waxy to 3 horny kernels. This ratio

represented the average proportion of each when the ears of all

the plants were lumped together. The F, progeny of each selfed

8 Collins, G, N., and Kempton, J., II, 1912, ‘‘ Inheritance of Waxy Endo-

sperm in Hybrids of Chinese Corn,’’ IV° Conf. Internat. de Genetique, 1911,

P. 347; also Cire, No. 120, Bur. of P, I., U. S. Dept. of Agr., 1913.

190 THE AMERICAN NATURALIST [Vow. XLVIII

F, plant when taken by itself gave some ears as low as 13.7 per

cent. waxy, while others exceeded the expected proportions and

gave ears as high as 33.3 per cent. waxy. The investigators point

out that this variation is not the result of the laws of chance as

the deviation is far greater in many cases than the probable error.

Therefore, says Swingle,

there can be no doubt but that their varying percentages represented

real differences in the hereditary composition of the first generation

plants. It would be hard to find a more conclusive case since there could

be no doubt as to the purity of the parents and what is more rare no

possible doubt as to whether a given kernel had a waxy or a horny

endosperm.

Mendelians are said to be unaware how fatal this phenomena is

to some of the chief tenets of modern theories of heredity, and

they are also accused, somewhat unjustly, I believe, of applying

the term ‘‘imperfect dominance’’ to this and to the Citrus

phenomena. :

In this case, both parents were undoubtedly homozygous for

their respective endosperm characters, so that heterozygosity will

not account satisfactorily for the deviations. But this is a dif-

ferent phenomena than Swingle found in his Citrus hybrids, for

here one is dealing with a fluctuation in a proportion or ratio

involving the same character, while in his experiments the diffi-

culty was the variation in presence and absence of distinct and

often new characters, indicating an extremely heterozygous

parentage.

As an explanation or working hypothesis for his own and

similar data, Swingle advances a somewhat new and suggestive

chromosome theory on the assumption that it fills an urgent need.

The theory of zygotaxis, as it is called, may be summarized as

follows:

Maternal and paternal chromosomes probably persist side by

side in the cells, unchanged in quality and number throughout

the whole development of the F, organism. This being true,

Swingle, in order to explain his data, assumes that the influence

in character formation exerted by chromosomes on the F, hybrids,

is in some cases due to their relative positions in the nucleus, and

that these relative positions result from accident or at least are

determined at the moment of nuclear fusion in fertilization, and

remain unchanged in succeeding cell generations. He further

No. 567] NOTES AND LITERATURE 191

assumes that those chromosomes lying nearest the nuclear wall

(peripheral) are better nourished than those centrally located,

and hence they exert more influence in character formation, and

dominating synapsis, produce gametes similar in their hereditary

character to the cells of the first generation hybrids, whose char-

acter in turn was determined at fertilization by the configuration

the chromosomes took in the fusion nucleus. On this theory,

reversions, sports, etc., may result from sudden changes in the

nuclear configuration.

Three types of nuclear configuration are assumed to occur in

higher organisms, the character and effects of which are synop-

tically outlined below.

1. Interspecific Hybrids—Usually sterile and intermediate.

Chromosomes repel each other and occupy opposite sides of the

F, zygote nuclei, exerting equal influence in the ontogeny of F,

organisms, explaining why first generation hybrids of this char-

acter are always intermediate, little variable and usually sterile.

Synapsis often impossible.

2. Mendelian Crosses—Abnormally inbred races of domesti-

cated animals and plants. F, generation usually intermediate,

fertile, dialytic at synapsis. Dominance of certain characters in

these hybrids is due to the inherited potentialities of the chromo-

somes rather than to their nuclear positions.

3. Normal Cross-bred Species—Probably normal in wild

Species. Hybrids usually vigorous, fertile, and variable. Free

intermingling of chromosomes in the fusion nucleus at fertiliza-

tion. Nuclear configuration permanent for each individual.

Synapsis normal.

This elaborate and attractive theory, based admittedly to a

great degree on assumptions, is advanced by Swingle in the belief

that it will help to clarify the problems of heredity, even though

he acknowledges it does not help one to arrive at satisfactory

explanations. In the reviewer’s opinion, however, the field of

genetics is already burdened with enough theories of this par-

ticular type and the somewhat unnecessary but ever-increasing

new additions serve to confuse rather than clarify the ideas of

the average student of genetics. Besides, Swingle’s assumption

that maternal and paternal chromosomes in the cells of F, hybrids

repel each other and do not mingle in the F, zygote cells is not

borne out by the few cytological facts at our command. Rosen-

192 THE AMERICAN NATURALIST [Vou. XLVIII

berg’s* work on species hybrids of Drosera, Moenkhaus’s* investi-

gations of species hybrids in fish and some work on certain

hybrids in the Echinodermata group give us facts that directly

oppose such an assumption. As a further criticism, one may say

that most biologists who have had experience with pedigree cul-

tures would decidedly criticize the synoptic outline and the nar-

row sphere assigned to Mendelian phenomena.

Aside from the theoretical considerations, these two papers con-

tain descriptions of Citrus-like species new to occidental horti-

culture, together with a somewhat detailed account of the various

Citrus hybrids and their hardiness and practical value, showing

the truly fine results achieved by the workers in this field toward

moving the Citrus belt northward and adding new varieties of

this genus to the world’s horticulture.

ORLAND E. WHITE

BROOKLYN BOTANIC GARDEN

December 4,

8 Rosenberg, O., ‘‘Cytologische und Morphologische Studien an Drosera

longifolia X D. rotundifolia,’’ Kungl. Svenska Vetenskapsakademiens Hand-

linger., 43, N: ou, pp. 1-64, 1909. 4 Tafn.

4 Moenkhaus, W. J., ‘‘The Development of the Hybrids between Fundulus

heteroclitus and Menidia notata with especial reference to the Behavior of

the Maternal and Paternal Chromatin,’’ Amer. Jour. of Anatomy, 3; 29-65,

1904. Plates I-IV.

VOL. XLVIII, NO. 568,“ APRIL, 1914

THE

AMERICAN

NATURALIST

A MONTHLY JOURNAL

Devoted to the Advancement of the Biological Sciences with

Special Reference to the Factors of Evolution

CONTENTS

Page

I. The Origin of X Capsella Bursa-pastoris arachnoidea. Dr. HENRI Hus - 193

- - 236

Dr. A. FRANKLIN SHULL

Biology of the Thysanoptera. II.

II. Shorter Articles and Discussion: Barriers as to Distribution as regards Birds

and Mammals. JOSEPH GRINNELL. Yellow Varieties of Rats. Professor

W. E. CASTLE soo ooo oo a G ee a a a

IV. Notes and Literature: Woods on Heredity and the “ Influence of Monarchs.”

M LK a8 wi i 8 ž es Sas = = a = - - 255

THE SCIENCE PRESS

LANCASTER, PA. GARRISON, N. Y.

NEW YORK: SUB-STATION 84

The American

Naturalist

S. intended for publication and books, etc., intended for review Cares be

sent to the Editer of THE AMERICAN NATURALIST, Garrison- -on-Hud yter y

Short maries of

articles containing sum

w York.

research work bear gp the

problems of organic evolution are peneers welcome, and will be given pe Paors

in pais atio

rea reprints of spider are supplied to authors free of charge.

e hundr

Further reprints will be supplied at cost

riptions and advertisements should be sent to the gma The

anadian postage twenty-five cents additional.

postage is fifty c and

eign ents

The cha mee for single aust is

forty cents. The advertising rates are Four Dollars for a pa

THE SCIENCE PRESS

NEW YORK: Sub-Station 84

Lancaster, Pa.

Entered as second-class matter, April 2, 1908, at the

Garrison, N. Y.

Post Office at Lancaster, Pa., under the Act of

Congress of March 3, 1879

FOR SALE

ARCTIC, eee nine GREENLAND

DS’ S

Well Herth ‘eo Prices

Particulars of

ESEN, Bird Collector

Husavik, North Iceland, Via Leidle, England

WANTED TO PURCHASE

a set of BIRDS OF AMERICA by J. J. Audubon,

7 or 8 volumes, please report, stating cash price, stat-

ing condition, binding and dates of volumes.

F. C. HARRIS,

Box 2244 Boston, Massachusetts

TO OOLOGISTS

and ENTOMOLOGISTS

W. F. H. ROSENBERG

Importer of Exotic Zoological Specimens

57, Haverstock Hill, London, N. W., England

Begs to announce the publication of a new Price

List (No. 20) of et Sg — k over 400 species

from various parts of the world.

These will be mailed free on application, as well

as any of the following lists:

Birdskins (over 5000 species); Birds’ Eggs (over

1,100 species) ; Reptiles, Batrachians and Fishes

(over Sg species) ; Exotic Lepidoptera (over 8,000

species

Largest stock in the world of specimens in all

branches of Zoology.

All Museums and Amateurs

should write for these Lists

All specimens sent on approval.

Please state which lists are required and give

name of this periodical.

Photographic Films Developed

5 cents per Spool, Prints 3 cents up, Post

Cards 5 cents.

8x10 Enlargements 25 cents mounted.

Send for Samples and Price List.

FRANK R. BARBEAU y

89 West Bridge Street Oswego, N. Y.

The Sge of Chicago

Arts, Literature,

Commerce and phe #

Law, Medicine, Educa-

. Instructs

embers of the

4 tion,

Summer enas, 1914

15-July 22

oo Tem July 23-Aug. 28

announcements will be

sent upon application.

The University of Chicago

Illinois

(L Mitchell Tower Chicago Lo n

THE

AMERICAN NATURALIST

Vor. XLVIII April, 1914 No. 568

THE ORIGIN OF x CAPSELLA BURSA-PASTORIS

` ARACHNOIDEA

DR. HENRI HUS

UNIVERSITY OF MICHIGAN

Sivce Jordan! described a number of elementary spe-

cies of Capsella Bursa-pastoris, their constancy has been

a subject of cultural experiment. Herbarium material

demonstrates the existence of numerous apparently unde-

scribed forms. The finding of strikingly distinct forms,

such as Capsella Heegeri? and, more recently, C. Viguieri,?

the work of Almquist‘ and that of Shull have added to

the interest which this species holds for the investigator.

It was Shull who determined the zygotic constitution of

various forms. To be able to demonstrate this with ex-

actitude is of the greatest value since Bateson and Lotsy

expressed their doubt as to the homozygocity of de Vries’s

nothera Lamarckiana. It was left to Nilsson® to clearly

Show its necessarily heterozygous character. The inter-

est aroused by this paper® leads me to believe that an

1 Jordan, A., ‘ Diagnoses d’espéces nouvelles ou méconnues pour servir

de matériaux à une flore réformée de la France et des contrées voisines.’’

Paris, 1864.

_*Solms-Laubach, H. Graf zu, ‘‘Cruciferen studien. I. Capsella heegeri,

eme neuentstandene Form der deutschen Flora,’’ Bot. Zeit., 55: 167, pl.

7, 1900,

*Blaringhem, L., ‘‘Les transformations brusques des êtres vivants.’’

Paris, 1911.

*Almquist, E., ‘Studien über die Capsella Bursa-pastoris (L.),’’ Acta

Horti Bergiani, 4: No. 6, 1907.

5 Heribert-Nilsson, N., ‘Die Variabilität der @inothera Lamarckiana und

das Problem der Mutation,’’ Zeitschr. f. ind. Abst. u. Vererb., 8: 89, 1912.

* Lotsy, J. P., ‘‘ Fortschritte unserer Anschauungen über Deszendenz seit

Darwin und der jetzige Standpunkt der Frage,’’ Progressus Rei Botanicæ,

4: 361, 1913.

193

194 THE AMERICAN NATURALIST [Vou. XLVIII

account of certain cultures of Capsella, in which muta-

tions were simulated, would be of timely interest.

During the winter of 1908-1909, I collected in a green-

house at Ann Arbor, Michigan, and at the disposal of the

Botanical Department of the University of Michigan,

twelve rosets of Capsella Bursa-pastoris, the leaves of

which showed certain more or less striking morphological

differences. With the hope of isolating certain biotypes,

the rosets were placed in pots and permitted to flower.

No measures were taken to prevent the accidental trans-

ference of pollen, but the pots were placed about six

inches apart. This, as will be shown later, is the only

precaution necessary to guard against cross-pollination,

provided the cultures are carried on in a greenhouse and

during the winter months. After a portion of the seed

had ripened, the plants, the majority of which retained

their climax leaves, became herbarium specimens. More

recently, after constant association has enabled me to

detect minute differences, it has been possible to identify

some of these plants with two of the biotypes described

by Shull,” to wit, rhomboidea and simplex. At the time

of collection, the differences were sensed, but could not be

described technically, since the extent of the influence

wielded by fluctuating variability was an unknown quan-

tity. Never before had I so fully realized the truth of de

Vries’s statement.’

We are trained to the appreciation of the differentiating marks of

systematic species. ... Our minds are turned from the delicately

shaded features which differentiate elementary species.

The seed obtained was sown in sterilized soil during

the spring of 1910. From each seedpan 60 individuals

were transplanted to flats. As the plants grew older, it

was found that, with a single exception, the seedlings in

each of the flats were uniform, but that the seedlings in

the different flats were not alike, three types being dis-

tinguishable. The interest in these types, for the isola-

7 Shull, G. H., ‘‘ Bursa bursa-pastoris and Bursa Heegeri: Biotypes and

Hybrids,’’ Publ. No. 112, Carnegie Institution of Washington, 1909.

8 de Vries, Hugo, ‘‘Species and Varieties,’’ 689, 1905.

No. 568] X CAPSELLA BURSA-PASTORIS ARACHNOIDEA 195

tion of which these cultures had been undertaken, soon

yas overshadowed by the behavior of the seedlings bear-

ing the number 4,108.6 and which were the offspring

yielded by a plant of a type not described by Shull and

which I have named X Capsella Bursa-pastoris Setchelli-

ana, in honor of Professor William Albert Setchell.

FIG. 1. APPEARANCE OF A LINEAR-LEAVED FORM AMONG SEEDLINGS OF Capsella

Bursa-pastoris.

During the time that the seedlings remained in the seed-

pan, no deviations from the expected course of develop-

ment were noted. However, after the seedlings had been

transplanted to flats and had remained there a week or

So, it became evident that some of the seedlings were not

making the expected growth. Their development ap-

peared most insignificant compared with that of the

majority. A closer examination showed the cotyledons

to be somewhat larger than normal and the leaves proper

to be exceedingly small and almost linear. Nor did they

attain the same length as the leaves of the rosets belong-

ing to other types.

196 THE AMERICAN NATURALIST [Vou. XLVIII

An explanation of this peculiar development was sought

in a possible attack on the part of either fungi or bacteria

or in soil conditions. But the latter were uniform for the

entire flat. Neither fungi nor bacteria could be demon-

strated nor did the underground portion of the ‘‘ab-

normal’’ plants look unhealthy or underdeveloped.

Fic. 2. SEEDLINGS or æ C.... Setchelli AND « C.... arachnoides

At this stage the flat presented the appearance shown

in Fig. 1. At the time but three types were distinguished,

the first of these constituted by plants which showed an

incision of the blade, the second composed of those which

apparently had entire leaves, and a third, comprising the

small and linear-leaved rosets, which, because of the spider-

like appearance of the latter, has been designated X Cap-

sella Bursa-pastoris arachnoidea. There also appeared a

No. 568] X CAPSELLA BURSA-PASTORIS ARACHNOIDEA 197

single individual which, while closely resembling the form

arachnoidea, differed from it in having somewhat spatu-

late leaves. This plant, a plant of arachnoidea and two

of Setchelliana, are shown in Fig. 2.

After photographs had been taken, the plants were

potted and placed in the frames. None of the plants

made a growth as vigorous as that of the Capsellas grow-

ing in the open. The plants of the form arachnoidea de-

veloped leaves with a greatest length of 15 mm. and a

greatest width of a little over 1 mm., causing the plant to

retain its spider-like appearance. The roset with spatu-

late leaves appeared somewhat more vigorous, the aver-

age leaf measuring 22 mm. in length, with a greatest width

of 2.5mm. In later generations I have been able to ob-

tain rosets of arachnoidea with a greatest leaf-length of

100 mm. and a greatest width of 6 mm.

In the frames, flowering shoots made their appearance,

those on arachnoidea being remarkable chiefly because of

their small size, reaching a length not exceeding 12 em.

The flowers were small but well-formed. No well-devel-

oped pollen could be demonstrated. Seed did not form

and the capsules retained their original form, typical of

non-fertile capsules in Capsella Bursa-pastoris, remind-

ing one of the capsules of Capsella Heegeri. They do not

resemble the fertile capsules of C. procumbens. In the

next generation I saw a single capsule formed on arach-

noidea as the result of cross-fertilization, and in this case

it differed in no manner from the normal capsule such as

we know it in Capsella Bursa-pastoris.

The ‘‘normal’’ plants, i. e., all those not belonging to

the form arachnoidea, matured a large amount of seed.

No measures were taken to prevent cross-pollination, but

no other plant of Capsella Bursa-pastoris, within a radius

of twenty feet, was in flower.

At this time, another attempt was made to group the

plants. It was found that the criterion used earlier, i. e.,

the incision of the blade, no longer could be relied upon,

Since plants, which at the time of the previous count, had

Shown an entire margin, now were more or less incised.

198 THE AMERICAN NATURALIST [Vou. XLVIII

Unfortunately, after the seed had been collected, the plants

were destroyed, having lost their climax leaves. An attempt

to group them later withtheaid of photographs failed, be-

cause photographs of all plants were taken during the

earlier stages only, i. e., before the appearance of the cli-

max leaves. Another classification, for which climax leaves

are not essential, and which is based upon the relative

width of the first six or eight leaves, yields for 54 plants

the proportion: ‘‘wide’’ 31, ‘‘narrow’’ 16, ‘‘linear’’ 7, the

ideal proportion, as since worked out, being 33:16: 16.

The fact that the number for ‘‘linear,’’? which represents

the form arachnoidea, is too small by 9, may be ascribed

to various circumstances, among others the fact that the

last row in the flat did not appear in the photograph upon

which the count was based. It is in the last row of a flat

one ordinarily meets with the smaller or at least less vig-

orous individuals and it is very probable that in this last

row occurred a large percentage of individuals belonging

to arachnoidea. Furthermore, not all the seedlings, but

only sixty, were taken in each case. Almost unconsciously

one selects the largest individuals when transplanting

from seedpan to flat. It is probable that in this process

there were eliminated a greater percentage of seedlings

of the linear form than of any of the others. Hence no

great weight can be attached to the proportion obtained.

The collection of seed brought the work for 1910 to a

close. As far as I was aware, no forms similar to arach-

noidea had been either noted or described by any one who

had devoted his time to culture experiments with Cap-

sella. Neither Shull in America, nor Almquist in Swe-

den, nor Lotsy® in Holland, has made mention of such

forms in their publications. The fact that no seed was

produced by the aberrant form seemed to hold out little

hope for the continuation of the cultures, and the sole

trace left by this new form, if taxonomic form it was,

threatened to consist of but a few photographs and some

aleohol specimens. A single possibility presented itself.

9 Lotsy, J. P., ‘‘Vorlesungen über Deszendenztheorien,’’ 1: 180, Jena,

1906.

No. 568] X CAPSELLA BURSA-PASTORIS ARACHNOIDEA 199

Whether the parent plant was of a hybrid character or

whether the parent plant was mutating, and the new form

or forms were to be looked upon as mutants, in either

case there existed the possibility, if not the probability,

that from the seeds obtained from those plants of the

second generation which appeared ‘‘normal,’’ a third gen-

eration might be obtained which would again present the

abnormal form. Such indeed proved to be the case.

3 Ur

Fic. 3. EARLY STAGES IN THE DEVELOPMENT OF BROAD-LEAVED, NARROW-LEAVED

AND LINEAR-LEAVED FORMS OF Capsella.

The seed for the next generation was obtained from 19

plants. The seed was sown separately in pots of steril-

ized soil. Certain of the parent plants, which we now

identify with Shull’s simplex and rhomboidea, produced

a uniform, broad-leaved offspring. Others behaved like

the parent, the form arachnoidea appearing in 197 indi-

viduals out of a total of 979, which does not include the

713 which bred true to the broad type. (For an illustra-

tion of these types see Fig. 3.)

It is unnecessary to go into details as to the various

theories which suggested themselves as a solution of the

200 THE AMERICAN NATURALIST [VoL. XLVIII

origin of the linear-leaved form which, because of its

striking appearance, concentrated the attention upon it-

self. That perhaps we were dealing with a mutation was

a thought which most naturally obtruded itself upon

the mind of one who, for years, had fruitlessly tested

a large number of species in the hope of discovering a

case analogous to that of @nothera Lamarckiana.’® The

possibility of a cross between a local form and either

Capsella Heegeri or C. procumbens, suggested itself.

However, the seedling stage of either of these two forms

does not bear the remotest resemblance to that of Cap-

sella arachnoidea. At the same time there was slight

reason for believing that either Capsella Heegeri or Cap-

sella procumbens ever had been grown in Ann Arbor.

During 1911 and the greater part of 1912, the problem

rested here, no satisfactory explanation being found.

But pedigree cultures were continued until, on the one

hand, we succeeded in placing the plants in optimum sur-

roundings for the production of climax leaves, and on the

other began to distinguish between the various biotypes.

THe BIOTYPES

As has been noted previously, it was possible to use two

criteria for the classification of the rosets. Leaving out

of consideration the rosets of the linear-leaved arach-

noidea, it was found that after dividing the rosets accord-

ing to the ‘‘broad’’ or ‘‘narrow”’ character of the earlier

leaves (Fig. 3), it was possible to further subdivide each

group on the basis of the marginal indentation of the

leaves subsequently formed.

I. The “Broad”? Group.—Here the first four or five

leaves possess a blade which is approximately twice as

long as broad. Up to this stage the margin remains

entire. When the sixth leaf appears one ordinarily can

begin to distinguish between two types. These are:

Type 1.—In this, the first of the two broad-leaved

forms, the margin of the first eight leaves remains entire,

10 Hus, H., ‘‘The Origin of Species in Nature,’’ AMERICAN NATURALIST,

45: 646, Nov., 1911.

No. 568] X CAPSELLA BURSA-PASTORIS ARACHNOIDEA 201

showing at most a very slight crenation (Fig. 4). Usually

the ninth leaf, though sometimes it is the eighth and some-

times the tenth, shows a more marked indentation, though

seldom of a depth of more than 2 mm. on each side of

the leaf and slightly below the middle. Subsequent

leaves show an increase in the number and depth of the

FIG, 4. DISSECTION OF XOVRG Rosers or C. . . . simpler AND Ọ. . . . rhom-

boidea, ‘Suowixe THE ‘*‘ BROAD” Cniniceen OF THE Reena Leaves

AND THE DISTINCTIVE CHARACTER OF THE FIRST SIN

indentations, the maximum for both being reached in the

climax leaves. which usually show five indentations reach-

ing about midway from margin to midrib. In those of

the earlier leaves which show a marked incision the lobes

are obtuse. In the later leaves the lobes become acute.

It may be stated as a general truth, that an increase in

the depth of the sinus carries with it an increase in sharp-

ness of the lobe. There is no secondary lobing, but some-

times the margin of the sinus shows a slight denticulation.

While in the earlier leaves the sinuses separating the

terminal lobe from the rest of the blade are the deepest,

202 THE AMERICAN NATURALIST [Vou. XLVIII

the converse is true in the later leaves, where the sinuses

separating the terminal lobe are the most shallow. I

have identified this form with Shull’s simplex! My

plants also agree fairly well with the illustration of onto-

genetic succession of leaf forms in Bursa ... simplex,

shown by Shull.!2

Type 2.—In the second of the two forms distinguished

because of the greater relative width of their first leaves,

the margin of the first five leaves remains entire, as in

the case of those of type 1 (simplex). The sixth leaf,

however, ordinarily shows a marked indentation, at least

3 mm. deep and slightly below the middle of the blade

(Fig. 4). This indentation may appear in one margin or

in both. The lower margin of the sinus ordinarily is at

right angles to the midrib, the upper margin making an

angle of 45 degrees with the midrib (Fig.7,b). Even when

it has become difficult to distinguish between types on the

basis of relative width of the earlier roset leaves, italways

is possible to distinguish between type 2 (rhomboidea)

and type 4 (Setchelliana and Treleaseana), by means of

the character of the sinus. In type 4, the lower margin

of the sinus makes an angle of 45 degrees with the midrib,

while the upper margin makes an angle of between 30

and 45 degrees with the midrib. Hence the first sinus in

C.... Setchelliana and C. ... Treleaseana is at least

90 degrees, while the first sinus in rhomboidea measures

seldom more than 45 degrees and frequently less.

The seventh leaf of plants belonging to type 2 ordi-

narily shows two indentations on both sides of the leaf,

dividing the blade into a lower portion, two central lobes

and a terminal lobe. The depth of the incision amounts

to about three-fourths of the width of the blade from mid-

rib to margin.

It is possible to delay the appearance of the first inden-

tations by transplanting from seedpan to flat either too

early or too late. In such cases, the indentations appear

in the seventh leaf only, or even later, and are rather

11 Loc. cit., 25, and Pl. 2, Fig. 2.

12 Shull, G. H., ‘‘ Verh. d. naturf. Ver. in Briinn,’’ 49, Pl. 4, 1911.

No. 568] X CAPSELLA BURSA-PASTORIS ARACHNOIDEA 203

shallow, reaching a depth of three-fourths of the width of

the leaf from margin to midrib in the eighth, ninth or

tenth leaf. However, once the indentations have made

their appearance, the leaf next produced ordinarily shows

two sinuses on both sides of the blade, usually the upper

set, rarely the lower, being the deeper of the two, and

almost reaching the midrib. The succeeding leaves show

an increase in the number of lateral lobes from two to

six. Since the incisions almost, if not quite, reach the

midrib, both lateral lobes and the terminal lobes are well

defined. Upon the lateral lobes secondary lobes appear, `

both on the distal and proximal margins. It is to be

noted that only the climax leaves of well-grown specimens

of the homozygotic form distinctly show the lobing of the

proximal margin and this only on the middle lobes. The

lobing of the primary lobes results in the setting off of

a small terminal portion of each lateral lobe, which

possesses a more or less rhomboidal form. This terminal

lobe of the primary lobe can be observed to advantage

only in the climax leaves of well-developed specimens.

I have no hesitation in identifying type 2 with Shull’s

rhomboidea.18

Capsella Bursa-pastoris simplex and C. Bursa-pastoris

rhomboidea, described, respectively, as types 1 and 2,

agree in having the first five or six leaves twice as long as

broad, thus contrasting sharply with the plants to be de-

scribed under types 3 and 4, which constitute the ‘‘nar-

row’’ group.

II. The ‘‘Narrow’’ Group—tIn the plants belonging

here, the first five or six leaves possess a blade which is

from 2} to 3 times as long as broad. Usually after the

appearance of the seventh leaf, sometimes not until the

appearance of the tenth leaf, it is possible, on the basis of

marginal indentation, to separate the plants with ‘‘nar-

row” roset-leaves into two groups, designated respec-

tively types 3 and 4.

Type 3.—Rosets of plants belonging to type 3 can not

be distinguished from those of type 4, until after the

78 Shull, Verh., Pl. 2; Biotypes, Pl. 1, Fig. 2.

204 THE AMERICAN NATURALIST [Vou. XLVIII

seventh leaf has appeared (Fig. 5). Itis to be noted that

for the first six leaves of type 4, the ratio between mean

length and width is 6:2, while for the corresponding

leaves of type 3, the same ratio is 5:2. Once the seventh

—

Fic. 5. DISSECTION oF YOUNG RosETS OF @ C. .. . Setchelli AND g O. ...

ep pet SHOWING THE “ NARROW ” CHARACTER OF THE EARLIER

ES AND THE DISTINCTIVE CHARACTER OF THE FIRST SINUS.

leaf has appeared, a distinction readily can be made, since

in type 3, no sinuses appear, and the leaves, from the

seventh to the tenth, might be mistaken for those of

simplex (Fig. 5). Later leaves readily can be distin-

guished from those of simplex, by the pointed apex, the

very shallow sinuses, ending in a sharp tooth, and by the

fact that the greatest width of the blade lies above the

middle, about one third the length from the tip (Fig. 6).

This form, which because of its morphological charac-

ters on the one hand, and its behavior in breeding on the

other, can readily be distinguished from all others, I

designate X Capsella Bursa-pastoris attenuata.

Type 4.—Not only do the first leaves of plants, belong-

ing to this type, differ in relative width from the first

leaves of plants of rhomboidea and simplex, but there also

No. 568] X CAPSELLA BURSA-PASTORIS ARACHNOIDEA 205

is a difference in the apex of the leaf, the apices of leaves

of this type, like those of type 3, being decidedly pointed,

while those of types 1 and 2 are rounded."

At the sixth or seventh

leaf stage, the marginal

indentations make their

appearance, at first as

slight crenations, then as

long and shallow sinuses,

and finally, in the eighth

or ninth leaf, as a sinus

on one or both sides of

the midrib and about the

middle of the blade (Fig.

5). The lower margin of

the first sinus ordinarily

makes an angle of 45

degrees with the midrib,

while the upper margin

makes an angle of from

30 to 45 degrees with the

midrib. This renders the

first sinus ordinarily

greater than 90 degrees

(Fig. 7, a). The depth

of the first sinus is ap- Had con eel tee ee

proximately one half the... . attenuata axv © i

distance from margin to

midrib. In subsequent leaves the depth increases, so that

in the 11th leaf the sinuses almost reach the midrib. In

T'releasi, one of the two forms, which together constitute

type 4, the climax leaves show incisions to the midrib, and

a well-marked terminal lobe, while in the other the sinuses

are less deep but the terminal lobe still is well marked

(Fig. 8). The number of sinuses increases in propor-

14 It is to be noted that in my cultures there appear, from time to time,

plants of rhomboidea of which the leaves have sharply pointed lobes. What

relation these plants bear to others classed with them under rhomboidea, I

am at present unable to say.

206 THE AMERICAN NATURALIST (Vou. XLVIII

tion to their depth. If the seventh leaf has one sinus

in each margin, the eighth and ninth usually have two,

the tenth and eleventh, three, and so on, until the mean

of six is reached. As the lobes increase in number, they

Fic. T. EARLY ROSET LEAVES OF @ Fig. 8. CLIMAX meag OF &

. «s+ Setehelt ARD C. .; rhom- ares patchetii's AND & ory & & dh

boidea.

not only become narrower but the sinuses do likewise.

This is the result of a gradual increase in the angle

between the lower margin of the sinus and the midrib.

In the eighth leaf the lower margin forms an angle of

about 90 degrees with the midrib, causing the formation

of a primary lobe, triangular in shape and with an upper

angle of about 45 degrees, instead of the 90-degree angle

found in the first lobe. In older leaves the angle between

lower margin of sinus and midrib may increase to 110 or

even 120 degrees. The climax leaves therefore get to

resemble more and more those of rhomboidea, especially

since the distal margin of the sinus, from the tenth leaf

on, exhibits a number of denticulations which, in older

leaves, especially of one of the forms (Treleaseana),

tend to become incisions, so that secondary lobes are

No. 568] X CAPSELLA BURSA-PASTORIS ARACHNOIDEA 207

formed. However, the end of the lobes of early

leaves of type 4 always are sharply pointed (Fig. 9),

while the lobes of early leaves of rhomboidea are ordi-

narily rounded at the ends (Fig. 4).

RGH

{iei

FIG. ¢

hasidin ae

UPPER Row: 3 SETS or LEAVES FROM AS MANY PLANTS OF @ C....

Treleasi. Lowrer Row: 4 SETS OF LEAVES FROM AS MANY

LANTS OF g C. Setchelii.

From a morphological point of view these leaves are

entirely different from any form described by Shull, the

differences being most marked and very readily recog-

nized once our attention has been called tothem. But it

is especially the behavior of the plants on breeding which

leads me to recognize them as most distinct hybrid forms

and which,I have designated X Capsella Bursa-pastoris

Setchelliana in honor of Professor William Albert

Setchell, and X Capsella Bursa-pastoris Treleaseana, in

honor of Professor William Trelease.

Type 5—Capsella Bursa-pastoris arachnoidea. This

form, which readily is recognized from the first by its

linear leaves, does not require an elaborate description

at present, since it will be discussed in detail later. It has

been illustrated in Figs. 1, 2 and 3.

208 THE AMERICAN NATURALIST — [Vou. XLVIII

The above descriptions apply only to plants grown

under fairly uniform conditions, in a light soil in a green-

house, and treated in such a manner as to offer the plant

the most favorable conditions for development. By leav-

ing the plants too long in the flats, so that crowding re-

sults, by keeping them too moist and warm, etc., it is

possible to produce abnormal climax leaves in which the

typical differences can be recognized with difficulty only.

By leaving plants too long in the seedpans, by keeping

them too dry, it may be brought about that plants flower

without having produced climax leaves. There will be

doubtless many who, because of this, will refuse recogni-

tion to the segregates just described. ‘‘Quacunque dixi,

si placuerint, dictavit auditor.” Fortunately, the differ-

ences of behavior on breeding are such, we must recognize

their distinct genotypic constitution.

GENOTYPIC CONSTITUTIONS

_ Shull, in the papers above quoted, made one of the most

important of recent contributions to science, since he de-

termined with exactitude the relations existing between

some of the lesser forms which, because of their alleged

constancy or inconstancy, have been a bone of contention

since the days of Jacquin. Making extensive cultures of

Capsella, Shull was able to distinguish four forms (Fig.

10), to wit, heteris, with leaves divided to the midrib, with

Fic. 10. CLIMAX LEAves OF C. .. . heteris, C. . . . tenuis, C. . . . rhomboidea

AND 0. . , , simplex.

No. 568] X CAPSELLA BURSA-PASTORIS ARACHNOIDEA 209

elongated primary lobes, a marked secondary lobe, in the

distal axil of the primary lobe and a well-marked terminal

lobe; rhomboidea, with leaves divided to the midrib, with

an unelongated primary lobe, with an incision in the distal

margin setting off a secondary lobe and a corresponding

incision on the proximal margin of the primary lobe, set-

ting off, in well-grown specimens, a terminal portion of

each lateral lobe, generally of rhomboidal form; tenuis,

with the elongated primary lobe of heteris, but with a

sinus which usually does not reach the midrib, terminal

lobe clear cut; simplex, with lateral lobes obtuse, never

attenuated, the incisions being shallow and never reach-

ing the midrib.

Shull recognized here the presence and absence of two

factors, one (A) responsible for the sharp primary lobe

of heteris and the attenuation of the lobes in tenuis, while

the other (B) is responsible for the division of the leaf

to the midrib, the definite terminal lobe and the second-

ary lobes. On this basis Shull was able to represent the

biotypes by conventional Mendelian symbols, thus:

heteris, AB; rhomboidea, aB; tenuis, Ab; simplex, ab.

That this conventional presentation gives us a reliable

working basis, my experiments have shown most satis-

factorily. With the aid of these symbols I have been able

to solve the origin of Capsella arachnoidea, the experi-

ments showing that, without question, forms presenting

the spider-like appearance of the rosets typical of this

plant are of hybrid origin.

THE Zycotic Constirution oF 4,108.6

The problem to be solved was that of the zygotic con-

stitution of the original parent, the plant which in my

notes is recorded as 4,108.6. Among its offspring neither

heteris nor tenuis made their appearance, while both

rhomboidea (aB) and simplex (ab) were met with. Hence

the parent was homozygotie for (a), but heterozygotic

for (B). Therefore, its zygotic constitution, in part, must

have been aaBb.

Besides rhomboidea and simplex there appeared two

210 THE AMERICAN NATURALIST [Vou. XLVIII

forms, referred to as types 3 and 4, the latter being ca-

pable of further subdivision. Neither of these was de-

scribed by Shull. At least one difference between rhom-

boidea and simplex, on the one hand, and types 3 and 4, on

the other, could be noted at once, i. e., the relative width

of the leaf. As has been shown above, the former have

their first leaves twice as long as broad, the latter three

times as long as broad. The idea suggested itself that

there might exist a factor which determined these charac-

ters. Since the original parent belonged to type 4, the

narrow character of the earlier leaves must be dominant

over the broad character. Also, since the original parent

produced both ‘‘narrow’’ and ‘‘broad’’ types, it must

have been heterozygotic for this character. Using (N)

to indicate the gene, we get for the zygotic construction

of the parent plant aaBbNn.

aBN aBn abN abn

1 2 3 ar |

aBN aBn abN abn

“DNI BN | ON | bX | BN |

5 6 7 8 ||

aBN aBn abN abn

aBn aBn aBn aBn aBn

2) Cae 11 12

abN | aBN aBn abN abn

~ i) @bN abN abN abN

| 13 14 15 16 |

aba | aBN aBn abN abn

| abn abn abn abn

pea eee Sa }

Fig. 11. DIAGRAM TO ILLUSTRATE THE NATURE OF THE OFFSPRING OF @ C....

Setchelli (aaBbNn).

Since self-fertilization is the rule in Capsella, it was an

easy matter to test the validity of the theory. A form

aaBbNn, one with unelongated primary lobes, sinuses

reaching the midrib and with early leaves of a ‘‘narrow’’

type should yield, on self-fertilization, the following com-

binations: 1.bbnn (square 16), a plant of which, accord-

ing to our definition, the earlier roset leaves should be

No. 568] X CAPSELLA BURSA-PASTORIS ARACHNOIDEA 211

broad and of which the later leaves shall lack incisions

reaching to the midrib, a plant, in short, which should

have all the characteristics of Shull’s simplex. Further-

more, on being selfed, it should yield a uniform offspring,

in all respects resembling the parent.

Such plants actually were encountered. Of the plants

grown to maturity, twelve were selected as seed-bearers.

All bore the simplex character. Ten of these plants were

selected from among the first generation of plants of the

supposed zygotic constitution BbNn, while one parent

(yielding No. 25,712) was derived from a plant bearing

the simplex character and another (yielding No. 31,112)

was derived from a plant which was shown to have the

zygotic constitution bbNn.

TABLE I

EVIDENCE OF HomMozyGoTic CHARACTER OF Simplex (bbnn)

| | Character of

a, a tg Index Number of Parent Grand-

Parent parent

25,712 78 Gis 8,112BR12P9 bbnn bbnn

25,912 22 G 8,212BR3P1 bbnn BbNn

26,312 4 G 8,212CR5P1 bbnn BbNn

26,512 187 G 8,212F R3P3 bbnn BbNn

26,712 G 8,212HR7P7 bbnn BbNn

30,012 276 O 8,212CR5P1 bbnn BbNn

30,112 108 O 8,212H R2 bbnn BbNn

30,212 Oo 8,212GR6P8 bbnn BbNn

30,312 162 G ,212GR6P8 bbnn BbNn

30,712 oO 8,212HR3P6 bbnn BbNn

31,112 50 Oo 26,012AR7P6 bbnn bbNn

3,113 207 O 26,912BR1P3 _ bbnn | BbNn_

1,399 |

This table offers an excellent illustration of the small

danger of an accidental cross, even if the plants are not

guarded, always, of course, when the proper precautions,

indicated above, are taken. Numbers 26,312 and 30,012,

as well as numbers 30,212 and 30,312, respectively, offer

instances of uniform inheritance in plants possessing

recessive characters only and of which the parents in the

one case were left unguarded, in the other caged. Had

15 In this column ‘‘G’’ indicates that the parent plant was guarded,

“O” that the plant was open- -fertilized. In other tables the same abbre-

Viation will be used.

212 THE AMERICAN NATURALIST [Vou. XLVIII

crossing taken place in the case of the unguarded flowers,

this would, because of the purely recessive characters

possessed by simplex, have become apparent at once. In

all cases the parents were checked by means of herbarium

specimens or photographs, or both.

2. bbNN (square 11). According to our hypothesis, a

plant of this zygotic construction should have the earlier

roset leaves narrow and the climax leaves should lack

incisions to the midrib. It also should breed true. A

plant fulfilling these conditions has not been encountered,

or rather, its recognition was delayed until the offspring

of the corresponding heterozygote bbNn could be observed.

As will be shown, the zygotic combination bbNN yields

a plant with the external characteristics of arachnoidea.

3. bbNn (squares 12 and 15). A plant of this zygotic

constitution should have narrow early leaves and the

climax leaves should lack incisions to the midrib. On

self-fertilization it should yield 25 per cent. bbNN, 50

per cent. bbNn and 25 per cent. bbnn.

bN bn

bN bN

bN bn

bn bn

Several plants were found which fulfilled the require-

ments as to leaf characters. Such plants, on being selfed,

yielded approximately 25 per cent. simplex, which we

know to have the zygotic constitution bbnn, while about

50 per cent. bore the parental characters, supposedly rep-

resented by bbNn. The remaining 25 per cent. clearly

belonged to the type arachnoidea. In all, 12 plants were

selected as seed-bearers, some being guarded, others re-

maining uncaged. The results are given in Table II.

The totals closely approximate the Mendelian ratio,

yielding, respectively, bbNN 24 per cent., bbNn 49 per cent.

and bbnn 27 per cent. Having established the identity

of bbnn (simplex) and bbNn (attenuata), we are forced

to recognize bbNN as the zygotic construction of arach-

noidea. It would be a comparatively easy matter to test

No. 568] X CAPSELLA BURSA-PASTORIS ARACHNOIDEA 213

this directly, provided the form arachnoidea produced

seed. Though I have grown several hundreds of these

plants, I have obtained in all but eight seeds, and these as

the result of hybridization. Hence the test must be made

indirectly through crossing of forms yielding the desired

gametic combinations.

TABLE II

EVIDENCE OF HETEROZYGOTIC CHARACTER OF attenuata (bbNn)

Number of Plants |

| 7 bcs Char.

Index es ON N bb Nn bbnn Char. | G :

Num-

E Index Number A OP o

| -

w $ w v os 5 of Parent Parent |0 Gran “

a 'o fz] © =| 1o | paren

(de (3/48 | 2 | as |

m £ | = 2 es 2

26,012| 19 | 25.50| 56| 51.00| 27| 25.50 eel bbNn | G | BbNn

26,412| 8| 7.50} 14| 15.00] 8| 7.50 DR1OP8 | bbNn | G | BbNn

31,212| 16| 14.75| 27 | 29.50| 16 | 14.75 26'0124R2P1 bbNn | O | bbNn

31,312| 22 | 15.00) 19 f 30.00] 19 | 15.00 | 26,012AR1P2 | bbNn | O | bbNn

31,412) 10 | 14.25) 25 | 28.50| 22 | 14.25 | 26,012AR1P3 | bbNn | O | bbNn

| 47 | 51.00| 3 50 R

31,512) 24 | 25.50! 47| 51. 1 | 25.50 | 26,012AR1P4 |bbNn | O |bbNn

31,612; 2 .50 | j 4 0 | 26,012AR6 bbNn | O | bbNn

31,812} 26 | 28.25) 54| 56.50| 33 | 28.25 | 26,012BR1P3 | bbNn | O | bbNn

31,912| 11 | 16.50) .00| 15 | 16.50} 26,012BR1P6 |bbNn | O |bbNn

3,213| 61 | 64.25 130 |128.50| 66 25 | 26,912DR2P4 |bbNn | O | BbNn

3,313} 70 | 61.50| 117 |123.00| 59 | 61.50 | 26,912ER6P6 Nn | O | BbNn

3,513| 27 | 35.50! 71| 71.00| 34 | 35.50 | 26,912FR6P4 | bbNn |O | BbNn

Total .| 296 [308.50 | 604 (617.00 | 334 (308.50

Of the twelve parent plants concerned in the above ex-

periment, five were selected from among the first genera-

tion of a plant having the supposed zygotic constitution

BbNn, while seven were the direct offspring of No. 26,012,

which had been shown to yield the three forms, arach-

noidea, attenuata and simplex, as indicated in Table II.

The simplex, obtained by selfing a plant of bbNn, breeds

true, as indicated in Table I, No. 31,112, a simplex, yield-

ing a uniform simplex offspring, consisting of 50 indi-

viduals.

4. BBnn (square 6). A plant of this supposed zygotic

constitution should resemble, in all respects, Shull’s

rhomboidea, the earliest roset leaves being broad, and the

incisions of the climax leaves reaching the midrib. It-

Should breed true. Five lots, involving four parents, were

grown. Again it was shown, in the case of No. 26,812 and

214 THE AMERICAN NATURALIST [Vou. XLVIII

No. 30,612, that the fact that plants are left unguarded

does not affect results. The parents, in all cases, were

selected from among the first generation of plants having

the supposed zygotic constitution BbNn. The results are

given in Table ITI.

TABLE III

EVIDENCE OF HoMozyeoric CHARACTER OF rhomboidea (BBnn)

Index Number No. of Plants Index Number of Parent Char. of Parent Goro

25,812 20 8,212BR1P6 BBnn G

26,812 80 8,412BR3P2 BBnn G

27,012 6 8,412ER10P6 BBnn G

27,112 210 8,412FR13P12 BBnn G

30,612 96 8,412BR3P2 BBnn O

In all cases- the off-

spring was uniformily of

the rhomboidea character.

Bn Bn

Bn bn

bn bn

5. Bbnn (squares 8 and

14). Plants of this zygotic

constitution should resem-

ble those of the preceding

group, but on being selfed

should yield 25 per cent.

homozygotie rhomboidea

(BBnn), 50 per cent. het-

erozygotic rhomboidea

(Bbnn) and 25 per cent.

simplex (bbnn).

These three forms were

found to constitute the

Fic. 12. Crax Leaves or a Her- Offspring of a single plant,

omenie i p bei ee i San eee

spring of a plant of the

supposed zygotic constitution BbNn. This plant, from

16 Bursa... 39.

No. 568] X CAPSELLA BURSA-PASTORIS ARACHNOIDEA 215

the first, was classified as a rhomboidea. At the present

time, a photograph of the young roset confirms this

classification. But two climax leaves, which, in the

earlier part of these experiments, were deemed suff-

cient, show that the sinuses do not quite reach the midrib

(Fig. 12). Unfortunately, Shull, in the description of

his No. 054.28,% does not mention this point, though he

does point out that ‘‘the later rosette-leaves had some

of the secondary lobes acutish, but not elongated.’’? In

the older climax leaves, even of a homozygous rhom-

boidea, I find that the secondary lobes disappear. Shull,

in the description just referred to, is so specific as to the

typical rhomboidea character of the heterozygote that I

have hesitated to classify the heterozygotes and the homo-

zygotes. But the homozygotie rhomboidea, obtained as

the extracted recessive of a selfed plant of the supposed

zygotic constitution BBNn, always has sinuses which

reach the midrib. In other combinations, also, one can

distinguish between BB and Bb by the relative depth of

the sinus. For the present, then, we will rely upon this

character. In the case under discussion (26,612, the off-

spring of 8,212HR1P3, guarded) there were among the

39 plants 6 which clearly were simplex, the heterozygotic

rhomboidea was represented by 22 individuals, and the

homozygotic rhomboidea by 11 individuals, the calculated

ratio being 9.75:19.50:9.75. The percentage of simplex

is far too low, 15.4 per cent., instead of 25 per cent., but,

considering the small number of individuals concerned,

the total outcome is fairly satisfactory. It is almost un-

necessary to add that in this, as in other cases, the off-

spring of the various plants is being tested as fast as

time and facilities permit.

Type 4.—Having shown the presumable correctness of

our supposition as to the zygotic constitution of the initial

plant (BbNn), as far as the presence, appearance and be-

havior on breeding of simplex, rhomboidea and attenuata

are concerned, there remains to identify the major group of

combinations which, in a simple di-polyhybrid, constitutes

216 THE AMERICAN NATURALIST [Vou. XLVIII

nine sixteenths of the total offspring and may be uniform

in appearance, the constituents being separable only by

breeding, ‘‘eine heillose Arbeit,’? as Baur has it. For-

tunately, in this case, it is possible to distinguish readily

between the various combinations.

One of the combinations, BBNN (square 1), should

breed true, being homozygotic for both characters con-

cerned. We would expect such a plant to have narrow

first leaves and climax leaves with incisions to the mid-

rib. Thus far I have not encountered such a plant, some-

thing which at one time led me to consider the possibility

of gametic repulsion, in this instance the gamete BN

being incapable of existence. This supposition seemed

the more plausible since the two genes B and N well might

be supposed to be antagonistic, the one being responsible

for an incision of the leaf to the midrib, the other tending

to make the leaf, especially the earlier leaves, narrow.

Were this assumption correct, none of the zygotic combi-

nations found in squares 1, 2 and 5, 3 and 9, and 4 and 13,

would be formed, though we would expect the same com-

bination as occurs in squares 4 and 13 to make its appear-

ance as the result of the fusion of the gametes bN and

Bn (squares 7 and 10).

Were this supposition correct, we should have a case

similar to that of the sweet pea ‘‘ Purple Invincible,’’ and

we could not expect the gamete (bn) to be formed. Since,

however, simplex (bbnn) appears in our cultures, this

theory must be rejected. Recently also, in culture No.

30,412, an instance was found in which the guarded parent,

supposedly of type 4, yielded, not simplex, rhomboidea,

attenuata, arachnoidea as well as the parental type, but

only arachnoidea, rhomboidea and the parental type, and

in proportions closely approximating a ratio 1:1:2.

A plant which yielded 25 per cent. rhomboidea and no

simplex, must have been homozygotie for B, and since it

yielded also 50 per cent. of type 4, must have been hetero-

zygotic for N, its zygotic constitution therefore being

BBNn. Such a plant, on self-fertilization, should yield

25 per cent. rhomboidea. Provided the homozygote and

No. 568] X CAPSELLA BURSA-PASTORIS ARACHNOIDEA 217

the heterozygote have the same appearance, the remaining

75 per cent. should resemble tle parent (Fig. 9, b).

BN | Bn

BN | BN

| |

| BN | Bn

[2 Bae: St ae

But in one case (30,412), the parent being 8,412BR9P9,

and open fertilized, the offspring consisted of 26.3 per

cent. rhomboidea, 46.2 per cent. of the parental type and

27.5 per cent. arachnoidea. If our supposition as to the

zygotic constitution of the parent is correct, then the

zygotic constitution of the arachnoidea in this offspring

must be BBNN. In the case of a selfed attenuata, we

found that approximately 25 per cent. of the offspring

was composed of arachnoidea of the probable zygotic con-

stitution bbNN. Is it possible that any Capsella, homo-

zygotic for N, would have the appearance of arachnoidea?

This seems more than probable, and other evidence, to be

adduced later, appears to support this view. The history

of the BBNn is as follows:

During 1912 I grew No. 8,412 from seeds of a plant

which resembled the grandparent 4,108.6. It was com-

posed of 1,079 individuals, among which various types,

such as ‘‘broad,’’ ‘‘narrow’’ and ‘‘linear,’’ could be rec-

ognized. Not all plants were thus classified, a fourth

group of ‘‘intermediates’”’ being formed, indicating that

some of the plants, while in certain respects resembling

simplex and especially rhomboidea (deep lobing, second-

ary lobes), in other characters more closely approximated

the “narrows,” since their early leaves had been noted

as “‘narrow.’’ In the light of recent experience, it is easy

to see why the distinction was made, though at the time

the conception of the differences was most hazy. Several

of these ‘‘intermediates’’ were grown, and of these a

single one yielded the seed for the next generation. This

plant had been permitted to flower unguarded, but after a

number of capsules had developed on the main stalk, this

was decapitated and the sideshoots were allowed to de-

218 THE AMERICAN NATURALIST [Vou. XLVIII

velop. At this time the entire plant was caged. Subse-

quently the seeds of the open fertilized and of the guarded

flowers were sown separately, with the following results:

30,412. Open Fertilized 80,512. Guarded

Plants Plants

Per Cent. MOG GON toe re ee oS

i Found | Expected Found | Expected

Arachnoidea... . 27.5 40 | 36.25 21.15 52 | 61.50

Naw Nas 46.2. 67 | 72.50 36.15 89 | 123

Rhomboidea..... 26.3 38 | 36.25 42.70 105 | 61.50

The figures are given separately to again call attention

to the fact that open fertilization is no hindrance to pedi-

gree work in Capsella. Since the seeds came from the

same parent, we may add the results, which gives us

arachnoidea 23.50 per cent., ‘‘narrow’’ 40 per cent. and

rhomboidea 36.50 per cent. The fact that the percentage

for ‘‘narrow’’ is too low and that for rhomboidea too

high, while the percentage for arachnoidea is within the

limits of probable error, is probably due to errors in

classification, since greater weight was laid upon lobing

of the adult leaves than upon comparative width of the

earlier ones. The value of this culture lay chiefly in its

suggestion of a zygotic combination BBNn, which prior

to that time, on account of the gametic repulsion theory,

was not supposed to exist. In consequence, a number of

cultures were made, with the following result:

TABLE IV

EVIDENCE OF HETEROZYGOTIC CHARACTER OF Treleaseana (BBNn)

_ Number of Plants

BBNN BBNn BBnn Char.

Index Index No. of Char. | O of

No. Parent

Ex-

pected

Found

Ex-

pected

Found

Ex-

pected

38 6 26 | 30 30,412AR2P6 |B

3,913 | 36 | 44.75) 87 | 89.50 56 | 44.75) 30,412AR4P3 | BBNn

14 | 15.25} 28 | 30.50; 19 | 15.25) 30,412AR6P3 | BBNn

15 | 28.50! 65 34 | 28.50 | 30,412BR2P6 |B.

4,313 | 33 | 33 62 | 66 37 | 33 30,412BR6P5 | BBNn

4,413 | 37 | 45.25| 102 | 90.50; 42 | 45.25) 30,412BR9P2 | BBNn

Oooo

Total .| 173 [196.75 | 400 (393.50 | 214 (196.75

No. 568] X CAPSELLA BURSA-PASTORIS ARACHNOIDEA 219

The ‘‘narrows’’ in question, then, fulfilled our expecta-

tion on the basis of a zygotic constitution BBNn. In

some cases the percentages are too high, in others too low.

The total yields fairly satisfactory results, to wit: BBNN

22 per cent., BBNn 51 per cent. and BBnn 27 per cent.

Two tests of the extracted recessive, a homozygotic rhom-

boidea, were made. The cultures, No. 3,713, from a

guarded rhomboidea (30,412AR2P3) and No. 4,113, from

an unguarded rhomboidea (30,412AR8P3), both derived

from plants of the supposed zygotic. constitution BBNn,

_ yielded, respectively, 54 and 207 plants, all of which bore

the typical rhomboidea characters.

In the cultures just tabulated, the plants of the sup-

posed zygotic constitution BBNn resembled the parent in

all respects. The form arachnoidea, in this case, must

have the zygotic formula BBNN. Unfortunately, in this

case also, it proved unfertile.

A better acquaintance with plants of the zygotic consti-

tution BBNn led us to formulate certain differences be-

tween them and our original ‘‘narrow.’’ Plants of the

BBNn character, readily can be segregated from those of

the BbNn character by somewhat narrower primary

lobes, split to the midrib and the development, in climax

leaves of well-grown specimens, of a secondary lobe, not

pronounced but recognizable (Figs. 8, 9).

On the basis of these morphological differences, as well

as because of the behavior of the plant on breeding, I

propose to segregate it from type 4 under the name

X Capsella Bursa-pastoris Treleaseana. This form is

homozygotic for B, while Setchelliana is heterozygotic

for B. Both are heterozygotic for N. They may be ex-

pected to look alike during the early stages. Later they

Show a difference, since the form containing Bb does not

develop sinuses as deep as the form containing BB. The

form Treleaseana, when young, can readily be distin-

guished from a heterozygotic rhomboidea (Bbnn) by the

relative width of the early leaves; later such a distinction

is difficult (Figs. 4,5, 7). If any distinction at all is to

be made, it should be made on the basis of the rounding

220 THE AMERICAN NATURALIST [Vou. XLVIII

of the lobes, those of T’releaseana being sharp, those of

the heterozygotie rhomboidea rounded.

I am fully aware that in thus naming genotypes, I am

departing from all rules laid down by systematists. But

a rule is useful only as long as it serves a purpose. For

the geneticist, the rules of systematists are of small value.

Subspecies, variety, form, are, after all, but very general

terms, almost incapable of definition because of too fre-

quent abuse. But once we have determined the zygotic con-

stitution of any plant, we have placed ourselves on a firmer

basis. Behavior in breeding is the proper criterion. And

while I recognize that this, for systematic purposes, is

impracticable, at the same time I assert the right to use

a trinomial for any organism of known zygotic constitu-

tion, this being, at the present time at least, the easiest

way of designating any particular form. Some day we

shall have formulas, corresponding to those of chemistry,

to designate the lesser forms.

The increase in the number of named forms, a neces-

sary consequence, need cause no alarm, since they concern

only him who occupies himself with one species exclu-

sively. But we must go even further than this. Squarely

facing the issue, we find ourselves placed in a position

which necessitates the naming of heterozygotes. Obvi-

ously, numerous objections could be urged. But since it

has been shown, on the one hand, that certain forms can

exist only in a heterozygous form (Baur’s Antirrhinum)

and, on the other, that not only the difference between the

homozygote and the heterozygote is as great as that be-

tween many of our ‘‘systematic’’ species (for instance,

attenuata, bbNn, and arachnoidea, bbNN), but that a

homozygotic condition for a single gene gives the same

result, whatever the condition of the other known genes,

at least as thus far determined (arachnoidea occurs as

aaBBNN, aaBbNN and aabbNN), the advantage of nam-

ing all forms of different zygotic constitution must be

granted. :

Thus far we have not encountered a plant of the zygotic

constitution BbNN, at least as far as can be judged from

No. 568] X CAPSELLA BURSA-PASTORIS ARACHNOIDEA 221

breeding experiments. On being selfed such a plant

should yield:

BN | bN

BN | BN

BN | oN

bN | bN

It has been shown that plants of the zygotic constitu-

tion BBNN and bbNN exhibit the arachnoidea type. At

least 50 per cent. of the offspring then should show this

character. But if the suggestion made above is the cor-

rect one, i. e., that all plants homozygotic for N exhibit

the arachnoidea type, then the parent and its entire off-

spring should bear this character. The unfortunate in-

fertility of arachnoidea prevents us from submitting this

hypothesis to direct experimental proof. But there exist

indirect means for establishing the probable truth of our

contention. In the first place, we may cross two plants,

the identity of which can be established beyond doubt, to

wit, attenuata (bbNn) and Treleaseana (BBNn). Sucha

cross would yield:

BN Bn

Of these, we would recognize Bbnn because of its rhom-

boidea character, 50 per cent. would be recognized as

Setchelliana (BbNn), while the remainder, if our surmise

is correct, would consist of arachnoidea. Experiments to

determine this are under way. At the present we have

another, though by far less accurate, means of testing our

hypothesis. If the combination NN always results in a

form arachnoidea, the offspring of a plant of the zygotic

constitution BbNn would be composed of :

4 Setchelliana (BbNn),

2 Treleaseana (BBNn), -

2 attenuata (bbNn),

4 arachnoidea (1 BBNN, 2 BONN, 1 bbNN),

222 THE AMERICAN NATURALIST [Vou. XLVIII

3 rhomboidea (1 BBnn, 2 Bbunn),

1 simplex (bbnn).

Since BbNn, BBNn and bbNn, in the earlier experi-

ments, might have been confounded in the later stages,

and since there is little doubt as to the earlier stages,

these three forms have been combined in Table V.

TABLE V

RESULTS FROM SELECTED Setchelliana (BbNn)

“ Narrow ” | arachnoidea rhomboidea simplex

Index | EES EROS eee ro

26,912 | 134 | 157.6 | 94 | 78.8 | 68 | 59.1 19 | 19.7

3,613 94 89 45 45 33 33.75 8 11.25

This, especially in the case of No. 3,613, is a fairly close

approximation to what we might expect. When in No

3,613 we attempt to distinguish between Setchelliana,

Treleaseana and attenuata, we get the following num-

bers, the expected numbers following in parentheses:

BbNn 39(45), BBNn 21(22.50), bbNn 34(22.50), the last

number being far too high. When the experiments were

begun, we distinguished only between ‘‘narrow,’’

‘“broad’”’ and ‘‘linear.’’ To-day we know that the ‘‘nar-

rows” include Treleaseana, Setchelliana and attenuata,

that the ‘‘broads”’ include rhomboidea and simplex, while

the linears are identical with arachnoidea. In this light

it is of interest to go back to the first generation of 1910.

Our data yield the figures given in Table VI.

TABLE VI

“ Narrow ” “ Linear” | “ Broad ”

Index No. pe | |

| Found Expected | Found Expected | Found Expected

7,911 | 34 oa 9 15.25 | 18 15.25

8,111 ba SF 35.50 16 mwi B 17.75

8,311 | 66 61 Ye 30.50 | 24 30.5

8,711 | page 13.50 | 14 13.50

8,811 | 49 46.50 20 98.55: — 24 23.25

9,011 | 4 7.50 4 3.75 | 7 3.7.

9,511 t oe 6 +o 38 Let 38

voni 28 23.50 | 5 176 |. l 11.75

Total i... 38 308.50 | 133 154.25 _156 154.25

Per cent........ | 63.2 50 LS 25 | gee 25

No. 568] X CAPSELLA BURSA-PASTORIS ARACHNOIDEA 223

It must be granted that the approximation is fairly

close, and that, taken in consideration. with the others, it

offers ample support for the correctness of the diagnosis

of the zygotic constitution of the original plant. It at

least offers a working basis. One would be tempted to

accept it as a final solution were it not for the fortunate

appearance of a plant which does not fit into our scheme

and which, provisionally, has been named Capsella Bursa-

pastoris orbicularis.

CAPSELLA BURSA-PASTORIS ORBICULARIS

“This form differs from any other plant encountered

in my cultures. While in a general manner resembling

simplex, it differs in being more robust, having larger

flowers (though not as large as those of C. grandiflora),

and in having orbicular first

leaves (Fig. 13). All leaves

are covered with stout hairs. S) Lk)

It is a plant which tempts us Y-

to draw a parallel between it CTE A

and @Œnothera gigas, a name A

which I have not used for the

sake of avoiding an implied

comparison.

The first plant of this type appeared in a culture of

attenuata (26,012BR3P5) and was of sufficiently striking

appearance, though but four or five leaves had developed,

to call for a special note and a photograph. Later the

plant was potted and finally seed was gathered from the

unguarded plant. From this seed four seedlings were

obtained. At least three of them closely resembled the

parent, the fourth having somewhat narrower leaves.

Later the differences between these plants and those of

simplex became more apparent (Fig. 14). Those of my

students to whom the differences have been pointed out

have not the slightest difficulty in distinguishing between

the two forms. It is hoped that later, when by means of

prolonged cultures I shall have made myself more familiar

with this form, it may be made the subject of a distinct

paper where histological and cytological studies will find

Fig. 13. chorea oF C.

. . orbicul

224 THE AMERICAN NATURALIST [Vou. XLVIII

a place. One would be inclined to look upon orbicularis

as a mutation. But the fact that at first we classed

arachnoidea as such, later to prove it of hybrid origin,'*

Fic. 14. Four SEEDLINGS oF C. .. . orbicularis AND (THE LOWER) Two SEED-

LINGS OF O. . . . simplez.

would tend to make us cautious, and lead us to attempt to

find a solution for the origin of orbicularis in the disso-

ciation or combination of certain ‘‘units.’? While I should

not care to go quite as far as M. Heribert Nilson’? ‘‘das

ganze Mutations phänomen durfte unter einen gemein-

samen Gesichtspunkte: der Mendelschen Neukombination

eingeordnet werden können,” yet it is probable that here

the majority of alleged mutations may be classed.

17 Baur’s (Vererbungslehre, 189) narrow-leaved Melandrium album is

perhaps susceptible of the same explanation.

18 Zeitschr. f. ind. Abst. u. Bererb., 8: 89, 1912.

No. 568] X CAPSELLA BURSA-PASTORIS ARACHNOIDEA 225

An examination of the herbarium material placed at

my disposal reveals the fact that plants, apparently

identical with C. orbicularis, occur in Europe. In the

Engelmann herbarium of the Missouri Botanical Garden

are two sheets (No. 3,661 and 3,664) containing specimens

which undoubtedly must be classed here. The latter sheet

bears the label: Thlaspi Bursa-pastoris humile. Heidel-

berg. April 1828.

A culture of Capsella, derived from seed of a single

plant, unfortunately not preserved, escaped from culti-

vation in the Experiment Garden, and consisting of 182

individuals (Ehlers, No. 4,813), appears to be composed

entirely of orbicularis. And while I have never encoun-

tered the plant in nature, these two facts lead us to another

possible explanation. Perhaps the appearance of orbic-

ularis in the original culture was due to an accidental

admixture, such as is almost impossible to guard against

when experimental plants are grown in a greenhouse used

for a variety of purposes.

The exact relation which orbicularis bears to the other

types of Capsella here described can, of course, be de-

termined only after a series of experiments has been car-

ried out. However, the delay in the completion of the

manuscript, caused by the unfortunate destruction, by

fire, of the botanical laboratories of the University of

Michigan, enables me to add that a third generation of

orbicularis, the parent being No. 32,012R1P3, shows at

least two and possibly three types, of which one is espe-

cially interesting in having rather narrow leaves, at least

as compared with those of typical orbicularis. The con-

trast between the two forms is increased by the fact that

in the narrow-leaved form the foliage is entirely glabrous,

while in the typical orbicularis the leaves are covered

with numerous stiff, almost bristle-like, hairs.

X CAPSELLA Bursa-PASTORIS ARACHNOIDEA

By this name is designated the linear-leaved form, the

appearance of which induced us to undertake the cultiva-

tion of Capsella Bursa-pastoris Setchelliana.

226 THE AMERICAN NATURALIST (VoL. XLVIII

Fic. 15. tOSETS ILLUSTRATING TEE Two TYPES

Already the leaves which immediately follow the coty-

ledons serve to distinguish plants of this type from all

others. At the ten-leaf stage even the casual observer is

able to segregate them at once from the other rosets.

The leaves are acicular and the cotyledons far larger

than those of the seedlings of the other forms. The

greater size of the cotyledons may be attributed to the

insufficiency of the subsequent leaves.

If one removes the terminal bud of seedlings of Atri-

plex hortensis or one of its color varieties, it will be found

that the cotyledons increase in length far beyond normal,

sometimes reaching a length of 8 em. Under favorable

conditions the leaves of X C. arachnoidea may reach a

length of 100 mm., with a greatest width of 6 mm. (Fig.

16). The stem ordinarily is weak, having a diameter of

only 1 mm. It may reach a length of 30 em. (Fig. 17)-

No. 568] X CAPSELLA BURSA-PASTORIS ARACHNOIDEA 227

THUS FAR RECOGNIZED 1N C... . orbicularis.

The flowers are small, the petals especially so. The

anthers shrivel up early and as a rule are devoid of pollen

grains. Occasionally a few can be demonstrated. The

ovary, though small, contains what appear to be ovules

capable of being fertilized. Thus far I have collected

eight seeds contained in 6 capsules on unguarded plants

of arachnoidea (Fig. 18). Two of these germinated, the

one yielding a plant whieh looks like simplex, though

having a lar ge amount of red coloring matter in the peti-

oles, while the other is an arachnoidea. Attempts to arti-

ficially fertilize arachnoidea have failed absolutely.

As has been shown above, one may distinguish, on the

basis of genotypic constitution, three forms of arach-

noidea, viz.: BBNN, BbNN and bbNN. Externally no

228 THE AMERICAN NATURALIST [Vou. XLVIII

differences can be noted. A single exception perhaps may

be made to this statement. It had been noted that speci-

mens of arachnoidea frequently showed fasciation. This

fasciation seems most marked in plants of the zygotic

constitution BBNN (Figs. 19, 20, 21).

Fic. 16. Roser or æ CC... . arachnoidea.

While it is hoped that later a more extended report may

be made upon this plant, at present it may be stated that

there exists the probability that it may throw some light

upon the nature of fasciations. In earlier publications’®

I have brought together some of the known facts bearing

upon this teratological character. Though a large por-

19 “í Fasciation in Oxalis crenata and Experimental Production of Fascia-

tions,’’? Rep. Mo. Bot, Gard., 17: 147, 1906; ‘‘Fasciations of Known

Causation,’’ AMERICAN NATŪRALIST, 42: 81, 1908; ‘‘Inheritance of Fascia-

tion in Zea Mays,’’ The Plant World, 14: 1911; ‘‘The Origin of Species

in Nature,’’?’ AMERICAN NATURALIST, 45: 641, 1911; ‘‘ Frondescence s

Fasciation,’’ Plant World, . 2910. Passion in Ozalis crenata,’

Botanical Journal, 2: 111, 1913.

No. 568] X CAPSELLA BURSA-PASTORIS ARACHNOIDEA 229

tion of the experimental garden is devoted to cultures of

fasciated races, nothing further has been determined than

that the fasciated character is inherited, that it is trans-

mitted through non-fasciated individuals, that its ap-

parentness depends upon nutrition, that it behaves as a

fÍ

Fig. 17. Two HERBARIUM SPECIMENS OF @ ©... . arachnoidea.

recessive character and that the fasciated character of the

stem appears to be associated with split leaves and cup-

Shaped leaves. Ina paper read before the Research Club

of the University of Michigan on March 16, 1910, and an-

nounced under the title ‘‘The Identity and Inheritance of

Teratological Characters,” I showed that split leaves,

ascidia, certain disturbances in the arrangement of the

230 THE AMERICAN NATURALIST [Vou. XLVIII

flowers, supernumerary locules in the fruit, etc., may

safely be taken as an indication of the presence of the

fasciated character. More recently, Kajanus,” working

Fic. 18. SHOOT or gz C.... oe a WITH A beet NUMBER OF INFERTILE

AND FE ERTILE CAPSU

with different material, has fully confirmed the views

which I expressed at the time. This is of particular in-

20 Kajanus, B. Sieh seins r Fasziation bei Trifolium pratense L.,’’

Zeitsch. f. ind. Abst. u. Vererb., T: 63, 1912; ‘‘Ueber einige vegetative

Anomalien bei cafes Siac L., ibid., 9: 111, 1913.

No. 568] X CAPSELLA BURSA-PASTORIS ARACHNOIDEA 231

terest in connection with Capsella arachnoidea, since

many of the plants which do not show a fasciated stem

do show split leaves (Fig. 20) and a most peculiar whorl-

ing of the flowers (Fig. 21).

Fic. 19. FASCIATED PLANT OF ¢ C. .. . arachnoidea.

The spatulate condition of the leaves of the seedling

shown in Fig. 2 is believed to have been due to fasciation.

Capsella Bursa-pastoris arachnoidea, then, bears all the

earmarks of a fasciated race. All of the three zygotic

combinations which yield the arachnoidea type are homo-

zygotic for N. The recent work of East and Hayes, and

of Emerson on Zea Mays has shown that the fasciated

232

THE AMERICAN NATURALIST

[ Vou. XLVIII

+ ')

Cal

YJ i

\ |

Fig. 21. ABNORMAL WHORLED ARRANGEMENT OF

CENCES OF @ 0. ...

Frio: 20. LEAVES oF #0... arachnoidea.

TE y VA

j4 an

r =% *

BS ri

{ f TF A `

tf? }

Dri

SAF

Xo aA

, J

X “i

THE FLOWERS IN INFLORES-

arachnoidea.

No. 568] X CAPSELLA BURSA-PASTORIS ARACHNOIDEA 233

character is dominant, though Mendel, in his experiments

with Pisum umbellatum, has shown it to be recessive.

No fasciation, thus far at least, has been noted in the

other forms used in these experiments.

SUMMARY $

A culture of Capsella Bursa-pastoris proved heterozy-

gotic, yielding certain new forms (X C. Bursa-pastoris

Setchelliana, XC.Bursa-pastoris Treleaseana, XC.Bursa-

pastoris arachnoidea and X C. Bursa-pastoris attenuata),

as well as certain forms already described by Shull (C.

Bursa-pastoris rhomboidea and C. Bursa-pastoris sim-

plex) in the proportion 4:2:4:2:3:1. The distinction

between simplex and rhomboidea, both inter se and be-

tween them and the other forms, is readily made by any

one familiar with Shull’s investigations. These two

plants agree in having the earlier leaves broad (Fig. 4).

The climax leaves of rhomboidea and simplex show

marked differences, especially. as far as the incision of

the blade is concerned. These incisions, in simplex, reach

a depth equal to approximately one fourth of the width

of the blade (Fig. 10). In rhomboidea the incisions are

deeper, reaching the midrib in the homozygous form

(Fig. 12). The leaves of the latter also show marked

secondary lobes.

The distinction between X C. Bursa-pastoris Setchelli-

ana, X C. Bursa-pastoris Treleaseana and X C. Bursa-

pastoris attenuata is made with greater difficulty. They

agree in having long and narrow first leaves. The climax

leaves of Treleaseana and Setchelliana show marked

incisions, exceeding one fourth of the width of the blade,

and which may reach the midrib (Fig. 9). The latter

form also may show marked secondary lobes.

Besides the phenotypes here mentioned occur two

others, the one, X C. Bursa-pastoris orbicularis, with an

almost orbicular first leaf (Fig. 13) and a climax leaf

greatly resembling that of simplex (Figs. 14, 15), though

differing i in texture. This form has not been sufficiently

studied, but is believed to be identical with one known to

234 THE AMERICAN NATURALIST [VoL XLVIII

occur in Europe. Finally there is X C. Bursa-pastoris

arachnoidea, a sterile, linear-leaved form, with a weak

stem and which frequently shows fasciation (Figs. 17-21).

To facilitate a distinction between these forms, a key is

appended:

a. Early leaves broads

b. Early leaves orbicular. orbicularis.

bb. Early leaves twice as long as broad.

c. Climax leaves incised to midrib. rhomboidea.

cc. Early leaves not incised to midrib. simplex.

aa. Early leaves long and narrow.

b. Early leaves acicular. arachnoidea.

bb. Early leaves 24-3 times as long as broad.

c. Climax leaves not incised to midrib. attenuata,

ce. Climax leaves incised to or almost to the midrib.

d. Secondary lobes pronounced. Treleaseana.

dd. Secondary lobes absent. Setchelliana.

It was found that, besides the genes A, B, C and D,

whose existence was shown by Shull, there exists another

gene, N, responsible for the narrow character of the

earlier leaves. For the various forms, mentioned here,

the following zygotic constitutions have been tentatively

determined: simplex, bbnn; rhomboidea; BBnn and Bbnn;

Setchelliana, BbNn; Treleaseana, BBNn; attenuata,

bbNn; arachnoidea, BBNN, BbNN and bbNN. The zygo-

tic constitution of orbicularis has not been determined.

As to the probable origin of X C. Bursa-pastoris Setch-

elliana, little can be said. It most probably results from a

cross between rhomboidea and attenuata (BBnn X bbNn).

This seems the most plausible explanation since, judging

from herbarium specimens, both attenuata and rhom-

boidea occur throughout the United States. Unfortu-

nately such an assumption necessitates an explanation of

the origin of attenuata.

My thanks are due to the regents of the University of

Michigan for the facilities placed at my disposal, to head-

gardener Adolph Weiner for his constant care of the ex-

perimental plants, to Messrs. J. H. Ehlers, A. Povah, C.

Oberlin and A. W. Murdock for assistance in classifica-

tion of the seedlings and to the director of the Missouri

Botanical Garden for the loan of herbarium material.

No. 568] X CAPSELLA BURSA-PASTORIS ARACHNOIDEA 285

CONCLUSIONS

1. Besides the genes (4, B, C, D) discovered by Shull,

there exists in Capsella a gene N, responsible for the nar-

row character of the early leaves of certain forms.

2. Absence of the gene N results in the formation of

early leaves of a ‘‘broad”’ character.

3. The form designated arachnoidea is of hybrid origin,

as are the forms Setchelliana, Treleaseana and attenuata.

4. X Capsella Bursa-pastoris arachnoidea is formed

whenever the plant is homozygotic for N, whatever the

constitution of the remainder of the zygote (BBNN,

BbNN, bbNN), i. e., a homozygous condition for the pres-

ence of a single factor may overshadow the influence of

others.

5. Homozygocity for a single factor may be responsible

for total, or almost total, sterility.

6. A knowledge of the early stages, as well as of the

climax leaves, is essential for the classification of the

phenotypes of Capsella Bursa-pastoris.

BIOLOGY OF THE THYSANOPTERA. II

DR. A. FRANKLIN SHULL

UNIVERSITY OF MICHIGAN

II. SEX AND THE LIFE CYCLE

INTRODUCTION

From observations made on the abundance of males in

several species, Jordan (1888) was led to believe that

there might be among Thysanoptera, as in aphids, an

alternating life cycle; that is, that there might be a series

of parthenogenetic generations during the summer, fol-

lowed by a generation of males and sexual females in the

latter part of the summer or in the fall. Coupled with

this he suspected that there were winged forms in the

parthenogenetic part of the cycle, and at least occasional

wingless individuals in the sexual phase.

Uzel (1895), however, was unable to detect any indica-

tions of such a cycle. He held that there could be no

question of parthenogenesis in a species in which males

were abundant all the time or at intervals. Only in spe-

cies in which the males were too rare to impregnate all

the females would he admit parthenogenesis. To prove,

in such a species, an alternating cycle like that of the

aphids, it must, in Uzel’s opinion, be shown that the males

are abundant only at certain seasons. As Uzel was ac-

quainted with no European species in which males were

plentiful at but one season, he rejected Jordan’s sugges-

tion regarding an alternating cycle, and his view seems

to have been accepted by thysanopterists since that time.

To Uzel’s argument it may be objected that the pres-

ence of males, and even the occurrence of copulation, is

no proof that parthenogenesis is wanting. For among

the aphids and rotifers, the parthenogenetic and sexual

females exist side by side. Nor is parthenogenesis in

these two groups facultative (optional), as Uzel appears

to assume for Thysanoptera; a female is either only

236

No. 568] BIOLOGY OF THE THYSANOPTERA 237

sexual or only parthenogenetic. Moreover, in the roti-

fers, females incapable of fertilization copulate as fre-

quently as do those requiring fertilization, as was first

shown by the work of Maupas (1890) on the rotifer

Hydatina.

Presence of males and occurrence of copulation are,

therefore, no proof of sexual reproduction. But even if

we accept, as Uzel does, this criterion of sexuality, Jor-

dan’s view that there may be an alternating cycle would

receive some support if it could be shown that males are

more abundant at one season of the year than at other

times. Casual observations made by me several years ago

seemed to indicate this seasonal variation in the abund-

ance of males. As the data then available were meager,

no conclusion was drawn, but I subsequently undertook

to obtain such data on a larger scale, by making extensive

collections at all seasons of the year to determine the sex

ratio. The following pages give these data, along with

other observations bearing on sex or the life cycle.

I desire to acknowledge the assistance of my wife, by

whom much of the labor of determining species and

counting the sexes was done.

THE Sex RATIO IN Various Species or THYSANOPTERA

In making collections for the purpose of determining

the sex ratio, the food plants were examined very care-

fully, torn apart if necessary, and every individual cap-

tured. This precluded the possibility of obtaining an

erroneous sex ratio because one sex was more easily dis-

turbed than the other. A few individuals escaped, but

they could not have affected the sex ratio very greatly,

and it was known from their size that they were some-

times of the one sex, sometimes of the other.

The sex in the suborder Terebrantia is readily deter-

mined by the presence of an ovipositor in the female and

the rounded end of the abdomen in the male. In the sub-

order Tubulifera, the sex in Anthothrips verbasci was

determined by the presence of two short, heavy spines,

238 THE AMERICAN NATURALIST [Vou. XLVIII

one on each side of the abdomen of the male, near the end.

As the specimens, when placed on a microscope slide,

nearly always lie either on the dorsal or ventral side,

these spines are nearly always readily visible if present.

I used this criterion (mentioned in the re-description of

the species by Hinds, 1902) only after having taken eleven

pairs of this species copulating in nature, and observing

in every case that the male possessed these spines, and

that in the female they were wanting. In other Tubu-

lifera, e. g., Anthothrips niger, sex was determined by the

longitudinal chitinous rod in the next to the last abdomi-

nal segment of the female. When the specimens were too

opaque to observe this rod, they were cleared by boiling

in caustic potash.

The data from these collections are given in the accom-

panying table. Unfortunately the collections could not

all be made in one year, nor in the same locality. Those

made from July 1 to September 18, 1912, were made at

the University of Michigan Biological Station, Douglas

Lake, Michigan; all others were made at Ann Arbor,

Michigan. It is not probable that the results are greatly

modified by collecting in two regions within the state.

In this table the larve of all species are combined, as I

am unable to distinguish with certainty the larve of

several of the species here mentioned.

The important facts contained in this table are, it seems

to me, the following:

Euthrips tritici appeared in spring at first only in the

female sex. Males were first collected nearly a month

later, and not until about the time fairly large larvæ were

found elsewhere. Once the males appear, though their

number fluctuates in the individual collections, they fur-

nish a fairly constant proportion of the whole number

(about one third).

The males of Anthothrips verbasci appear in the earli-

est collection of this species, and in considerable numbers

throughout the season. The total proportion of males is

23 per cent., and the only considerable increases over

No. 568] BIOLOGY OF THE THYSANOPTERA 239

this percentage in individual collections are in the three

collections made in August, and on October 7. Consider-

ing the large majority of females taken September 12, the

abundance of males October 7 may be due in some way to

TABLE I

SHOWING NUMBER OF MALES AND FEMALES OF THE COMMONER SPECIES OF

THYSANOPTERA CAPTURED AT INTERVALS DURING THE ACTIVE SEASON

Chir

Euthrips | aneno an me- | Thrips anapi | Thrips ae trips

tritici verbaski | niger | tabaci | striatus Physopus, mani- | Larve,

Date | catus «all H

Aoa —j— Species

Q So ae eee | g Q | P Poo te a

Apr. 30, 1911 9| 0 | |

May 5 T O | |

10 2- 0 |

17 i 0 |

22 22) 1 2:0 213

24 30| 2/174! 30) 20; 0 | 6 0 7

June 1 41| 17 18 0 6

7 40| 62| 52| 10) 42) O 4| 7 3

15 21| 27| 47| 8 19 0 1 2 25

21 7| 11 410 | 25} 0] 1! 0} O| 1 1

Ti) ô 40 Sa) Piet OS

July. 3,1912 | 14| 25 1 0 46

Se 7 zo 20 72

5 3| 0| 82| 29) 2| 0 23 0 1

11 0 3 0 | 95| 18 187) 1! T6

16 7-9 4i 9 T OLI® O 3

17 6| 1 $1] 0 | 42) 0

19 12) 29 2| 0 | 35| 6 5 Ol 3

26 58| 15 2

27 a 0 45! 13 4

29 2 0 4 0 10| 9 12

30 1| 51

42) 15 1

Aug, 5 154/151 28) 10| 1 39

8 1| 30| 20 0,10; 1

9 16 1 0

12 2103) 2

13 26' 48 1

18 60 16 20 i 0

20 18' 17 39| 48 2| 36) 21

40| 3 10 0 13

Sept 60) 7 27| 2 | 60| 33 10

2, 1911 | 35| 15 801190 22 2

88) 1 18

ig 3 15 0 1 o 4

18, 1912 183| 21

Oct. 7,1911 |32 13| 23! 27 19 0 12

4 18} 4 11 0 2; 0

25,1912 | 56] 19 12| 0 15| 4 2

Nov. 9 27 3 12| 0 1 0

E a ee o

FoM o 879 441 641 200/162’ 0 226! 0 1530:174 50 | 36 (203/191

240 THE AMERICAN NATURALIST [Vou. XLVIII

the dying of their food plants; but the greater proportion

of males throughout August is probably significant. It

should also be stated that I have collected adults of this

species, of both sexes, from dead mullein spikes in late

winter.

Anthothrips niger was found only in the female sex.

There are no records of males of this species, so far as I

am aware, in any published work.

Thrips tabaci was taken almost exclusively in the

female sex, the two males found September 2 being the

only ones I have ever collected.

In Anaphothrips striatus the total number of males is

less than 25 per cent. On August 20 and September 2 the

proportion of males is considerably greater than 25 per

cent., especially on the former date, while at other times

the proportion was nearly always less. The collection on

August 20 can hardly have been erroneous by chance, for

the figures given for that date are combined figures for

two collections from different localities. In one of these

collections there were 13 females and 14 males, in the

other 26 females and 34 males. This strengthens the

probability that the excess of males is significant.

Thrips physopus was collected in small numbers, but

shows a fairly constant proportion of males.

Chirothrips manicatus presents curious phenomena.

All the collections up to the end of July were made on

timothy heads in a small patch a few feet square near the

laboratory. On July 11 careful search revealed numer-

ous females, but only one male. By July 19 almost all

the thrips of this species were gone; only 5 specimens

were obtained, and these were females. Less than two

weeks later, however (July 30), on other timothy heads

in the same small patch, there were found 51 males and

but 1 female. No living thrips were taken here later, as

the timothy died; but subsequent collections elsewhere,

from timothy and bluegrass, show again almost ex-

clusively males.

No. 568] BIOLOGY OF THE THYSANOPTERA 241

AppitionaL Data BEARING ON THE Lire CYCLE AND Sex

In view of the fact, to be discussed later, that Anapho-

thrips striatus has hitherto been known almost exclu-

sively in the female sex, and is known to reproduce par-

- thenogenetically, and the fact that in the collections here

recorded the males constitute nearly 25 per cent. of the

total, the question arises, are these males functional? If

not functional in this species, are the males functional in

other species? A number of observations and experi-

ments I have made bear on these questions.

A single pair of Anaphothrips striatus was found copu-

lating in nature, which Uzel would have considered proof

that parthenogenesis did not occur. The testes of the

males are plainly visible without dissection. Suspecting

that they might not be fleshy organs at all, but chitinized

structures, perhaps vestiges of testes, I boiled a number

of specimens in caustic potash. The testes disappeared,

from which I judge they are not merely chitinous bodies.

I can say nothing of their cellular nature, owing to the

loss of material killed and fixed for that purpose. Nu-

merous sections of another species Anthothrips verbasci,

however, reveal well-developed testes. Cell divisions

(probably the spermatocyte divisions) and nearly mature

spermatozoa in bundles were observed in these sections.

Though the number of chromosomes could not be deter-

mined, it is an interesting fact that spindles in side view

usually showed a lagging chromosome.

Finally, with further regard to the functioning of

males, I have attempted to breed several species par-

thenogenetically. The results in the case of Euthrips

tritici were so far encouraging that two larve appeared

on the plant on which virgin females had been previously

placed. But in these cases I could not be certain that the

food plant was uninfected. Experiments with Anapho-

thrips striatus and Anthothrips verbasci gave negative

results, but in each case failure to obtain young by par-

thenogenesis may have been due to the conditions.

Some observations on the place of pupation may also

242 THE AMERICAN NATURALIST [Vou. XLVIII

be here recorded. The rarity with which the pupe of

most species are discovered in collecting suggested that

they might not pupate on the food plant of the larve.

Some species of thrips, for example, the pear thrips

(Euthrips pyri), are known to pupate in the ground

(Moulton, 1912). Since many of the species included in

Table I may be found on white clover, which was abun-

dant at Douglas Lake, the place of pupation of these

species was tested in the following manner. A mass of

the flowers of white clover was collected. The flowers

were gently squeezed for some time to drive out all the

adults. They were then placed in a vessel under cover.

After two days, when the flowers were thoroughly dried,

they were again gently crushed to make sure that all

adults were driven out. At intervals from one to two

weeks afterward, 15 adult thrips appeared on the inside

of the glass cover. These were of three species, Euthrips

tritici, Thrips tabaci and Anthothrips niger.

I have also frequently observed the pupe of Antho-

thrips verbasci in mullein spikes, those of Sericothrips

cingulatus on white clover, the pupa of Trichothrips tri-

dentatus under the bark of the white oak, where the larve

and adults live, and that of an undescribed species on

willow galls along with larve of the same species. I

judge from these observations that the majority of thrips

pupate on the plants on which the larve live, and that

their rarity in collections is due merely to concealment

and sluggish habits.

Discussion OF THE Resuuts IN RELATION To THE LIFE

CYCLE

From the data in Table I and the observations given

above it is evident that there is considerable diversity in

different species with regard to the life cycle, and diver-

sity within the same species at different times or in

different regions. First, as regards the mode of passing

the winter, it would seem that in Euthrips tritici only the

females survive that season. The reason for so believing

No. 568] BIOLOGY OF THE THYSANOPTERA 243

is that males could not be found in the spring until the

females had been active long enough to have produced

one generation of offspring. Males occur late in autumn,

but must perish before the end of winter. Likewise,

neither eggs nor larve live over winter, or larve would

appear earlier in spring. In Thrips physopus, on the

other hand, males were found as early as the females;

hence, in the absence of any collection earlier than May

22, and in ignorance of the time required for develop-

ment, I should assume that both sexes survive the winter.

Both sexes of Anthothrips verbasci have been seen on

dead mulleins in winter.

In species, like Euthrips tritici, whose males do not

survive the winter, if fertilization of the early spring

females takes place at all, it must occur in the fall. I do

not regard my breeding experiments as proof of par-

thenogenesis in this species, but it is by no means improb-

able that parthenogenesis occurs. More rigorous experi-

ments are needed.

As regards the mode of reproduction during the rest of

the year, there is nothing in the sex ratio, as given in

Table I, to suggest an alternating cycle in Euthrips tri-

tici. In other species, it would be possible to interpret

certain facts to mean that an alternation of partheno-

genesis and sexual reproduction occurs, or did once

occur. There is a well-marked increase in the proportion

of males in Anaphothrips striatus, for example, in Au-

gust. This is a particularly interesting species. Hinds

(1902) saw only the female of this species, though he

mounted and examined over a thousand specimens, and

he bred it parthenogenetically in the laboratory for

months. What purported to be the male was described

by Cary (1902), from Maine, but the specimens described

were evidently those of another species. The first males

ever recorded were described by Shull (1909), two speci-

mens among probably two hundred females. It is re-

markable, therefore, that in the vicinity of Douglas Lake

there should be nearly 25 per cent. of males. Whether

244 THE AMERICAN NATURALIST [Vou. XLVIII

the presence of numerous males is dependent on climatic

conditions, or whether it is a racial difference, there is at

present no way of deciding. The weather was unusually

cold during the summer in which these records were

made, and it is desirable that the effect of temperature be

experimentally determined. The presence of males in

goodly numbers throughout the summer, the occurrence

of copulation in nature, and the failure of an attempt to

breed the species parthenogenetically, leave, as the only

reason for suspecting that it may have been partheno-

genetic at Douglas Lake, the fact that it is parthenoge-

netic elsewhere. But if the species is parthenogenetic in

one region and sexual in another, it is not difficult to be-

lieve that it may be both parthenogenetic and sexual in

the same region. It is difficult to decide whether the well-

marked increase in the proportion of males in August

and early September should be regarded as evidence of

such an alternation, or as due to a period of cold weather

or other climatic factor, or as a hereditary remnant of

the sexual phase of an alternating cycle once possessed

by the species. Only experiment, and perhaps cytological

study, can decide this question.

A similar but less marked increase in the number of

males is seen in Anthothrips verbasci, also in August. In

that month the proportion of males rose from about 20

per cent. to 40, or even nearly 50 per cent. In this species

the increase may be due to the late date at which the first

brood of larve becomes mature. The life history of this

species is longer than that of most of the suborder Tere-

brantia, and may appear to be still longer because ene-

mies destroy many of the larger larve. For these rea-

sons, in the region of Douglas Lake, the first generation

of larve may not become mature until nearly August. If

this assumption is correct, the proportion of males found

prior to August is the proportion that survive the winter.

This explanation receives support from the cytology of

the germ cells. As stated above, there is a lagging chro-

mosome in the spermatocyte divisions, which suggests

No. 568] BIOLOGY OF THE THYSANOPTERA 245

the probability that there are two classes of sperm asso-

ciated with sex, as in the bugs and many other animals,

and that Sherefoe the sexes should be approximately

equal in numbers. The 40 to 50 per cent. of males in

August accord fairly well with this explanation.

This explanation would not, however, account for the

increase in the number of males in late summer in a spe-

cies whose life history is much shorter than that of

Anthothrips verbasci. Thus, in Anaphothrips striatus,

Hinds states that the entire life history is passed through

in 12 to 30 days. Even in a cold season, such as that of

1912 at Douglas Lake, therefore, the life history can not

have been so long that the first adults would emerge in

the middle of August. The increase in the number of

males of Anaphothrips in August and September is not

to be explained, therefore, as due to the first appearance

of a new brood at that time.

Thrips tabaci likewise affords interesting, even if mea-

ger, evidence regarding the seasonal occurrence of males.

In this species males are exceedingly rare. Hinds (1902)

redescribed the male in quotation marks, from which it

is to be inferred that he did not have specimens. In my

own collecting, though the females were quite common, I

never saw a male until the summer of 1912. Then two

specimens were taken September 2, as shown in Table I.

These irregularly occurring males can hardly be fune-

tional, so that Thrips tabaci is still probably to be re-

garded as wholly parthenogenetic. But their appearance

-in late summer may be the vestige of a former sexual

phase, and may be caused now, as the sexual phase prob-

ably was in part formerly caused, by climatic conditions.

Chirothrips manicatus presented, at Douglas Lake, an

anomalous condition. As shown in Table I, and stated

more explicitly above, females were abundant in a given

small area early in July, but practically no males were

present. Then, so far as I could determine by painstak-

ing collections, the females disappeared; almost no adults

of either sex, and not many larve, were to be found. Two

246 THE AMERICAN NATURALIST [Vou. XLVIII

weeks later, however, males were found in the same area

in large numbers. As these males were wingless, they

had probably not immigrated. The only other explana-

tion that occurs to me is that the larve were present in

considerable numbers at the time of the earlier collec-

tions, but in the flowers, not among the spikelets of the

timothy, so that I did not discover them; and that the

female larve reached maturity much earlier than the

males. In any case, it is difficult to see how the males can

have been functional, when the two sexes occurred at dif-

ferent times. If such conditions recur frequently, Chiro-

thrips manicatus, even though it produces many males,

must be parthenogenetic.

SUMMARY

The principal conclusions reached in the second part of

this work may be stated as follows:

Some species of Thysanoptera pass through the winter

in both sexes, in others the males perish. In none of

those studied does the egg or larva live over winter.

Pupation of most of the species of Thysanoptera stud-

ied occurs on the food plants where the larve live, not-

withstanding that the pupæ seldom appear in collections.

From the determination of the sex ratio, Huthrips

tritict shows no indication of an alternating life cycle. It

is probably sexual throughout the active season, though

this is not proven.

Chirothrips manicatus occurred abundantly in both

sexes, but the two sexes appeared at different seasons.

The explanation of this phenomenon is doubtful.

An increase in the number of males in Anthothrips

verbasci in late summer may be explained as due to the

great length of the life history and to selective mortality

during the winter, without assuming an alternating life

eycle.

Anaphothrips striatus, a species which has hitherto

been known almost wholly in the female sex, produced

about 25 per cent. of males at Douglas Lake. This may

No. 568] BIOLOGY OF THE THYSANOPTERA 247

be due either to climatic conditions or to racial differ-

ences. Sexual reproduction was not wholly proven, but

seems probable. An increase in the number of males in

late summer in this species and in Thrips tabaci might be

interpreted as indicating a sexual phase, or the vestiges

of a sexual phase that existed in the species formerly.

Jordan’s belief in an alternating life cycle, which was

rejected by Uzel, thus receives some measure of jus-

tification.

BIBLIOGRAPHY

Cary, L. R. 1902. The grass thrips (Anaphothrips striata Osborn). Maine

Agr. Exp. Station, Bull. 83, June, pp. 51-82

912

Jones, P. R. 1912. Some new California and Georgia Thysanoptera. U. S.

Dept. Agr., Bur. Ent., Tech. Ser. No. 23, Part 1, 24 pp., 7 pls.

Jordan, K. 1888. PORS und Biologie der Physapods. Zeit. wiss.

Zool., Vol. 47, pp. 541-620.

Hinds, W. E. 1902. Contribution to a monograph of e insects of the

order ppp top ie & North America. Proc. U. 8. Nat.

Museum, Vol. 26, No. , December 20, pp. 79-

Maupas, E. 1890. Sur i Te a de 1”Hydatina ata Ehr. Comp.

Rend. Acad. Sci. Paris, Tome 111, pp.

Moulton, Dudley. 1911. Synopsis, catalog and ‘bihttography of North

Aniiesn Thysanoptera. U. S. Dept. Agr., Bur. Ent., Tech. Ser. No.

21, 56 pp.

1912. Papers on deciduous fruit insects and insecticides. IV. The pear

thrips and its control. U. S. Dept. Agr., Bur. Ent., Bull. 80, Part IV,

pp. 51-66.

Shelford, V. E. 1911. Physiological animal geography. Journ. Morph.,

Vol. 22, No. 3, vig Stns pp. 551-618.

Shull, A. F. 1909. me apparently new Thysanoptera from Michigan.

Entom. News, a "20, No. 5, pp. 220-228.

1911. A biological survey of the sand dune region = the south shore of

Saginaw Bay, Michigan. Thysanoptera and Orthoptera. Mich. Geol.

and Biol. Survey, Pub. 4, Biol. Ser. 2,

Uzel, H. 1895. Miocrithis der Chiat Piopi. Königgratz,

privately published, 482 pp., 1

SHORTER ARTICLES AND DISCUSSION

BARRIERS TO DISTRIBUTION AS REGARDS

BIRDS AND MAMMALS

THE geographical range of any species of animal may be

likened to a reservoir of water in a mountain canyon. The con-

fining walls are of varying nature. A concrete dam, absolutely

impervious, may retain the water at one end. Along either side

the basin’s walls differ in consistency from place to place. The

substratum varies in porosity, at some points being impervious

like the dam, at others permitting of seepage of water to a greater

or less distance from the main volume. The water continually

presses against its basin walls, as if seeking to enlarge its area.

And it may succeed in escaping, by slow seepage through such

portions of its barrier as are pervious or soluble, or by free flow

through a gap in the walls, if such offers. The area occupied

by the water will extend itself most rapidly along the lines of

least resistance.

Every species has a center or centers of abundance in which

favoring conditions usually give rise to a rate of reproduction

more than sufficient to keep the critical area stocked. A tendency

to occupy a larger space results, because of competition within the

species: individuals and descent-lines multiply and travel radi-

ally, extending those portions of the frontier where least resist-

ance is offered. Such radial dispersal takes place slowly in some

directions, more rapidly in others, according to the degree of

passability of the opposing barriers. These barriers consist of

any sort of conditions less favorable to the existence of the

species than those in the center of abundance.

Theoretically, sooner or later and in all directions, every

species is absolutely stopped. But as a matter of undoubted

fact most barriers are continually shifting, and the adaptability

of the animals themselves may be also undergoing continual

modification ; so that perfect adjustment is beyond the limits of

possibility so long as topography and climate keep changing.

The ranges of species may thus be constantly shifting. Descent-

lines may move about repeatedly over the same general region,

like sparks in the soot on the back of a brick fireplace.

Yet, in all of our studies, of but a few years’ duration, the

248

No. 568] SHORTER ARTICLES AND DISCUSSION 249

time element is reduced almost to a negligible quantity, and we

may look upon the areas occupied by each species as, for the

time of our observation, fixed. We are thus enabled to compare

one with another, and because of the large number of the species,

we can infer a good deal as to the nature of barriers in general,

at least as regards birds and mammals. It is even conceivable

that; with sufficient refinement in methods, the inquirer might in

time find himself able, from a comparative study of the ranges

of rodents, for example, to establish the identity of all of the

external factors which have to do with the persistence of each of

the species; in other words to analyze the ‘‘environmental com-

plex’’ into its uttermost elements—as regards the existing species

of rodents in their recent development.

The most obvious kind of barrier to distribution is that con-

sisting of any sort of physical, or mechanical, obstruction. Such

obstruction affects directly the individuals of a species en-

countering it, either by stopping their advance or by destroying

outright such as attempt to cross it. As barriers of this nature, .

are to be cited land in the case of purely aquatic mammals, and

bodies of water to purely terrestrial, especially xerophilous,

mammals. In each case the width of the barrier has to do with

the degree of impassability. Oceans and continents are most

perfect, and affect a large proportion of the species. The com-

paratively narrow Colorado River is a barrier of the first rank,

but only to a certain few desert rodents. Mechanical barriers,

where they exist at all, are clearly recognizable.

It is to be observed, however, upon considering the birds and

mammals of a whole continent, that by far the greater number

of species are delimited in range without any reference to actual

land and water boundaries; more explicitly, their ranges fall far

short of coast lines. The barriers here concerned are intangible,

but nevertheless powerful. By their action the spread of species,

genera and families is held in check as surely as by any tangible

obstruction.

By these invisible barriers the individual may not necessarily

be stopped at all, as with animals of free locomotion; but the

species is affected. For example, the mocking bird in its Cali-

fornian distribution is closely confined to those parts of the state

Possessing certain definite climatic features; but vagrant indi-

viduals, especially in autumn, occur far beyond the limits of -

these restrictive conditions. Carnivorous mammals are well

250 THE AMERICAN NATURALIST [Vou. XLVIII

known to be subject to sporadic wanderings on the part of indi-

viduals, but the species is kept in set bounds by some potent but

invisible set of factors. The very fact that individuals are quite

capable of temporarily transgressing these bounds and yet do

not overstep them en masse emphasizes all the more the remark-

able potency of this category of barriers as regards species and

higher groups.

Our geographic studies lead us to designate among these rela-

tively intangible barriers: (1) increase or decrease in prevailing

temperature beyond certain critical limits, according to the species

concerned; (2) increase or decrease in prevailing atmospheric

humidity beyond certain limits; (3) modification in food-supply

and appropriate breeding and foraging ground. The limits set

by each of these factors will vary with the physiological pecul-

iarities of the organism considered; in other words the inherent

structural equipment of each animal figures importantly. In

these three sorts of barriers will be recognized what have been

called ‘‘zonal,’’ “faunal” and ‘‘associational’’ delimitation, each

of which I will now try to define.

Two schools of faunistice students are represented among Amer-

ican zoo-geographie writers of the present day. One, of which

C. H. Merriam is the most prominent exponent, sees in tempera-

ture the chief cause controlling distribution, and deals with the

ranges of species in terms of ‘‘life zones.’’ The other school, of

which C. C. Adams, A. G. Ruthven and Spencer Trotter are

active advocates, assigns to temperature but a minor rôle, look-

ing rather to a composite control, of many factors, resulting in

ecologic ‘‘associations,’’ of which plants are essential elements,

and which are to be further explained on historical grounds.

The two sets of areas thus defined do not by any means corre-

spond. Yet the reviewer can not fail to note, here and there,

places where boundaries coincide, and such coincidences are so

frequent as to be suggestive of real concordance in some signifi-

cant manner. Is it not probable that both schools are approxi-

mately correct, the difference in mode of treatment being due to

different weights given the different ae of evidence, or, in

other words, to difference in perspective

Every animal is believed to be Sisto in distribution zonally

by greater or less degree of temperature, more particularly by

that of the reproductive season. When a number of animals

(always in company with many plants similarly restricted)

No. 568] SHORTER ARTICLES AND DISCUSSION 251

approximately agree in such limitation they are said to occupy

the same life zone.

The observation of this category of distributional delimita-

tion is particularly easy in an area of great altitudinal diversity

like that comprised in the southwestern United States. The

writer is led to wonder if those authors who minimize the impor-

tance of temperature have ever been privileged to travel exten-

sively, and carry on field studies, outside of the relatively uni-

form eastern half of North America!

Study of any area which varies widely in altitude and hence

provides readily appreciable differences in daily temperature

from place to place brings conviction of the very great effective-

ness of temperature in delimiting the ranges of nearly all species

of animals as well as of plants. Particular attention may

called to the pertinent results of Merriam’s survey of Mount

Shasta. ;

But temperature is not to be considered the only delimiting

factor of environment, though its possible overemphasis by the

Merriam school seems to have led some other persons to believe

that this view is held. In fact it becomes evident, after a con-

sideration of appropriate data, that very many species are kept

within geographic bounds in certain directions only by an in-

creasing or decreasing degree of atmospheric humidity. By

plotting the ranges of many animals as well as of plants coin-

cidence in this regard is found in so many cases as to warrant

the recognition of a number of ‘‘faunal areas’’—on the causa-

tive basis of relative uniformity in humidity. It is probable that

every species is affected by both orders of geographic control.

The reader may enquire as to the grounds for employing the

widely used terms zone and fauna in the restricted sense here

prescribed. In reply, it may be said that this is not an inno-

vation, but is an adoption of a usage which has come about his-

torically among a certain group of workers in the geography of

vertebrate animals in North America. The writer recognizes the

fault in imposing restricted meanings upon old terms, but he

also hesitates at coining new wo

As to which is the more apòrta; assembled data seem to

show that more genera and higher groups are delimited by zonal

boundaries than by faunal boundaries. The arresting power of

temperature barriers would therefore seem to be relatively the

greater .

252 THE AMERICAN NATURALIST [Vou. XLVIII

In the third category of distributional control there is a con-

spicuous association of the majority of so-called adaptive struc-

tures of animals (often of high taxonomic value) with certain

mechanical, or physical, features of their environment. n

animal may thus intimately depend upon certain inorganic or

organic peculiarities, or both, of a given area, and be unable to

maintain existence beyond the limits of occurrence of those

features of the environment. Tracts of relatively uniform en-

vironmental conditions, including their inanimate as well as

living elements, are here called associations.

After a consideration of all the birds and mammals occurring

both within the state of California and elsewhere as far as the

writer’s knowledge goes, associational restriction appears to be

governed by the following three factors, of relative importance

in the order named.

1. Kind of food-supply afforded, with regard to the inherent

structural powers of each of the animals concerned to make it

available.

2: Presence of safe breeding places, adapted to the varying

needs of the animals, in other words depending upon the respect-

ive inherent powers of construction, defence and concealment

in each species concerned.

3. Presence of places of temporary refuge for individuals,

during daytime or nighttime, or, while foraging, when hard-

pressed by predatory enemies, again correlated with the respec-

tive inherent powers of defence and concealment of each species

involved

It is believed that the geographical distribution of any animal

is correctly diagnosed in terms of each of the three main group-

ings here suggested. In other words an animal belongs simul-

taneously to one or more zones, to one or more faunas, and to

one or more associations. No one of these groupings can be

stated in terms of the other, any more than a person can com-

pute liquids by ecandle-power, or weight in miles. The constit-

uent species within each of these groupings always belong to

the other two. To illustrate: the southern white-headed wood-

pecker inhabits the coniferous forest association of the San

Bernardino fauna of the Transition zone; the Abert towhee be-

longs to the mesquite and the quail-brush associations of the Colo-

rado Desert fauna, of the Lower Sonoran zone; the Pacific shrew

belongs to the upland riparian association of the northern coast

redwood fauna of the Transition and Boreal zones.

No. 568] SHORTER ARTICLES AND DISCUSSION 253

CLASSIFICATION OF BARRIERS TO SPECIES AS REGARDS

BIRDS AND MAMMALS

Barriers:

A. Intangible.

(a’) Zonal (by temperature).

(b’) Faunal (by atmospheric humidity).

(c’) Associational.

) By food supply.

(2) By breeding places.

(3) By temporary refuges.

(Each of these three with regard to the inher-

ent structural characters of each species

concerned.)

B. Tangible (mechanical).

(a”) Land to aquatic species.

(b”) Bodies or streams of water to terrestrial species.

The above categories are believed to include all the factors

commonly involved in checking the spread of species of birds and

mammals. It is possible that inter-specific competition may

sometimes occur where associational homologues meet. But even

here it becomes a matter of relative associational fitness which

determines supremacy and consequent ultimate limits of inva-

sion of the forms concerne

A mountain range, sivaithiin ally speaking, is no barrier at all,

per se, as frequently alleged. Only as it involves zonal or faunal

barriers does it affect distribution. The same is true of a valley

or a desert.

As far as contemplation of cases has gone, the writer’s experi-

ence has led him to believe that the outlines of the ranges of all

birds and mammals may be accounted for by one or more of the

factors indicated in the analysis here presented. And as de-

tailed knowledge of the facts of geographical distribution accu-

mulates, the delimiting factors become more and more readily

detectable. By such a study, of comparative distribution, it

seems possible that the ranges of birds and mammals may become

Subject to satisfactory explanation.

When considered in its historical bearing, the problem of

barriers concerns itself intimately with the origin of species. It

is believed by the writer that only through the agency of barriers

Pias multiplication of species, in birds and mammals, brought

about.

254 THE AMERICAN NATURALIST [Vou. XLVIII

The present contribution is abbreviated from a general discus-

sion of certain distributional problems which forms part of a

paper to appear from the University of California press and

which treats in detail of the birds and mammals of the lower

Colorado Valley, in California and Arizona.

JOSEPH GRINNELL

MUSEUM OF VERTEBRATE ZOOLOGY,

UNIVERSITY OF CALIFORNIA

YELLOW VARIETIES OF RATS

In a recent number of the Naturauist I described a yellow

variety of the common rat (Mus norvegicus) which in recent

years made its appearance in England and is now a recognized

variety among fanciers. Dr. John C. Phillips and Professor L.

J. Cole have both called my attention to a fact which I had over-

looked; namely, the occurrence of a yellow variety in another

species of rat (Mus rattus). Bonhote described the occurrence of

this variety in Egypt in 1910 and has since found by experiment

(1912) that the yellow variation of Mus rattus is recessive in

heredity precisely as it is in Mus norvegicus. The fact that the

yellow variation in mice is dominant in heredity, but can not be

obtained in a homozygous condition, stands, therefore, as a phe-

nomenon all the more singular and striking.

W. E. CASTLE.

BUSSEY INSTITUTION,

March 3, 1914.

NOTES AND LITERATURE

HEREDITY AND “THE INKLUENCE OF

MONARCHS”

IN ‘‘The Influence of Monarchs’’ (xiii and 422 pp., 1913, The

Maemillan Co., New York, $2.00) Dr. Frederick Adams Woods

makes a second and firmer step along the path entered on with his

interesting ‘‘Mental and Moral Heredity in Royalty’’ published

in 1906. Dr. Woods’s goal in beginning and continuing his an-

alysis of the character of royalties and the circumstances of their

reigns is one probably not immediately to be reached but also

probably one not impossible of attainment. It is indeed not one

goal that he has before him, but two, the ways to which lie close

together and parallel. One is the establishing of a new science of

history to be called historiometry; the other is the making ap-

parent of the dominance of heredity over environment in deter-

mining human fate.

That the methods and even the aims of most historical study

are not satisfying to all historical students is made obvious by

the constant complaining of historians to and of each other.

There are two conspicuous groups of these protestants, one de-

manding more interest, more imagination, a more literary treat-

ment of historical fact, and the other demanding a more signifi-

cant, more inductive, more scientific treatment. The former

wants more ‘‘humanity,’’ the latter more biology, in history.

Dr. Woods is of the latter group.

But Dr. Woods is not primarily of any historical camp. He is

biologist, especially evolutionist and student of heredity. How-

ever, he marches very boldly into the ranks of the students of his-

torical human history—to distinguish thus the last few thousand

years of human history from the earlier many thousand years

of it—with the new methods and results of his historiometry,

just as Pearson, several years ago, invaded the biological camp

with his biometry. Something of historiometry in history there

has always been, just as there has always been something of

biometry in biology. But these reformers want to make history

and biology wholly, or, at least, most importantly, sciences of

measure. And each of them finds that his use of measure in them

leads him to discover that the facts that he is measuring offer, in

the new significance they are thus made to yield, a special argu-

ment for some particular one of the major factors in evolution.

255

256 THE AMERICAN NATURALIST [Vou. XLVIII

Biometry emphasizes the enormous importance and significance of

variation in all living things; historiometry reveals the enormous

importance of heredity in human life and the affairs of society.

After an introductory chapter stating the need of a new inter-

pretation of history and of new methods of getting at this inter-

pretation, and a following general chapter further elaborating

and expanding his views concerning ‘‘the philosophy of history

and historiometry,’’ Dr. Woods plunges into a series of compact

histories of France, Castile, Aragon, United Spain, Portugal, The

Netherlands, Denmark, Sweden, Russia, Prussia, Austria, Turkey,

Scotland and England. In each of these he presents a swift sum-

mary of the economic and political conditions (success in wars,

increase in territory and prestige, prosperity, advance, failures in

war, loss of prestige, poverty, retrogression) of these nations in

the various reigns of a period of about 500 years for each country,

together with a statement of the personal traits of each monarch.

In all, three hundred and sixty-eight monarchs, regents or other

rulers, royal or non-royal, and correspondingly, three hundred _

and sixty-eight sets, or periods, of national conditions, are pre-

sented.

From these data is derived the very positive and important

conclusion that the dominant causal influence in determining the

character of national, political and economic conditions has been

the personality of the monarchs, and that the prime determinant

of this personality is heredity and not environment.

A host of possible criticisms and objections to the method, its

results and their interpretation, leaps into every one’s mind.

Well, they are all—or all that I have so far been able to formu-

late—anticipated, and ingeniously, and usually convincingly,

answered. At least they are anticipated and discussed. In this

the book reminds one of Darwin’s ‘‘ Origin of Species.”’

To all who have read ‘‘ Heredity in Royalty”? this new book of

Dr. Woods will need no recommendation of its interest and im-

portance. To those who have not, and are interested either as

historian, biologist, or natural philosopher in human history and

the bionomic factors that control it, ‘‘The Influence of Monarchs’’

may be strongly recommended as an original and very suggestive

treatment of the subject. To students of heredity the book is a

necessary library addition.

Wi Ey Be

STANFORD UNIVERSITY,

CALIFORNIA

VOL. XLVIII, No. 569

The American Naturalist

intended for pr gece ere etc., intended for review should be

MSS.

sent to fo igi of THE AMERICAN

rticles arashi summaries of

ALIST, Garrison- -on-Hudson, New York.

research work bearin on th

pro seat = Fecha oer evolution are especially welcome, and wiil be given preference

red reprints of sph herbie are supplied to authors free of charge.

Further pimi will be supplied at c

Subscription

— sements should be sent to the publishers.

subscription price a Nore

Canadian posta

a yea

enty- five pei additional.

The

y cents and

The ahaa for single copies is

Foreign postage is fift

forty cents. The maven ticieg rates are Four Dollars for a pa

THE SCIENCE PRESS

NEW TORR: : Sub-Station 84

Lancaster, Pa.

A.elaad matter, April tth

Garrison, N. Y.

Post Office at Lancaster, Pa., under the Act ot

ARE of March 3, 1879.

FOR SALE

ARCTIC, ee and GREENLAND

BIRDS’ SKIN

Well sung re Low Prices

DINESEN, pok Collect

Husavik, "North Iceland, Via Leidle, pR

WANTED TO PURCHASE

a set of BIRDS OF AMERICA by J. J. Audubon,

7 or 8 volumes, please report, stating cash price, stat-

ing condition, binding and dates of volumes.

F. C. HARRIS,

Box 2244 Boston, Massachusetts

P a

TO OOLOGISTS

and ENTOMOLOGISTS

W. F. H. ROSENBERG

Importer of Exotic Zoological Specimens

57, Haverstock Hill, London, N. W., England

Begs to announce the publication of a new Price

List (No. 20) of Mammals, including over 400 species

from various parts of the world.

will be pe free on application, as well

as any ot i the following lists

Birdskins (over 5000 species); Birds’ Eggs ( cee

,100 species) ; Reptiles, Batrachians and Fish

lone 400 species); Exotic Lepidoptera (over 8, 000

apecies

Largest stock in the world of specimens in all

branches of Zoology.

All Museums and Amateurs

should write for these Lists

All specimens sent on approval.

Please state which lists are required and give

Photographic Films Developed

5 cents per Spool, Prints 3 cents up, Post

Cards

8x10 Enlargements 25 cents mounted.

Send for Samples and Price List.

FRANK R. BARBEAU

89 West Bridge Street Oswego, N. Y.

a ee ee

name of this periodical.

eT SE

The paca ad of Chicago

fiers instruction during the Sum-

a Quarter on the same basis he

ing the t

academic year.

ngan colleges, the

ia “ungen and the. profes

kent

wi fedi

given y regular members of =

po ee 7 tal which is augmen n

the summer by, aS Dn

Summer Quarter, 1914

The Univesity of Chicago

Mitchell Tower. Chicago, Minois as

THE

AMERICAN NATURALIST

VoL. XLVIII May, 1914 No. 569

ECTOPARASITES OF MAMMALS

PROFESSOR VERNON LYMAN KELLOGG

STANFORD UNIVERSITY, CALIFORNIA

THE wingless permanent ectoparasites of mammals

are chiefly of two groups, namely, the Mallophaga, or

biting lice, which feed on the hair and dermal scales, and

the Anoplura, or sucking lice, which feed on the blood.

Certain mites and ticks, a few of the Pupipara (degener-

ate flies) and almost all of the fleas are also ectoparasites

of the mammals, but the fleas, numerous and economically

important as they may be, are not permanent parasites,

for they live as larve not on the host of the adult, but in

cracks and crevices in floors, or in the soil and elsewhere

that the organic detritus used by them as food may be

found. The adults, too, hop on and off their host, and

often change from one‘ host individual to another, and

even from one host species to another. So that the prob-

lems of distribution and species-forming with which I

am particularly concerned in my studies of the ecto-

parasites are not at all the same in such impermanent

form as the fleas as in those truly permanent forms, the

Mallophaga and Anoplura.

Tn these latter there occurs an extraordinary limitation

of the parasite individuals and their immediate progeny

and future generations to specific and even individual hosts

(and their progeny and future generations), so that the

Mallophagan and Anopluran fauna of any mammal usually

represents a closely inbred family strain biologically iso-

257

258 THE AMERICAN NATURALIST [Vou. XLVIII

lated from the rest of the individuals comprising the par-

ticular species represented by it. This brings about cer-

tain striking conditions of abundant small variation and

subspecific (or intraspecific) distinction, which, however,

because of the general similarity of habitat, food and

habit, do not tend to grow rapidly into large (specific,

generic, family) differences. The hundred or more species

of Mallophaga so far recorded from mammals have, until

very recently, all been ascribed to two genera, of which one

included nearly nine tenths of the total number of kinds.

There has been made a beginning—and not a particularly

convincing one—at breaking up this inclusive genus

(Trichodectes). It is a movement suggested more by the

needs of convenience than the needs of expressing a bio-

logical situation. Similarly, although not representing

so extreme a condition of likeness, the Anoplura, also

including about a hundred parasite species (occurring

only on mammals) have been, until recently, divided into

but half a dozen genera, with the great majority of the

species included in one. Certain aberrant forms found

on man, the monkeys, the elephant, and on seals and

walruses have always made necessary the recognition of

four or five quite distinct genera. Attempts, however, are

now being made to break up the unwieldly genus Hema-

topinus.

As this paper is, in effect, a continuation of my paper

on ‘‘ Distribution and Species-forming of Ecto-parasites’’

published in THe Amertcan Naturauist in March, 1913,

which devoted itself to a consideration of the Mallophaga

(some 1,400 species as so far known) found on birds, and

to the problems presented by their conditions of life and

their host and geographic distribution, I can dispense

with any further account of the special biology of these

parasites by referring the interested reader to this

former paper. In it I have set out rather fully the spe-

cial structural and habit features of the Mallophaga.

Except that the Anoplura take blood, rather than

feathers and hair, for food, and have specially modified

No. 569] ECTOPARASITES OF MAMMALS 259

mouth parts to do it with, and are perhaps even more

specialized in their physiological adaptations to their

host than the biting lice, most of the general remarks

made concerning the Mallophaga will apply to the suck-

ing lice also.

In their peculiar special relations to their hosts as per-

manent ectoparasites on them, wingless, and reluctant to

migrate even with opportunity, and so fitted physiologi-

cally to their parasitic life that they can not live for more

than a few hours (or, at most, and exceptionally, days)

off the bodies of their hosts, the Anoplura and Mallophaga

are alike. And hence the conditions and problems of

their distribution and species-forming are practically the

same for the two groups.

The thesis that I have maintained, on a basis of the

conditions presented by the bird-infesting Mallophaga,

I now wish to test by the conditions presented by the

mammal-infesting Mallophagaand Anoplura. This thesis

is, in fewest words, that the host distribution of these

wingless permanent ectoparasites is governed more by

the genetic relationships of the hosts than by their geo-

graphic range, or by any other ecologic conditions. The

fact, proved by abundant cases, that two host species of

wholly distinct geographic range and with no possible

opportunity for contact such as would permit of the

migration of wingless parasites from one to the other,

may have, nevertheless, one or more parasitic species

common to them both, is associated almost always with

the further fact that these common hosts are closely

related genetically. They are most often of the same

genus or of closely allied genera; they are almost cer-

tainly always of the same subfamily or family. The ex-

planation for the possibility and the reality of this inter-

esting host distribution I find in the hypothesis that the

common parasite species has persisted unchanged from

a common ancestor of the now divergent but allied host

kinds,

Also, if it be true that genetic relationship is the deter-

260 THE AMERICAN NATURALIST [Vou XLVIII

mining factor in accounting for the host distribution of

the parasites, then it is also true that the distribution of

the parasites will indicate in some measure the genetic

relationships of the hosts, and that occasional aid in

determining the genetic affinities of birds and mammals

of doubtful relationships may be had from a study of

their parasitic fauna. In my paper already referred to

I have pointed out some suggestive cases of this sort in

connection with the birds and their parasites.

In examining the conditions existing among the mam-

mals and their Mallophagan and Anopluran fauna, the

first necessity was the compilation of a complete wecord

or catalogue of mammalian hosts and their parasites,

together with the record of the actual locality of each

finding of parasites, together with a general record of the

geographic range of all the various hosts. This cata-

logue, or set of records, I have now completed, and

despite its meagerness compared with the similar cata-

logue of the bird hosts and their Mallophagan parasites

from which the notes for the former paper were drawn,

it contains enough records of interest to make worth

while a preliminary report on the condition obtaining

among the mammals and their parasites.

It is unfortunate that, although there are nearly one

fourth as many mammal species as bird kinds, only about

one hundred mammals figure in the Mallophagan host

list, while Mallophagan parasites have been taken from

over eleven hundred bird species. Also, only one hun-

dred different Mallophaga have been taken from mam-

mals, while about fourteen hundred have been taken

from birds. Of the Anoplura, which are found only on

mammals, records have been made from about one hun-

dred host species, these records referring to just about

the same number of Anopluran kinds. Thus the mam-

malian host catalogue with its list of parasites is a short

one; as far as it goes, however, it is thoroughly interest-

ing and suggestive.

In working up the records I have used Trouessart’s

No. 569] ECTOPARASITES OF MAMMALS 261

‘*Catalogus Mammalium’’ as an authority for the synon-

omy of the hosts, and my own judgment, based on a con-

siderable personal knowledge of the parasites and on a

careful consideration of all the more intelligible litera-

ture of the two groups, as a last court for the synonomy

of the Mallophaga and Anoplura species. The synonomy

of the parasites I have, however, not pushed far.

With so much of introduction and explanation we may

come to a swift résumé of the results of a scrutiny of

these records, proceeding by sequence of the mammalian

orders, and referring to either or both groups of para-

sites as they may happen to be represented in the para-

site records of the successive host groups.

The Marsupialia are represented in the host list by

half a dozen species of kangaroos and wallabies (family

Macropide) all from Australia, and a wombat, Phasco-

lomys ursinus (family Phalangeride), from Tasmania

(also S. Australia?). From all of these hosts only Mallo-

phaga are recorded, no Anoplura having yet been taken

from a marsupial. The six species of kangaroos repre-

sent three genera (Macropus, Petrogale and A¢prym-

nus), and their Mallophaga are of seven species, repre-

senting four genera. Four of the species belong to the

genus Boopia, and I strongly suspect are not all different.

In addition there is one Trichodectes, from Petrogale

penicillata, one Latumcephalum, from ‘‘wallabies,’’ and

one Heterodoxus, which is recorded from Macropus

giganteus in Australia as well as from the same host in

the Jardin des Plantes, Paris. It is also recorded from

an undetermined wallaby in Victoria and one in Queens-

land, as well as appearing in three other records from

‘‘kangaroo”’ or ‘“‘wallaby’’ from Australia. The para-

site of the wombat is a species of Boopia, and it has been

twice recorded from the same host. It is interesting that

the kangaroo in the Jardin des Plantes harbored, even

after some period of captivity, only its own proper para-

262 THE AMERICAN NATURALIST [Vou. XLVIII

sites without accepting new ones from its many, various

and closely pressing neighbors.

Of the four Mallophagan genera found on the kanga-

roo, three, namely, Boopia, Latumecephalum and Hetero-

doxus! are peculiar to them. The third genus, Tricho-

dectes, is represented by but a single species which has

been recorded but once. This is the common Mallophagan

genus of mammals generally. The record is perhaps a

good one, but its lack of confirmation by being unrepeated

either for the same species or for any other species of

Trichodectes, is suggestive. Heterodoxus, Latumcepha-

lum and Boopia are two-clawed genera; that is, they are

Mallophagan forms which belong to a family all the other

genera of which are confined to birds. The characteristic

- structural difference between the mammal-infesting

Mallophaga and the bird-infesting species is the presence

in the first group of a single claw on each tarsus, and in

the second of two claws. This difference is plainly an

adaptive one concerned with the fitting of the foot for

the seizing of hairs and scrambling about among them,

on the one hand, and the manipulation of feathers and

moving about on them, on the other. In examining living

specimens under the microscope the special use and fit-

ness of the feet, in the one case adapted to hairs and in

the other to feathers, is obvious. However, Heterodoxus,

Latumecephalum and Boopia, and, in addition, perhaps

one other doubtful genus, represented by one species, and

perhaps two or three species of another two-clawed

genus, constitute exceptions to the general rule. It is of

decided interest to note that the only genera of two-

clawed Mallophaga found exclusively on mammals are

limited to the Marsupials. The antiquity and isolation of

1 The single valid species of this genus—the two or three that have been

named are undoubtedly all the same—has also been recorded from dogs! In

fact specimens in my own collection were received with the record ‘‘from

Japanese dog.’’ And Enderlein has recorded it from a dog from China and

Neumann from a dog from Formosa. Yet dogs panera do not harbor

this parasite, and kangaroos and wallabies do. It seems necessary to be-

lieve that the dog host records SS eases of incite from kangaroos

in zoological gardens or menageri

No. 569] ECTOPARASITES OF MAMMALS 263

this host group strongly suggests that the one-clawed con-

dition common to all other mammal-infesting Mallophaga

is a derivative from the original two-clawed condition

characteristic of the parasites of birds and of these ancient

mammals. The two-clawed condition is, of course, the one

common to insects generally and is characteristic of the

Atropids, in whom I am inclined to see the ancestors, or

near-ancestors, of the Mallophaga. All of the Anoplura.

it may be added, which are exclusively mammal-infesting,

are one-clawed.

In this connection the suggestiveness of the fact that

in face of the examination of many specimens of half a

dozen species of kangaroos and wallabies, no Anoplura

have yet been found on the Marsupials, may be referred

to. I am coming strongly to believe that there is no such

wide ordinal separation of the Mallophaga and Anoplura

as our clinging to the fetich of ‘‘biting and sucking

mouth-parts’’ as basis for radical classificatory separation

has led us to effect. I believe, with Mjöberg, that the two

groups of parasites have a fairly near genealogical

affinity, their differences, which are particularly those of

mouth-parts, being adaptive rather than palingenetic in

character. The Anoplura have gone on from the Psocid-

Mallophagan condition to a more specialized parasitic

habit, and are the extremes of a general line of ecto-

parasitic evolution. The absence of sucking lice from the

kangaroos may mean that the Marsupials are older than

the Anoplura! No other considerable group of mam-

mals, except certain families of strong-smelling Carni-

vora, is free from the blood-sucking parasites.

There are but two Edentates in the host list, one, the

Cape Ant bear, Orycteropus afer (family Oryeteropo-

dide) of south and central Africa, harboring a sucking

louse, of genus and species peculiar to it, and the other,

the three-toed sloth, Bradypus tridactylus (family Brady-

podide) of eastern South America, harboring a Mallo-

phagan of species peculiar to it but of the genus Gyropus

which is the less scattered, although still rather catholic,

264 THE AMERICAN NATURALIST [Vou. XLVIII

genus of the two large ones characteristic of the

mammals.

The large order Ungulata, with its numerous domesti-

cated and semi-domesticated species, is a favorite host

group with both Mallophaga and Anoplura. Altogether?

about thirty Anoplura and two dozen Mallophagan spe-

cies are recorded from fifty host species representing nine

Ungulate families.

The family Elephantide is represented by the African

and Indian elephants, recognized as distinct species of

distinct geographic range. They both harbor a common

Anopluran species, Hematomyzus elephantis, of species,

genus and family peculiar to the elephants. Fahrenholz

has given the varietal name sumatranus to specimens of

these sucking lice taken from an Indian elephant in

Sumatra. Records show that the parasites have been

taken from their elephant hosts not only in Africa and

Asia, but in various zoological gardens, as Paris, Ham-

burg and Rotterdam.

The small family of Hyracide, or conies, is represented

in the host list by two species and perhaps a third one,

one of which, the Syrian coney of west and south Asia,

harbors one Anopluran and one Mallophagan, while from

the other, the Cape coney of South Africa, the same

Anopluran species is recorded as well as another of the

same genus. This record of a second species is from a

coney in the London Zoological Gardens. From the pos-

sible third species of Hyrax (taken in the African Congo

and perhaps, but not probably, also a Cape coney), a

second Mallophagan species is recorded of the same

genus, Trichodectes, to which that of the Syrian coney

belongs.

In the family Equide three species, the horse, the

donkey and Burchell’s zebra, all suffer from the infesta-

tion of a common Anopluran species, Hematopinus asini.

In addition, the horse and the zebra have a common

2 The synonomy in the parasite records, and indeed in the host records

as well, is a vicious tangle. I have done the best I can, for the present.

No. 569] ECTOPARASITES OF MAMMALS 265

Mallophagan parasite, Trichodectes parumpilosus, while

the horse and donkey have another common biting louse,

Trichodectes pilosus. Two varieties of Trichodectes

parumpilosus have been named by Piaget, one from the

zebra and another from ‘‘little horses of Java.’’

The pigs (family Suide), of which three wild African

species besides the familiar animal of the barnyard are

found in the host list, are infested by two (perhaps three)

species of Anoplura and one (a not too certain record)

Mallophagan. Hematopinus suis is found on the domes-

tic Sus all over the world, while Hematopinus latus of

Neumann, H. phocochoeri of Enderlein and H. peristictus

of Kellogg and Paine, which are almost certainly all one

species, are recorded from the wart hog, Phacocherus

ethiopicus from Nyasa-land, Africa, and probably also

from another wart hog species from Africa, and the Red

River hog, Potamocherus cheropotamus from Nyasa-

land, Africa. In addition Potamochewrus demunis (prob-

ably), from German east Africa, is credited by Stobbe

with a Mallophagan parasite peculiar to it, Trichodectes

vosseleri Stobbe.

The peccary, Dicotyles tajacu (family Dicotylide) of

Central America and southwestern North America, has a

Mallophagan species peculiar to it, belonging to the

smaller of the two large Mallophagan genera, namely,

Gyropus.

The dromedary, of north Africa and western Asia, and

the bactrian camel, of central Asia, harbor a common

sucking louse, Hematopinus cameli. A doubtful second

Species called H. tuberculatus (Neumann thinks it iden-

tical with cameli) has been recorded from a dromedary

imported from India into Australia. The ‘‘South Amer-

ican camel,’’ the llama, harbors an Anopluran species

peculiar to it, and two Mallophagan species, Trichodectes

breviceps Rudow and T. inequalemaculatus Piaget. Al-

though Rudow’s species are often suspect, I have just

had his breviceps from a llama of Peru (collector C. H.

266 THE AMERICAN NATURALIST [Vou. XLVIII

T. Townsend). With these llama Mallophaga there is

also a small Anopluran which I have not yet worked out.

The family Cervide is represented in the host list by

about ten species. They are infested by three species of

Anoplura, each peculiar to its host, and six species of

Trichodectes (Mallophaga) of which T. tibialis is com-

mon to the roe deer of Europe and Asia Minor, an

African Capreolus, and our own black-tailed deer of the

western states. Trichodectes longicornis is common to

the red deer of Europe and Asia Minor and the fallow

deer of south Europe, Asia Minor and north Africa.

The giraffe (family Giraffide) harbors a sucking louse,

Linognathus brevicornis, peculiar to it.

The great family Bovide, with its many buffalo, buck,

sheep, goat and antelope kinds, is represented in the host

list by five or six species of Bos, four African bucks,

three or four sheep, the ibex, chamois and two or three

goats, and five or six antelopes, or gazelles. The domes-

tic ox, Bos taurus, harbors three species of Anoplura and

one Mallophagan. Curiously, none of these species is

recorded from any other Bos. On the other hand, the

zebu, the Indian buffalo, and the American bison all

have the same Anopluran species (and no other, nor any

Mallophagan), while the yak of central Asia and the

Kaffir buffalo each have an Anopluran peculiar to it.

The four species of African reedbucks and duikerboks

have, according to the records, each a peculiar species of

sucking louse. These records need scrutiny. One of

them is my own, but I had to describe the species without

seeing the types of the others. The domestic sheep

carries two Anopluran species and one Mallophagan.

The latter occurs also on at least two wild species of

Ovis, one of west Africa and the other of north Africa.

The fat-tailed sheep has a record from German south-

west Africa of a Trichodectes of its own.

The domestic goat harbors one Anopluran and at least

one Mallophagan, the latter being common also to the

Angora goat, the chamois, and a wild (?) goat of Guinea,

No. 569] ECTOPARASITES OF MAMMALS 267

and a wild (?) goat of Java. <A recent description of a

second Mallophagan species from the domestic goat is

not convincing. The chamois has also an Anopluran, but

one, so far, peculiar to it. Three species of Gazella (or

Antilope) have three species of Trichodectes, of which

one is common to two host species, one of Arabia and

Syria and the other of north Africa and southwest Asia

generally. This same Trichodectes is also recorded from

the roan antelope, Hippotragus equinus, of east central

and south Africa. One species of Gazella carries an

Anopluran peculiar to it, as does also Tragelaphus gratus

of west Africa.

The order Carnivora is represented in the host list by

eight families and a total of fifty-four species. Only one

species of Anopluran, the common sucking louse of the

dog (not found yet even on the wolf or fox, both of which

have other records) is recorded from a Carnivore, outside

of the two families Trichechide (walruses) and Phocide

(seals and sea-lions). From these two families, on the

other hand, only Anoplura are recorded.

The family Felide is represented by three species, the

domestic cat, the California lynx and the tiger. The cat

and lynx have a common Mallophagan parasite, T'richo-

dectes subrostratus (and no other), while the tiger has a

biting louse presumably peculiar to it. The description

of this parasite is, however, very brief and unsatisfactory.

The family Viverride, mongooses, ichneumons and

genets, is represented in the host-list by eight species, of

which five are of the genus Herpestes. Two of these

Herpestes species, one of southern Spain, north Africa

and Asia Minor, the other of west, east and south Africa,

harbor a common Mallophagan parasite. A record of

the finding of Trichodectes subrostratus, the familiar

biting louse of the cat, on Herpestes pluto, comes from

the Kameroons (Africa). It is probably a case of

straggling, the mongooses being common enough in gar-

ens, and some of them fairly domesticated.

Of the family Canidex there are records from eleven

268 THE AMERICAN NATURALIST [Vou. XLVIII

species, including the domestic dog, a wild dog of South

America, two wild dogs of Asia, two foxes, and a wolf.

The domestic dog has a familiar sucking louse and is also

credited with that problematical normal or straggling

biting louse of a peculiar genus which I have referred to

in my account of the parasites of the kangaroos Tricho-

dectes latus, the common biting louse of the domestic dog,

is also common to the wolf, Canis lupus, of Europe and

Asia, and to the raccoon-like wild dog, Nyctereutes pro-

cyonoides, of Asia and Japan. The record of this last

came, it must be noted, from the Berlin Zoological Gar-

dens. There is no other record of commonness of para-

site to two hosts in the family. The English fox has a

single Mallophagan species, and the California fox has

another. The dhole, a wild dog of the Himalayas, has a

Mallophagan species, and the Magellan wolf of Patagonia

has another.

The family Procyonide is represented in the host-list

by two raccoons, the California ring-tailed cat, and two

coatis of Central and South America, respectively. The

two raccoons, Procyon lotor of North America and Pro-

cyon psora of California, harbor a common Mallophagan

parasite. In addition a German record (from a zoolog-

ical garden?) credits Procyon lotor with carrying also a

Mallophagan which is the characteristic parasite of the

badger. On the California ring-tailed cat, Bassariscus

astuta, have been found two Mallophagan species, one of

which is the characteristic parasite of the skunks of

North and South America. The two coatis, Nasua narica

and Nasua rufa, one of southwestern United States,

Mexico and Central America, and the other of South

America from the equator south, both harbor a common

Mallophagan species.

The family Mustelide, comprising the badgers, wea-

sels, martens, and skunks, an ill-smelling crew, offers no

attraction to blood-sucking parasites, but is represented

in the host-list by nearly twenty species from which ,

Mallophaga have been taken. The Old World badger has

No. 569] ECTOPARASITES OF MAMMALS 269

a characteristic species, Trichodectes crassus. The mar-

tens, weasels and ermine have also a characteristic spe-

cies, Trichodectes retusus, which is recorded from the

pine marten of Europe and northern Asia, the beech

marten of the same range, still another Old World mar-

ten, the weasel of Europe and Asia, the ermine of north

Europe, Asia and America, and the weasel and mink of

North America, in all six or seven species of Mustela and

Putorius of very wide geographic range. The skunks of

North and South America have also a characteristic

Mallophagan species, Trichodectes nephitidis, described

by Osborn from the common North American skunk,

Mephitis mephitica, taken in Nebraska. I have found

this parasite on the western skunk, M. occidentalis, in

California, and on M. macrura of Arizona. It has also

been recorded from the spotted skunk, Spilogale inter-

rupta, of the southern United States, Mexico and Central

America, and I have examples from a ‘‘skunk’’ of Bolivia.

It is also recorded from a Chilian Mustelid, Galictis

quiqui, which ranges over South America from the River

Plate south, and from another species of Galictis in

Brazil. Finally, examples of this ubiquitous pest are

recorded from Helictis everetti from North Borneo! The

_ last record comes from Neumann, a very careful and

well-informed student of the parasites, but his specimens

were taken from a skin in the Museum of Natural History

of Paris. The Old World otter, Lutra lutra, has a Tri-

chodectes of its own, as has also an African otter, L.

matschiei, and the North African Zorilla lybica.

Mjöberg records a species of Boopia (typical kangaroo

parasite genus) from Lutra pruneri of India. As the

record is an extraordinary one, being the only case of a

Boopia found outside of Australia or on a mammal other

than a Marsupial, it is well to note the exact circum-

stances of the record. The parasites (several examples)

were got by Mjöberg from the Hamburg Zoological Mu-

seum where they were ticketed as having been taken

from a ‘‘soeben frisch angekommenes Thier’’ of the

Species Lutra pruneri, the animal having been received

270 THE AMERICAN NATURALIST [Vou. XLVIII

from India. There are to be considered in connection

with this extraordinary record, first, the possibility of an

exchange of labels in the course of the several handlings

of the Mallophagan specimens, and, second, the possibility

of a favorable answer to the question: Is Lutra pruneri,

which does not appear in Trouessart at all, only Lutra

lutra, the common Old World otter, and was the speci-

men from which the Mallophagan came a resident in a

zoological garden in which kangaroos or wallabies also

lived, affording a bare chance of straggling? The similar

aberrant records from dogs of the kangaroo parasite

Heterodoxus have already been referred to.

The bears (family Urside) have, so far, but one para-

site record to present, a Mallophagan species, Tricho-

dectes pinguis, having been described from the Thibetan

bear, Ursus thibetanus, a century ago.

The walrus (family Trichechide) harbors a strange

Anopluran parasite of species, genus and family peculiar

to its host, as, indeed, might be expected of any ecto-

parasite daring enough to brave comrade life with wal-

ruses. Examples of the parasite have been taken from

walruses from Spitzbergen, Frobisher Bay (Davis

Straits), the Hamburg Zoological Garden, and I have

recently had them from a ‘‘ Pacifie walrus’’ from ‘‘south-

east of Siberia.’’

The family Phocide is represented in the host-list by

at least five species of seals and sea-lions carrying an

equal number of Anopluran species representing three

different genera, all of them peculiar to the seals. A

single parasite species, Echinopthirius phoce has been

repeatedly taken from the fur seal, Proca vitulina, from

both Old World and New World shores. The harp seal

of the Arctic is credited with the same parasite, as well

as another. Hooker’s seal of New Zealand and the Auck-

land Islands carries an Anopluran, Antarctopthirius

macrochir, of species and genus peculiar to it, while the

elephant seal of the south Pacific has another parasite

also of genus and species peculiar to it.

The large order Rodentia is well represented in the

No. 569] ECTOPARASITES OF MAMMALS 271

host-list, representatives of thirteen families, summing

about sixty species, being listed. Both Mallophaga and

Anoplura infest the rodents, but certain families are

parasitized almost or quite exclusively by Anoplura,

while Mallophaga are the only parasites of others.

The Sciuride (squirrels and spermophiles), for ex-

ample, with a dozen host species, are parasitized by a

dozen species of Anoplura with only a single Mallo-

phagan record; and a single record under such circum-

stances is always suspect. There is little commonness of

parasite species to two or more host species in this

family. Osborn’s Polyplax montana is recorded from

the eastern and western North American gray squirrels,

and his P. suturalis has been taken from two Spermo-

phile species, both, however, of the same general range.

The well-differentiated parasite genus Acanthopinus is

represented by one species from the common Old World

squirrel, Sciurus vulgaris, and another from the eastern

gray squirrel of North America. These species, though

close together, really seem to be different. In addition I

have just found the Acanthopinus species of the eastern

gray squirrel on Douglas’s squirrel in California, and

another (new) species on a California chipmunk. The

only Mallophagan species recorded from a Sciurid is

Gyropus turbinatus from the marmot, Arctomys mar-

motta, of the mountains of southern Europe.

From the beaver (family Castoride) a characteristic

Mallophagan species, Trichodectes castoris, has been

taken in America. The beaver, it may be noted, is the

host of the only beetle (Platypsylla castoris) that has

become a specialized permanent ectoparasite, passing its

whole life on the body of its host.

The Old World dormouse (family Gliride or Myoxide)

harbors a sucking louse, Polyplax pleurophea.

he large family Muride, including the rats, mice,

voles and lemmings, is represented by twenty host species

well scattered over the world. There are twenty-two

Anopluran species and two Mallophagan species in the

parasite list for the group. Both of these Mallophagan

212 THE AMERICAN NATURALIST [Vou. XLVIII

records are my own. One is a new species of Colpoce-

phalum (exclusively a bird-infesting genus) from a

‘“spotted rat,’? Uganda, Africa, sent me by Sjoestedt in

a collection made by the Swedish Zoological Expedition

to Kilimandjaro-Meru, Africa, in 1905-1906. It is un-

doubtedly a straggler from some bird taken at the same

time. The other is a poor specimen of Trichodectes from

Mus rattus, Canal Zone, Panama, sent me by Dr. Jen-

nings. It may bea good record—or it may be a deceiving

one. Both record and specimen need further scrutiny.

It is, perhaps, important to note that two specimens of

a wingless Psocid (Atropide) were sent with the lot

labeled ‘‘parasites from Mus rattus.” It would be very

interesting if we could know that these Atropids were

really living on the rats, feeding on their hair or dermal

scales. I have found Atropids in rats’ nests and birds’

nests living undoubtedly on the loose hairs, feathers and

dermal exuvie. It is my belief, based primarily on cer-

tain striking facts of morphology, that the Mallophaga

are degenerate descendants of the Rsocide.* Of the

murid Anoplura, two or three are common to several

hosts, as the well-known Polyplax spinulosa, recorded

from all over the world from the now cosmopolitan Mus

rattus and Mus decumanus, as well as from Mus syl-

vaticus of Europe and north Asia, and Mus alexandrinus

of south Europe and Asia Minor (perhaps only a variety

of Mus rattus), and Polyplax affinis (perhaps only a

variety of P. spinulosa) recorded from Mus agrarius of

eastern Europe, and Mus sylvaticus of Europe and north

Asia. Polyplax (Hoplopleura) acanthopus, the common

sucking louse of the mouse has been taken from the now

cosmopolitan Mus musculus, and also from Lemmus tor-

quatus, the lemming of Arctic Europe, Asia and America,

Microtus agrestis, the field vole of Europe, Microtus arva-

lis, another common vole of Europe and Asia, and Micro-

tus sp. from Iowa, U. S. A. The water rat, Hydromys

chrysogaster, of Australia, has a Polyplax species of its

own as has also Otomys bisulcatus of south and central

3 See Psyche, Vol. 9, 339, pp. 1902.

No. 569] ECTOPARASITES OF MAMMALS 273

Africa, Hesperomys leucopus of North America, Epimys

aurifer of the Malay Peninsula, Gerbellus indicus of

northern India and Afghanistan, and Holochilus sciureus

of Brazil and Peru. The common Old World mouse, Mus

minutus, harbors three Anopluran species, while Mus

musculus has but two. The Old World water vole, Micro-

tus terrestris, has a parasite differing from the two in-

festing respectively the two Old World land species of

Microtus.

In connection with this résumé of the Murid parasites,

I may say that I have now in process of working over

some two hundred vials of material collected last summer

from California mammals, which is going to add many

records to the Murid list of both hosts and parasites. It

will also add numerous records for the squirrels and

spermophiles (Sciuride).

The family Geomyide, gophers, is represented in the

host list by three North American and one Central Amer-

ican species. The Mallophagan species Trichodectes

geomydis oceurs on all of these hosts. The North Amer-

ican hosts are Geomys bursarius (Iowa), Thomomys

botte (California), Thomomys bulbivorous (California),

and the one Central American host is Macrotomys hetero-

dus (Costa Rica). T. bulbivorous may be a synonym of

T. botte. In addition, Geomys bursarius has yielded an

Anopluran species of genus and species peculiar to it.

The pocket rats, family Heteromyidæ, are represented

by a species of Perognathus (Baja California), and

Dipodomys merriami (Arizona). From both are re-

corded the same Mallophagan species, Trichodectes

californicus.

The jerboa, Dipus sp., is the sole representative of the

family Dipodidæ. From it is recorded an Anopluran

Species taken in Tunis.

The Octodontidæ are represented by three species

parasitized by one Anopluran and three different Mallo-

phaga. The three hosts are of three different genera, one

with an African range, the other two of South America.

The parasite species on each is peculiar to it. A third

274 THE AMERICAN NATURALIST [Vou. XLVIII

record, crediting the characteristic Trichodectes pilosus

of the horse to a coypou of South America (in the menag-

erie of the Jardin des Plantes in Paris), is certainly

either a false record or one of rather extraordinary

straggling. The two Mallophagan species from these

South American tuco-tucos belong to the genus Gyropus,

which is the Mallophagan genus especially characteristic

of the related South American families, the Caviide

(guinea-pigs), the Dasyproctide (agoutis), and the Chin-

chillidæ (chinchillas and vizeachas) (see following

paragraphs).

The guinea-pigs and mocos (family Caviide) are repre-

sented by three species, and are strongly parasitized by

Mallophaga. They have no Anoplura. The domesticated

form, which is variously held to be a species distinct

from any wild one now known, or a variety of the wild

species, Cavia cutleri, harbors two well-known species of

Gyropus, namely G. ovalis and G. gracilis (this latter is

held by some students to be of distinct genus). In addi-

tion, Piaget has described a species of Menopon (bird-

infesting genus) from it, and Paine and I have described

another Menopon from it from collections we have had

from Peru and Panama. We have also found this latter

species on the wild guinea-pig, Cavia cutleri, from Peru,

and from this host Paine has described a species of

Gyropus peculiar to this host. From the Brazilian moco,

Kerodon moco, has been recorded a variety of Gyropus

gracilis, one of the familiar species of the domestic

guinea-pig, as well as another species of Gyropus peculiar

to the moco. Recently Cummings has described a new

Mallophagan taken at Villa Rica, Paraguay, from the

wild guinea-pig, Cavia aperea. For this new species he

established a new genus called Trimenopon. As a matter

of fact the species is so much like Kellogg and Paine’s

Menopon jenningsi, except for its markedly larger size,

that I am not at all sure it should be added as a fourth

guinea-pig parasite.

A single agouti, Dasyprocta aguti, from Brazil, repre-

sents the family Dasyproctide. From it have been de-

scribed two species of Gyropus peculiar to it.

No. 569] ECTOPARASITES OF MAMMALS 275

The chinchillas and vizeachas (family Chinchillide,

or Lagostomide) are represented in the host list by two

species, to which I can add another (perhaps two others)

on the basis of material recently received from Dr. C. H.

T. Townsend, of Peru. From Lagidium peruanum Gay

long ago described a peculiar Gyropus, and I have speci-

mens of a Gyropus which may or may not be different

from Gay’s species. His description is very meager. In

addition I am about to describe, under the name Philan-

dria townsendi, another species, representing also a new

genus, specimens of which have been sent me by Dr.

Townsend from the same host. Also in this Townsend

sending are specimens of a small Polyplax species (Ano-.

pluran) from the same host.

The Cercolabide or Coendide, American porcupines,

are represented in the host lists by five species, three of

Central and South America and two of North America.

They harbor no Anoplura, but are parasitized by two

Mallophagan species, of which one, Trichodectes setosus,

occurs on all the host species in the list. The second

Mallophagan is a Trichodectes recently described by

Stobbe from Cercolabes nova-hispanieé of Mexico and

Central America. The other South American host poreu-

pines are Coendu (Cercolabes) prehensilis (northern

South America) and C. villosus (Brazil). The North

American hosts are Erethizon epixanthum (California)

and E. dorsatum (Nebraska).

Finally the family Leporide, hares and rabbits, ap-

pears in the host list with six (perhaps only five) repre-

sentatives, of which four, namely, Lepus timidus, of cir-

cumpolar arctic regions, Lepus cuniculus, native to

Europe and north Africa but introduced over the whole

world, Lepus europeus of Europe and Lepus campestris

of western Canada and United States, harbor the same

species of sucking louse, representing a genus peculiar to

hares and rabbits. I must note that this species, Hema-

topinus ventricosus Denny, is commonly referred to as

two species, of which one, H. ventricosus, is recorded

from the American host species and L. cuniculus, while

276 THE AMERICAN NATURALIST [Vou. XLVIII

the other, called H. lyriocephalus, is recorded from L.

timidus and L. europeus. But Neumann, an exception-

ally experienced student of the Anoplura, holds that the

two species are one. A deer-infesting Mallophagan, Tri-

chodectes tibialis, certainly a straggler, has been recorded

from Lepus europeus, and another Trichodectes (a very

old and uncertain record) from Lepus cannabinus.

The order Insectivora is represented by but two spe-

cies, the mole, Scalops argentatus, of North America,

and the shrew, Sorex araneus of Europe and Asia. Each

harbors an Anopluran species, that of the mole being a

curiously modified form and of species and genus peculiar

-to its host, while that of the shrew is of a species not

found on other hosts.

The order Prosimiæ, the lemurs, presents a single

record, that of a species of Mallophagan, Trichodectes

mjobergt Stobbe, described from the North Bornean

Nucticebus borneanus (family Nycticebide).

The order Primates is represented in the host list by

four families, the Cebide of the New World, the Cerco-

pithecide, the single family of apes, Simiide, of the Old

World, and the family of man, Hominid. The distribu-

tion of the ectoparasites of these groups is of unusual

interest to the special student and will likely prove equally

so to more general students.

The Cebide, platyrrhine, tailed, New World monkeys,

are represented by two species, the spider monkey and

one of the howling monkeys of Brazil, members of differ-

ent genera, each with a Trichodectes species peculiar to

it. In addition three species of Ateles, one of Mexico and

Central America, another of Guiana and Brazil, and the

third an undetermined species of the genus represented

by a specimen in a traveling menagerie in Europe, have

yielded three species of the Anopluran genus Pediculus,

otherwise characteristic of man and the anthropoid apes.

These three Pediculus species have been recorded and de-

scribed by three different students of the group, all careful

workers, and there can be no doubt of the generic refer-

ence. But it is to be noted that the specimens of all three

No. 569] ECTOPARASITES OF MAMMALS 277

parasite species were obtained either from host skins in

a museum (in one case the Zoological Museum of Ham-

burg, in another, the Berlin Museum) or from a live host

ina menagerie. In no case, therefore, is the possibility of

a straggling record wholly excluded, but the coincidence

of three discoveries makes the records practically safe.

Finally, in this connection it is to be noted (as I have

already pointed out in a brief paper*), that, although

Ateles is a tailed New World genus and -presumably

widely separated genetically from the anthropoids,

Friedenthal has affirmed, on a basis of blood and hair

comparison, that Ateles shows unmistakable differences

from other tailed monkeys, and resemblances with the

anthropoids, and he suggests that in Ateles we should see

monkeys, which, in a certain sense, replace, in the New

World, the anthropoids of the Old. It is, in any event, a

strange thing that Ateles differs from the other Cebide

and from the Cercopithecide as well, in not harboring the

Anopluran genus Pedecinus to which all monkey-infest-

ing Anoplura, except those of the simians, belong, but in

actually harboring parasite species of the genus found

elsewhere only on the simians and man.

The family Cercopithecide, catarrhine, Old World

monkeys, is represented in the host list by a dozen spe-

cies, from which one Mallophagan species, viz., my Tri-

chodectes colobi from a guereza monkey, Colobus guereza

var. caudatus (East Africa), and ten Anopluran species

have been recorded. Of the Anoplura nine species be-

long to the genus Pedecinus, long recognized as the char-

acteristic genus of the lower monkeys, as contrasted with

the genus Pediculus characteristic of the anthropoid apes

and man. For the tenth species, Fahrenholz establishes

the new genus Pthirpedecinus, just as for one of the

man-infesting species the separate genus Phthirius had to

be established, There are several cases of the common-

ness of a single Pedecinus species to two or three hosts.

P. breviceps Piaget is recorded from Macacus silenus of

*‘<Ectoparasites of the Monkeys, Apes and Man,’’ Science, N. S., Vol.

38, pp. 601-602, 1913.

278 THE AMERICAN NATURALIST (VoL. XLVIII

India, Cercopithecus mona of west Africa, and a third

Cercopithecus skin in the Zoological Museum at Ham-

burg. P. longiceps Piaget is recorded from Macacus

cyclopis of Formosa, Semnopithecus maurus var. cris-

tatus of Borneo, and Macacus cynomolgus of the Malay-

sian region. P. eurygaster Gervais has been recorded

from Macacus sinicus of India and on a macaque in the

Zoological Garden at Sydney, and another in the Zoolog-

ical Garden at Melbourne. A hamadryad (Paphio sp.)

of north Africa has a Pedecinus species peculiar to it, as

has a trachypithecus, of Malaysia, and the Barbary ape,

Macacus innuus, of northern Africa and Gibraltar. The

common Macacus rhesus carries one species of Pedecinus

peculiar to it, and that single species of Phthirpedecinus

already referred to. Macacus silenus also has recorded

from it two species both belonging to Pedecinus.

The family Simiide, anthropoid apes, is represented

in the host list by three species, namely, the chimpanzee

and two gibbons. One of these gibbons is Hylobates

syndactylus of Sumatra; the other is H. leuciscus of

orneo. A single species of Pediculus is common to

them both, and is not elsewhere recorded. The chimpan-

zee has also a single species of Pediculus which is pecul-

iar to it. No Pedecinus has been taken from a Simian.

Finally man, representing the fourth Primate family,

Hominide, is the host of three notorious Anopluran spe-

cies, two of which are species of Pediculus and the third

the only species so far known of another genus, Pthirius.

Neumann is inclined to see in Pediculus corporis only a

variety of Pediculus capitis. All of these parasites are

found on man in all parts of the world. Some curious

variations among the parasite individuals are shown,

perhaps the most curious being a plain tendency to a

darker coloration of the individuals occurring on the

bodies of men of the dark-skinned races. In my brief dis-

cussion elsewhere, already referred to, I have noted the

interesting significance of this possession by man and the

anthropoid apes of a common genus of Anopluran para-

sites, while the parasites of the lower monkeys belong to

No. 569] ECTOPARASITES OF MAMMALS 279

a well-distinguished other genus. There is no doubt that

the close physiological fitting of parasites to host makes

their host distribution significant of genetic or ‘‘blood’’

relationship, and this commonness of one type of parasite

to man and the apes, and its limitation to these hosts, and

replacement on the lower monkeys by another parasitic

type, is an added indication of the actual blood-likeness

of the Simians and man, a likeness apparently greater

than that between the Simians and the lower monkeys.

Ill

In the light of the plain statement in part I of this

paper of my belief gained from a study of the distribu-

tion of the bird-infesting Mallophaga, to the effect that

the host distribution of the permanent wingless ecto-

parasites of birds is determined more by the genetic rela-

tionships of these hosts than by geographic relationships

or any ecological condition, and the corollary of this,

which is that the distribution of the parasites may there-

fore often have a valuable significance as to the genetic

relationships of animals whose genealogic affinities are

in process of ascertainment, and in the light of the facts

of distribution for the mammal-infesting Mallophaga and

Anoplura as just set out in part II of this paper, I

hardly need to do more, in conclusion, than to point out

that the distribution conditions exhibited by the mammal

parasites, even in the face of the meager knowledge that

we yet have of the mammal-infesting forms, clearly, on

the whole, confirm this thesis. In fact, considering how

few mammal-infesting parasite species we yet know, it is

surprising how repeatedly the commonness of parasite

species to two or more related, although geographically

well separated, host species, is illustrated. All through

the order from Marsupials to Quadrumana this condition

is again and again exemplified. I am then, naturally,

made more certain of the essential truth of the thesis, and

can the more strongly recommend the attention of sys-

tematic zoologists to that practical application of it,

which I have stated in the form of a corollary.

REGENERATION, VARIATION AND CORRELA-

TION IN THYONE

PROFESSOR JOHN W. SCOTT

UNIVERSITY OF WYOMING

Ir is well known that many Echinoderms possess a re-

markable power of regeneration, and the results given

here show some interesting phases of this process in

Thyone briareus (Leseur). The problem was suggested

a few years ago in connection with class work in the

Marine Biological Laboratory at Woods Hole, Massachu-

setts. There itis a common practise for students who are

taking the invertebrate course to keep aquaria in which are

placed specimens brought in from various collecting trips

in the vicinity. Students are encouraged to study the

behavior of these animals, but their enthusiasm for col-

lecting frequently causes them to overcrowd their aquaria,

with disastrous results. After collecting Thyone, espe-

cially if they are kept in stagnant water, the student is

frequently amazed to find one or more of his specimens

that have undergone evisceration. In this process the

animal not only loses the principal feeding organs, the

tentacles, and the entire digestive system, consisting of

the esophagus, stomach and intestine; but it also throws

out a whole series of organs surrounding the esophagus

including the circlet of calcareous plates, the nerve ring

forming the central nervous system, the portion of the

water-vascular system known as the ring canal with its

attached stone canal and Polian vesicles, and the muscles

which serve as retractors for the set of organs surround-

ing and attached to the esophagus. We shall refer to

these muscles as retractors of the esophagus.

The remainder of the animal after evisceration con-

sists, principally, of the dermo-muscular integument, the

280

No. 569] REGENERATION 281

cloaca with its’ attached respiratory trees, the single

gonad, the radial canals of the water-vascular system and

the major portion of the dorsal mesentery by which the

intestine was suspended. Since this part of the animal

continues to give reactions, the student invariably raises

the question, ‘‘Can Thyone regenerate the lost parts???’

This question was the starting point of the following in-

vestigation. The work had not proceeded far when it

was discovered that important individual differences

occurred, and the question became, ‘‘To what extent, or

how completely, may these individual variations be re-

produced in the process of regeneration?’’ Curiously

enough, the most important differences between individual

Thyone involve structures which help to form the radia!

symmetry of the animal. Consequently the problem has

a bearing on the phylogeny as well as the ontogeny of

Thyone.

In general, the results show that regeneration of all lost

organs may occur and that there is a decided tendency to

even reproduce individual variations. It was found that

the Polian vesicles varied greatly in number, size and

location. The retractor muscles in a single radius were

single or multiple, and for each individual this variation

was closely correlated with a corresponding variation in

the number of Polian vesicles. Whether one or more

Polian vesicles are present, there is a strong tendency for

these to occur on the left side of the animal, a fact which

undoubtedly has a phylogenetic significance. A more

complete statement and a discussion of these results will

be given in the following pages.

GENERAL STRUCTURE OF THYONE

Thyone is functionally a bilateral animal. It has ante-

rior and posterior ends, dorsal and ventral surfaces, and

consequently right and left sides. The external opening

of the genital duct is located near the anterior end in the

mid-dorsal region. The structure and arrangement of

the tentacles is alike on both sides of the animal. Even

282 THE AMERICAN NATURALIST [Vou. XLVIII

the feeding reactions, as Pearse has pointed out, indicate

a bilateral type. The single genital gland is median in

position; the genital duct and the stone canal are in the

median dorsal mesentery; a part of the intestine and the

stomach are supported by the same structure. The respi-

ratory apparatus is also a bilateral structure, one branch

arising from each side of the cloaca.

Fig. 1. F >

A DIAGRAMMATIC DRAWING FROM A DISSECTION MADE BY TAKING A

Bopy WALL OF E

Fig. 1.

LONGITUDINAL CUT IN THE

TRAL LIN S he arrangement of the chief organs concerned in evisceration

and subsequent regeneration. B, w., body wall; cl., cloaca; c. p., calcareous

estine ;

LITTLE T LE THE MID-VEN-

canal; r, m., retractor muscles; r. t., base of respiratory tree; s., stomach; t.,

tentacles; m. d., mid-dorsal; 1. d., left dorsal; l. v., left ventral; r. d., right

dorsal, and r. v., right ventral, interradial spaces.

Fic. A DIAGRAM To SHOW THE RELATION OF RADIAL TO BILATERAL SYM-

METRY. The esophagus (e) is shown in cross-section, cut just anterior to the

stomach, and the view looks toward the anterior end. M., madreporite; r. ¢.,

ring canal, Other letters as in Fig. 1.

Notwithstanding this general tendency toward bilateral

symmetry, the most conspicuous differences between indi-

viduals involve structures of the radial type. Fig. 1 is a

diagrammatic drawing of a dissection to show the general

No. 569] REGENERATION 283

arrangement of some of the more important structures

studied in this experiment. The dissection was made by

making a longitudinal cut in the body wall a little to the

left of the mid-ventral line, and then pulling the flaps

apart and pinning the animal down on its dorsal surface.

The Polian vesicle is shown attached to the ring canal in

the position where it is usually found when only one is

present, that is in the left dorsal interradial space. It

will be noticed that the retractor muscles are simply

branches of the longitudinal muscles, and hence are radial

in position. At the time of evisceration the body wall

breaks a short distance posterior to the tentacles, the re-

tractor muscles separate at the point where they join the

longitudinal muscles and the intestine breaks off just in

front of the cloaca.

A better understanding of the radial type of structure

will be gained by a reference to Fig. 2. This figure is a

diagram to show the relation of the radial to the bilateral

symmetry. The dorsal side of the animal is represented

toward the top of the page, the esophagus appears in

cross-section, cut just anterior to the stomach, and there-

fore one is looking forward to the other organs shown.

The retractor muscles, showing the position of the radii,

are much contracted and thickened, a condition in which

they are usually found after evisceration. The stone

canal ending in the small madreporite is located in the

mid-dorsal interradial space. Passing around in a clock-

wise direction, the other interradial spaces are designated

as right dorsal, right ventral, left ventral and left dorsal.

Polian vesicles may be found in any of the interradii ex-

cept the mid-dorsal space which always bears the stone

canal. Although only one Polian vesicle is represented in

this figure, the mid-ventral retractor muscle is shown

double, a split condition which is characteristic when two

or more Polian vesicles are present. This description

will be sufficient to show the general relation between the

radial and the bilateral symmetry.

284 THE AMERICAN NATURALIST [Vow. XLVIII

E\VISCERATION

Only one method of producing evisceration was used.

By placing a number of Thyone in a small aquarium of

stagnant sea water, the supply of oxygen is soon ex-

hausted. The animals become greatly distended, they

crawl up on the sides of the aquarium when possible, and

extend the siphon toward and frequently above the sur-

face of the water. All of their behavior, including the

pumping of the siphon, indicates that respiration is in-

adequate. In the course of a day or two the water be-

comes very foul; soon some of the Thyone will eviscerate,

and a considerable percentage will do so as conditions

grow more unfavorable. Many, however, resist the un-

favorable surroundings and will not eviscerate though

kept for several days in foul water. But if the aquarium

is now placed where it will have a continuous stream of

water and air bubbles passing through it, the behavior of

the animals is somewhat different. They then tend to

contract to a minimal size, and sometimes assume a

volume not more than one fifth to one seventh of their

maximum distention. The respiratory movements are

practically discontinued; the animal seeks a position as

close as possible to the side and bottom of the aquarium.

Contraction does not always take place immediately. To

my surprise, after several hours I found Thyone which

had resisted the previous unfavorable conditions now dis-

charging their viscera. After remaining two or three

days in the running water, and the animals had appar-

ently become adjusted to this condition, I again set the

aquarium to one side partly filled with water. Then, by

repeating the conditions of the first experiment, as the

water became foul several more of the holothurians ap-

parently found life too strenuous to further retain their

internal organs. When the remainder of this lot of

Thyone was returned to running water, and again to

stagnant water, a few additional individuals underwent

self-mutilation. Out of a total of sixty-one specimens

used in this lot forty of them eviscerated. That is, autot-

No. 569] REGENERATION 285

omy occurred in at least sixty-five per cent. of Thyone,

under the conditions described. Probably one reason why

this process did not occur in a still larger number is that

some animals occupied more favorable positions in the

aquarium. A discussion of the cause of evisceration will

be given later.

When evisceration occurs it is sometimes hard to see

just how the process takes place. Pearse (’09) ascribes

the process to a ‘‘structural accident’’; that is, it is due

to a powerful contraction of the circular muscles at a time

when the calcareous ring is well forward. ‘‘But if the

tentacles are extended,” he says, ‘‘and the calcareous

ring is pushed forward a break may occur at b” (a point

in his Fig. 2 where the body wall joins the calcareous

ring) ‘‘as a result of the strong contraction of the circu-

lar muscles at that point, and the visceral organs are

forced out. . . . Whether this autotomy takes place or

not depends upon the breaking of the inner branch of the

longitudinal muscle bands, whose normal function is to

retract the caleareous ring. When the strain brought

about by the contraction of the circular muscles becomes

too great these inner bands are torn asunder, usually at

the point x’’ (inner end of the retractors of the calcareous

ring). While it is true that muscular contraction and

consequent pressure undoubtedly plays a prominent part

in the process, close observation has convinced me that

this is not the only factor causing evisceration. Upon

Several occasions I have watched carefully the breaking

of the body wall near its attachment to the calcareous

ring, and while there are times when the pressure appears

to be strong, especially when the animal is being irritated

mechanically, there are other times when the skin appears

to ‘‘melt away’’ or separate with very little or no pres-

sure present. Indeed, after the skin once breaks at one

side and the viscera escape through the opening, the pres-

sure is relieved. But one may observe that the skin con-

tinues to break until the calcareous ring is entirely sepa-

rated. This, of course, would not happen if the process

286 THE AMERICAN NATURALIST [Vouw. XLVIII

depended entirely upon an accidental structural defect.

Another thing noticed is of interest in this connection.

When splitting open the body wall of an animal that was

eviscerating, and thus relieving any internal pressure that

might be due to contraction of the circular muscles, some

of the retractors were seen still attached to the longitu-

dinal muscles. Under these conditions it would not be

possible for the retractors to exert any pull against the

pressure produced by the circular muscles, yet the re-

tractors were observed to constrict off or break away

from the longitudinal muscles by what appeared to be

purely a local disturbance. It is hard to see how this

could happen, or how the skin continues to separate

around the calcareous ring after the first break is made,

if the process of evisceration depends solely upon the

breaking of retractors and internal pressure. Indeed,

the view that local changes take place in the tissues is

supported by other facts. Leptosynapta, if left in stag-

nant water or under other favorable conditions, under-

goes repeated autotomous fission as the result of local

constrictions, and Pearse states that autotomy depends

upon the presence of the anterior portion of the body,

and presumably upon the presence of the circumoral

nerve ring. However, he found in Thyone that highly

irritating substances like acetic acid and clove oil did not

produce ejection of the viscera.

Nor were drugs like codene and atropine, which cause violent peri-

staltie waves of contraction to pass over the body, any more potent in in-

ducing autotomy. The same may be said of sodium chloride, atropine

and clove oil, although the injection of any of these substances was

often followed by a waving of the oral tentacles to perform feeding

movements, thus bringing about favorable anatomical relations for au-

totomy.

These results would indicate that the nervous system

is not primarily involved. Certainly the ejection of vis-

cera may occur in Thyone without cord visible external

stimulus.

The parts eviscerated in Thyone have already been

No. 569] REGENERATION 287

mentioned. However, sometimes evisceration is incom-

plete, as the following examples will show. On the morn-

ing of August 4, a Thyone, which we shall later speak of

as individual H, was found eviscerating in an over-

crowded aquarium jar. While the process usually re-

quires only a few seconds, or at most a few minutes,

the intestine in this case was not completely thrown out

until two or three hours later. This animal lived until

killed at the end of twenty-one days. In the afternoon

of the same day on which individual H eviscerated, an-

other Thyone was found with the process only partially

complete. Five hours later the intestine was still re-

tained, and scissors were used to cut it off at its anterior

end near the stomach. Though this Thyone received

equally good care it died at the end of two days without

further evisceration. A third specimen was found in-

completely eviscerated on the above date, but it was

allowed to stand until the next morning; at this time the

injured end was open, the intestine was still within the

body cavity and a part of one of the branchial trees was

protruding. The intestine was pulled out and broken

off, after which the branchial tree was retracted and the

injured end partially closed. This animal also died at

the end of two days. A fourth Thyone was seized and

by squeezing was forcibly caused to throw off the usual

parts except the following: a part of the stomach, most

of the intestines, and some of the retractor muscles

which had broken off near their esophageal end. The

next morning it had expelled the remainder of the

stomach and intestine, two complete retractor muscles,

and some débris which had escaped from the intestine

into the body cavity. The anterior end of the part re-

maining appeared ragged and imperfectly closed. It

died on the third day. It is probable that the two re-

tractor muscles last expelled were broken off at their

posterior ends by local constriction, not when the body

was under pressure. A fifth animal, which we shall

designate as individual M, was found partly eviscerated

288 THE AMERICAN NATURALIST [Vou. XLVII

late on the afternoon of August 6. The next day it still

retained the stomach and intestine and at noon the diges-

tive tube was clipped off with scissors in the region of

the esophagus. Nothing peculiar was noted in its behavior

until four days later, August 11, when it discharged the

remainder of the digestive tube. It lived and was killed

at the end of eighteen days. These results are typical.

The animal dies unless it is itself able to eliminate all

organs concerned in the process of evisceration, and

therefore regeneration does not occur unless all these

organs are eliminated.

The eviscerated animals show comparatively a low

degree of mortality. In an attempt to raise twenty-five

mutilated Thyone seven died; three of these were un-

able to complete the process of evisceration as described

above, and two more, since they lived for fourteen days,

probably owe their death to other causes. The sixth

specimen to die lived three days and had been slow in

eviscerating. The seventh did not receive the best of

care and died after three days. So considering the

amount of injury the mortality is extremely small where

proper care is taken and evisceration is complete.

It will not be inopportune to describe the subsequent

behavior of the different parts after evisceration. The

parts expelled lie on the bottom in a more or less inactive

condition until they die, which happens usually in the

course of a few hours. At first the tentacles frequently

expand and contract. They are highly sensitive, as one

would expect, and if touched withdraw quickly into the

esophagus and at the same time the retractor muscles

will undergo strong contraction. By supporting these

parts near the surface of the water, so as to insure plenty

of oxygen, an attempt was made to keep them alive. In

some cases the parts remained alive for two or three

days, so this experiment appeared to be partially success-

ful. Death is probably due to the direct exposure of

tissues to the sea water and to the attacks of minute

organisms. The dermo-muscular portion of Thyone is

No. 569] REGENERATION 289

much less sensitive than the expelled portion, just after

evisceration. This is due to lack of a central nervous

system.

Benavior Durinc REGENERATION

After evisceration each specimen was placed in a sepa-

rate jar of fresh sea water. The injured end of the body

turns in and closes up tightly, and the entire body is

somewhat smaller than before evisceration. Respira-

tion is slower and not so vigorous. If the water is stag-

nant, within a few hours the animal usually climbs up

on the side of the aquarium by means of its tube feet.

This part of the animal therefore is capable of respond-

ing to a lack of oxygen, and the reaction is independent

of the central nervous system.

The observations upon the following individual, re-

ferred to in my notes as Thyone A, will serve to illus-

trate the general behavior during regeneration:

July 14, A.m.—Animal eviscerated itself in the usual way. In the

afternoon it climbed up on the side of the jar and clung there evidently

for the purpose of respiration.

July 15-16.—Acts as on the afternoon of the fourteenth. Keeps

closed and well contracted at the injured end. Entire body somewhat

smaller than before evisceration, due in part to organs lost. Respira-

tion slower and not so vigorous as norm

July 17.—In the afternoon, after watar was changed, Thyone took

up position on the sand against the side of the jar farthest away from

the source of light.

July 18.—The next morning it was half buried in the sand in same

Tr, with a few pieces of débris pulled over it. Remained so all

ay.

July 23.—For some two days it has been slowly burrowing down until

only the two protruding ends of the body can be seen. When a piece

of débris that was being held over a part of the anterior end was

touched, this end retracted below the surface and the posterior end

withdrew until it could scarcely be seen. Later the posterior end re-

tracted when the shadow of my hand passed over it, the hand being held

about one foot away. The uninjured animal is even more sensitive to

shadow. The respiratory movements are growing stronger.

July 28.—For the past two or three days the Thyone has been slowly

moving through the sand in a posterior direction without uncovering

itself,

290 THE AMERICAN NATURALIST [Vou. XLVIII

August 2.—It is now oriented with respect to the direction of the

light and has reached probably the darkest portion of the jar.

August 7.—Has advanced still farther. Came about half way out of

the sand to do this.

August 8.—Reacts quickly to shadows by iar inlet and to jar-

ring tha ak Evidently is recovering its normal behav

August 10.—Has again come up about half way out a the sand.

Reacts quickly to shadows as before.

ugust 11.—Came entirely out of the sand. Spent the day on the

sand or on the side of the jar. Appeared restless.

August 12, 4 p.m.—Has been clinging to the side of the jar and mov-

ing about more or less all day. Respiratory movements are strong and

apparently normal. Has just now expanded the anterior end suffi-

ciently for me to see the new growth of tissue formed around a penta-

gonal opening. Fifteen minutes later it was observed to extend a set of

minute tentacles and go through feeding movements. The tentacles ap-

peared to be slightly more than three eighths of an inch in length. Its

behavior continued apparently normal until it was killed twelve days

later.

The actions of other Thyone were studied under the

same conditions, and we shall now give a general sum-

mary of their behavior during regeneration. The earli-

est reactions after evisceration take the form of contrac-

tions resulting in the closure of the wound, and move-

ments in response to lack of oxygen. If the oxygen

supply is sufficient Thyone will draw itself closely into

the angle between the side and bottom of the aquarium,

or if the supply is deficient, it clings close to the side of

the jar near the surface. In from three to seven days

an instinct to burrow usually asserts itself. There is a

tendency for the body to contract very noticeably at this

time, and the whole organism becomes rather inactive.

This condition is probably necessary for the formation

of new tissue. Pearse makes the statement that in bur-

rowing the normal Thyone will cover itself in from two

to four hours. My observations on the mutilated ani-

mals indicate that they require from twelve to twenty-

four hours, in one case forty-eight hours, to complete the

reaction. The process frequently stops for some hours

and occasionally is never completed. In the Thyone de-

No. 569] REGENERATION 291

scribed above the animal did not begin to orient itself

with respect to the source of light until about the twelfth

day, but in another case the response took place on the

second day, which shows that this reaction does not de-

pend upon the central nervous system. It should be

stated that normal Thyone similarly placed were used as

controls. Thyone A was quite sensitive to shadows and

to touch on the ninth day, but it reacted more quickly on

the twenty-fourth day both to shadows and to mechanical

disturbances. Whether this was due to the regeneration

of a new central nervous system, or to a more highly

developed specialization of function in the old tissue, I

am unable to say. It is quite possible that both factors

were involved. Respiration is undoubtedly correlated

with the activity of the animal, and feeding movements

do not occur until the regeneration of all organs is well

established, at about twenty-seven or twenty-eight days.

The internal changes that take place during regenera-

tion were studied in animals that were killed at different

stages in the process. Thyone N was killed nine days

after self mutilation. At the injured end there was a

very small plug of tissue representing the newly formed

esophagus; a thread-like continuation of this tissue, the

beginning of a new stomach-intestine, was also seen in

the mesentery. The calcareous ring and the ring canal

were not clearly defined. Another Thyone was killed at

about the same age after evisceration; India ink was in-

jected into the cloaca and into the opening at the ante-

rior end in an attempt to demonstrate a cavity in the

newly formed thread-like, stomach-intestine. The re-

sults were negative and the esophagus was found to be

tightly closed. However, the interesting observation was

made that the anterior end of each of the longitudinal

muscles had split off a very slender branch to form a

new retractor muscle (see Fig. 3). These newly formed

retractor muscles were not more than one fourth inch in

length; their anterior ends were attached in a normal

Position around the esophagus, but their posterior ends

292 THE AMERICAN NATURALIST [Vou. XLVIII

were attached only a short way back, much in front of

the position of attachment of the full-sized retractors.

In another animal killed when a day or so older, the same

conditions held with reference to esophagus, stomach

and intestine. At least three of the radial canals belong-

ing to the water vascular system had branched and con-

nected at their anterior ends in such a manner as to

form a part of a new ring canal (cf. Fig. 4). I was un-

able to find the rest of the ring-canal and perhaps it was

not yet mea

Fig. 3. Fic.

Fic, 3. DIAGRAMMATIC DRAWING TO SHOW THAT IN Ro oN. THE RE-

hera MUSCLES (r. m.) ARISE BY SPLITTING OFF FROM THE LONGITUDINAL

MUSCLE .m.). Dissected a little to the right of the mid-ventral line; d., dor-

sal squall ae suspending the intestine (i.); in., integument; e., region of

Fie. O SHOW THE DEVELOPMENT OF THE PATUN CANAL IN A THYONE

Kobi aa OR oe DAYS AFTER EVISCERATION, radial canal; p., pentagonal

The terior ends of the radial canals ak neisti and these

vegn pe erR ose to form the canal which later assumes a circular shape

around the esophagus.

Thyone F, which was killed twelve days after eviscera-

tion, showed minute calcareous plates which formed a

very small esophageal ring not more than one millimeter

in diameter. The esophagus continued posteriorly in the

form of a small tube, the stomach-intestine, which was

suspended in the dorsal mesentery. This new digestive

tube was about 0.5 millimeter in diameter and contained

small, colored, movable particles that could be seen with

the unaided eye. The ring canal was completely formed.

Another specimen, Thyone O, died at the end of four-

No. 569] REGENERATION 293

teen days and was in bad condition when examined. The

stomach had begun to expand and retractor muscles were

present. Probably owing to the condition of the speci-

men, no calcareous ring, ring canal, or Polian vesicle

could be found. Another individual killed at about fif-

teen days showed the stomach slightly enlarged, and the

intestine, retractor muscles, calcareous ring, tentacular

canals, and ring canal well formed. Two small Polian

vescicles each about one millimeter in length were pres-

ent. The position of the new intestine was described in

my notes as follows:

From the stomach the intestine follows the ventral edge of the dorsal

mesentery, lying ventral to the gonaduct. At the gonad it turned ven-

trally with the mesentery and then forward for about one half inch to

the left interradial space; here it turns rather abruptly backward, con-

tinuing in the mesentery below the left branchial tree to the anterior

ventral part of the cloaca.

At a little later stage in another specimen the intestine

passed from the left ventral interradial to the right ven-

tral interradial space; then posteriorly and again to the

left, following the ventral radial mesentery to the ante-

rior ventral side of the cloaca.

We see from the preceding description that all impor-

tant organs have been reproduced in form though not in

size, before the end of the fifteenth day. The first madre-

porite with its tiny stone canal was found some eighteen

days after mutilation. Twenty-one days after eviscera-

tion in one specimen the caleareous ring was about three

millimeters in diameter and the ampulle at the bases of

the tentacles were well developed. Within a week after

this time the regenerating animal begins active feeding.

Thyone A, killed at 41 days, was practically a normal

animal both in behavior and appearance, except for the

fact that the regenerated organs had not yet reached full

size. The stomach was about one third normal size, but

the Polian vescicles were better developed. The intes-

tine contained a small amount of food material and was

nine or ten inches in length; most of this growth had

294 THE AMERICAN NATURALIST (Vor. XLVIII

taken place posterior to the gonad. It was held in posi-

tion as previously described and had several additional

coils.

[INDIVIDUAL VARIATIONS

To all outward appearances any two Thyone are as

much alike as two peas. It was not until the internal

organs were studied that important differences were ob-

Fie. 5. see TO SHOW VARIATION IN POSITION AND SIZE OF THE POLIAN

VESICLES, P. v., Polian vesicles; m., madreporite; r. c., ring canal; a-d, with

one Polian v Manila a e-g, with two; h-k, with three, 1l., with four; c., d., f., 9. E-

with additional rudiments of these vesicles; j., with a branched vesicle.

served. While there are numerous minor differences,

the most conspicuous variations are found in the num-

ber, size and location of the Polian vesicles (cf. Fig. 5),

and in the number and arrangement of the retractor

muscles. On account of the radial structure of Thyone

not more than four Polian vesicles are present, since

No. 569] REGENERATION 295

a homologous structure, the madreporite and its stone

canal, occupies the dorsal interradial space. The num-

ber of vesicles varies in fact from one to four. By a

reference to Table I, it will be seen that out of 77 indi-

vidually examined, 41 had one, 20 had two, 14 had three,

TABLE I

To SHOW THE NUMBER OF POLIAN VESICLES PRESENT IN A GIVEN NUMBER

OF THYONE. ALSO TO SHOW THEIR LOCATION IN THE

INTERRADIAL SPACES, WITH REFERENCE TO THE

BILATERAL SYMMETRY OF THE ANIMAL

Number of Number of Individ- Left | Left | Right Right

Polian Vesicles uals Examined Dorsal | Ventral | Ventral Dorsal

1 41 les a: 0

2 20 Wooa o d 3 1

3 14 14 | 15 | 12 1

4 2 2o] 2o 2 2

TOM ee ee. Ta ee ee R 4

and 2 had four Polian bodies. If one is to test the matter

of regeneration, of course it is important to know

whether the variations or individual peculiarities will

be accurately reproduced. Another striking character-

istic comes out when we note in the same table the loca-

tion of these organs. Of the forty-one individuals which

had a single Polian vesicle, all were on the left side of

the animal, and 38 were in the left dorsal interradial

space. In twenty specimens with two Polian bodies each,

36 were on the left side and only four on the right side

of the body. A similar asymmetrical distribution of

these parts was found when three Polian bodies were

present. In one specimen, however, two vescicles were

found in one space, the left ventral interradius, the only

instance of this kind observed; on account of this dou-

bling, the right side lacked one of the number to which it

was entitled in the table. Where four Polian bodies are

present the arrangement is, of course, symmetrical on

both sides. Still another interesting fact comes out when

we examine the totals in the last line. Out of the 77 indi-

viduals, 71 had a Polian vesicle in the left dorsal inter-

radial space, 39 vesicles were found in the left ventral,

296 THE AMERICAN NATURALIST [Vou. XLVIII

17 in the right ventral, and only 4 in the right dorsal

space. That is, the total number on the left side com-

pared with the total number on the right side bears the

ratio of 110 to 21.. Not only is there this tendency for

the vesicles to be more abundant on the left side of

Thyone, but the totals show that the chances of a given

Thyone having a Polian vesicle in any given interradial

space decreases in a counter-clockwise direction, begin-

ning with the left dorsal interradial position. Coincid-

ing with the number of individuals examined, the maxi-

mum number of chances is found in the mid-dorsal inter-

radius, where the stone canal is always present. That is,

the stone canal with its madreporite is a more funda-

mental and stable structure than each or all of the

vesicles.

The conditions are none the less interesting when we

compare the Polian vesicles with reference to size and

location, as will be seen from the examination of Table

II. The Polian vesicles are here divided arbitrarily

into three groups, designated as large, medium and

small, and their respective locations are shown. In addi-

TABLE II

To SHOW THE POLIAN VESICLES WITH REFERENCE TO SIZE AND LOCATION

Size | Left Dorsal Left Ventral | Right Ventral Right Dorsal Total

| |

Largo. i... | 56 17 0 | 0 73

Medium..... | 17 22 5 | 1 45

Smal.. | 0 0 10 | 3 13

Rudiment... | 2 1 5 | 7 15

Toa... | 75 o a 20 | 11 146

tion some Thyone had the rudiments of other vescicles,

each too small to be considered a distinct pouch. These

are designated in the table as a ‘‘rudiment.’’ It will be

noticed that all of the large, and most of the medium-

sized vesicles are on the left side; that all the small

ones, and most of the rudimentary ones are on the right

side. The table as a whole shows that not only does the

number of Polian vesicles diminish in a counter-clockwise

No. 569] REGENERATION 297

direction, but their size diminishes following the same

law. These facts appear significant and without doubt

are suggestive of ancestral history.

If it is true that the radial symmetry of Echinoderms

is to be ascribed to a fixed stage in their ancestral his-

tory, we are led to suppose that the point of attachment

was on the right side of an originally bilateral animal.

The life history of Pentacrinus, the larval organ of Aste-

roidea, and a great many anatomical and embryological

facts support this view. While it is not within the prov-

ince of this paper to discuss the relative significance of

these matters, the evidence is so overwhelming that the

theory is generally accepted. It is also no doubt true

that some groups of Echinoderms took to a free-living

existence early in their ancestral history, and others re-

mained fixed until comparatively a late period. As proof

we may cite the embryological evidence that Holothurians

develop without any attached stage whatever, that the

Asteroids develop a larval organ and pass through a-

Sessile stage for a brief period in their development,

while the crinoids usually remain permanently fixed

throughout life. At least we can best account on this

theory for the deep-seated and fundamental radial sym-

metry of some forms; the longer the attachment the

more deep-seated would become the type of radial sym-

metry. Now if this theory is correct we can use it to ex-

plain the conditions described above for Thyone. The

ancestors of this form must have broken away from the

fixed stage very early, for we find the radial symmetry

not well established on the right side of the animal as

evidenced by both the position and size of the Polian

vesicles. Out of 118 large and medium-sized Polian

vesicles, 112 were on the left side, while in a total of 28

small or rudimentary Polian bodies, 25 were found on

the right side. The arrangement of these organs in

Thyone adds one more bit of evidence to support the

following statement of Lankester.

298 THE AMERICAN NATURALIST [Vou. XLVIII

It therefore appears that the Holothurian stock branched off from

the Pelmatozoa before complete pentamerous symmetry of the hydro-

coele and associated organs had arisen, before any definite caleynal sys-

tem had developed, while the gonads were still a simple strand opening

to the exterior by a single posterior gonopore.

The muscles used as retractors of the esophagus were

other organs in which there was considerable individual

variation. As a general rule each of the five retractor

muscles consists of a single band that takes its origin from

the longitudinal radial muscle about one third the way back

from the anterior end of the body and is inserted in front

into the wall of the esophageal ring. Such a retractor,

however, is frequently split up into several strands vary-

ing from two to five in number. A reference to Table IIT

TABLE III

To SHOW THE CORRELATION BETWEEN THE NUMBER OF POLIAN baar

AND THE TENDENCY FOR THE RETRACTOR MUSCLES TO DIVID

Number of Polian Vesicles [24 es soj a

| |

Retractor muscles; single.) es a ey et | 39 2 0 0

Retractor muscles, multiple. .................... ee. |17 5 2

Average n aeres retractor muscles, per individual. . .| 5.153 10.263) 12.400, 10-000

Average number retractor muscles, per radius. ..... eet 030) 2.06211 2: 480) 2 2.000

shows that in 76 individuals examined, 41 had retractor

muscles all in single bands, while 35 specimens had these

muscles subdivided or multiple in character. This vari-

ation is especially interesting when considered with

reference to the number of Polian vesicles. For in forty

cases where one Polian body was present thirty-nine bore

the unsplit or single retractor and there was only one

specimen with these muscles showing a multiple number.

In thirty-six cases where two or more Polian vesicles

were present, all but two had the retractor muscles in a

split or divided condition. If we consider each strand

as a separate retractor muscle, we may then obtain the

average number of retractors per individual for any

definite number of Polian vesicles. By a reference to

the fourth horizontal line of Table III, one finds that the

average number in individuals with one Polian vesicle is

No. 569] REGENERATION 299

just slightly in excess of five, the pentameric number,

and the average number when two Polian vesicles are

present is 10.263. This ratio is only partly maintained

when three vesicles are present, for the average number

is then 12.400, and in the two cases with four vesicles the

average was just twice the pentameric number. It is

therefore evident from the facts shown in this table that

with an increase in the number of Polian vesicles there

is associated a strong tendency for the retractor muscles

to take on a split character. If it were not for the fact

that the split character shows considerable variation in

the same individual one might suggest that the tendency

to divide is correlated with the greater functional activ-

ity of the water vascular system as evidenced by the in-

creased number of Polian vesicles and the location of the

longitudinal muscles that lie along and just internal to

the radial canals. About all one can say is that corre-

lated with a more complete radial symmetry with respect

to the Polian vesicles, there is a greater plasticity in the

retractor muscles, causing them to divide longitudinally

into separate muscle bands.

To what extent, or how completely, may these indi-

vidual variations be reproduced in the process of regenera-

tion? An answer was obtained in the following way.

First a close examination was made of all parts eviscer-

ated and a record was kept of all organs showing variable

structures. Special attention was given to Polian vesicles

and to retracter muscles. The mutilated specimens were

then placed in separate aquaria in which the water was

changed frequently to prevent it from becoming stale.

After a considerable interval these animals were killed

and the regenerated organs were compared with the lost

parts. Table IV shows several individuals compared in

this way. The number of retractor muscles found in each

radius is given in the order of the radii taken in a clock-

wise direction. A study of the table indicates that there

is a strong tetndency to reproduce individual peculiarities,

as shown by individuals B, E, G, H, M and O. This does

300 THE AMERICAN NATURALIST [Vou. XLVIII

not always hold true, for individual L reverted toward

the more radial type of symmetry. From these few cases

it would appear that individual peculiarities tend to pre-

dominate over ancestral influences in the process of re-

TABLE IV

To ILLUSTRATE THE RELATION BETWEEN REGENERATION AND ORIGINAL SYM-

METRY IN THYONE

Tnaividuai Original Symmetry Regenerated Symmetry

Uses Polian Vesicles Retractor Muscles | Polian Vesicles Retractor Muscles

B 2-—2-—2-—2-2 2 3-3-—2-3-3

E 2+ 3-3-2-2-2 2 2-83-2-2-2

G 1-1-1-1-1 2 1-1-1

A 2 1—2-—2-—2-1 2+ 2-2-2-—2—

L 1 1-1-1-1-1 2 2-3-2-2-2

M 1 1-1-1-1-1 7 1—2-1-1-2

O 3+ | 2-—2-2-2-2 ? 2-2-1-2-2

WwW Bh AA 2 2-2—2—3—4

> OE Aree E ad SE 2 2-2—2—2-2

Ee ea a a E a eo 2 2—2—1-—2-2

generation. Specimens W, X, Y, are included in this

table to show further the correlation between Polian ves-

icles and retractor muscles.

Discussion AND SUMMARY

There remains to be discussed the general bearing of

the foregoing experiments. First, the difference in the

number of Polian vesicles in different Thyone is partly

compensated by a variation in size, the fewer the number

the larger their size, though this ratio would not be an

exact one. In other words the total volume of the Polian

vesicles in any given specimen bears a general relation to

the size and functional activity of the animal. Notwith-

standing this functional relationship since the actual

number varies so widely it would be interesting to com-

pare the number found in other species of holothuria with

the conditions in Thyone. The data secured on this ques-

tion were meager and not very definite. For example,

Packard in one of the older text-books says in speaking

of Thyone,

No. 569] REGENERATION 301

There are three Polian vesicles, one fusiform and an inch in length,

the two others slenderer.

Clark (’02) gives the number for Thyone briareus

(Leseur) as usually one or two; for T. scabra (Verrill)

as usually single, and for T. wnisemita (Stimpson) as

one. He also mentions six other holothurians found in

the Woods Hole region and all have a single Polian ves-

icle except Cucumaria frondosa (Gunnerus), which usu-

ally has one. He says nothing of the position in which

these vesicles are found. In another paper (’01) Clark

mentions a large holothurian about 40-45 centimeters in

length (Holothuria mexicana Ludwig) in which there is

a great diversity in the number of tentacles and Polian

vesicles. The tentacles vary from 18 to 21, while the

Polian vesicles vary from 1 to 9. The number of speci-

mens examined, sixteen, was hardly sufficient to obtain an

adequate comparison; two had 1 vesicle each, two had 2,

five had 3, three had 7, one had 8, and one had 9. It is

probable that if one were to examine a large number of

individuals of each species, with reference to the number

and location of the vesicles, he would obtain further inter-

esting results. Lang (’96) cites a number of groups of

holothurians in which only one vesicle has been observed;

but states that there are a number of species in other

groups that have occasionally or usually more than one.

Where accessory vesicles oceur they vary greatly in number, and ap-

pear to have very slight, if any, systematic significance. Where only one

Polian vesicle occurs it lies in the left ventral interradius, very seldom

in the left dorsal interradius. Where two or more vesicles occur, they

are also mostly formed in the ventral region of the cireular canal.

Since Lang describes Cucumaria as the type specimen,

in which the Polian vesicle is said to be in the left ventral

region, it is possible that his generalizations were based

principally on this form. At any rate, the conditions in

Thyone seem to give a more definite significance to the

number and location of the Polian vesicles.

Various explanations of autotomy and evisceration have

been suggested, many of them having a teleological char-

acter. The view that the holothurian offers up the better

302 THE AMERICAN NATURALIST [Vow. XLVIII

part of itself to appease the hunger of its enemy lacks

confirmation, since the viscera are distasteful to fishes

and to some other animals. It may be that the autotom-

ous elimination of the Cuvierian organs serves a defen-

sive purpose, as pointed out by Ludwig and Minchin, and

Minchin suggests that the viscera may also be lost in this

process and thus incidentally be associated with a pro-

tective response. In the case of Thyone, however, evis-

ceration can hardly be considered defensive, and certainly

it is not a process of self-division for only one part pro-

duces a new individual. Clark (’99) in discussing self-

mutilation in the synaptas states the matter clearly in

the following terms:

I agree entirely with Cuenot (’91) in believing that autotomy is not

normal or defensive but is due entirely to pathological conditions. I

never saw a case of it in synaptas supplied with plenty of sand and an

abundance of sea water.

Lang (’96) points out one of these pathological condi-

tions, and recounts the fact that

A Stichopus was observed to come entirely out of its skin, i. e., the

whole integument dissolved into slime, so that only the dermo-muscular

tube enclosing the viscera remained.

In the present paper I have mentioned that Thyone at

times appears to undergo a similar softening of the

tissues in the region where the break occurs, and Pearse

(’09) showed that autotomy is due, at least in part, to a

structural arrangement which he considers is accidental

in character. My observations further show that local

constrictions undoubtedly have an important part in sepa-

rating the retractors from the radial longitudinal muscles.

All of these factors are pathological and are due to exter-

nal or internal stimuli. The external (extra-cellular)

stimuli, mechanical and chemical, as tried by Pearse

(’08), appear to be less effective in producing autotomy

than the purely internal (intracellular) stimuli such as

lack of oxygen and its associated phenomena. The chem-

ical (strychnine) that produced the largest percentage

of evisceration in Pearse’s experiments, probably affected

respiration, since it greatly increased the activity of the

No. 569] REGENERATION 303

animal; therefore the need of oxygen would be propor-

tionately greater than the supply, and the Thyone ren-

dered more susceptible to evisceration. Now while autot-

omy undoubtedly enables the animal to maintain its exist-

ence for a considerable period on a smaller supply of

oxygen, the times when this would become necessary in

nature are probably rare, and it would be futile to specu-

late upon what evolution yet has in store for the process.

According to Lang, the retractor muscles of the oral

region have been derived by the splitting up of the ori-

ginally simple longitudinal muscles, and this specializa-

tion became more marked as the oral tentacles became

more highly developed and required increasing protec-

tion. Species are to be found in the Dendrochirote in

which the separation and branching off of retractors from

the longitudinal muscles has not yet been perfected. In

regeneration the retractor muscles of Thyone are derived

in the same way, i. e., by splitting off from the longitu-

dinal muscles, and such progress is made that they are

fairly well developed by the time the tentacles take up the

function of feeding. The increasing sensitiveness and the

later activity of the regenerating animal are presumably

associated with the development of a new nervous system.

If we may regard the bilateral echinoderm larva as

representing an early phylogenetic stage rather than a

larval adaptation to a free-swimming existence, we will

now discuss the symmetry of Thyone. As stated above, it

is generally agreed that the radial arrangement of parts

of the echinoderm body is due to a fixed stage in its

ancestral history. Some holothurians and spantangoids,

show in their ontogeny first a free stage, second a radial

stage, and finally a bilateral adult. During the develop-

ment of asteroids that have a fixed embryonic stage, the

early bilateral symmetry is soon disarranged by the

development of organs on the left side of the animal.

For example, the left hydrocæle takes the form of an un-

closed water-vascular rosette which grows around the

esophagus to form the ring canal and its appendages, and

its connection with the dorsal pore gives rise to the stone

304 THE AMERICAN NATURALIST [Vou. XLVIII

canal. Excepting the echinoids and crinoids in which

there is either no distinct Polian vesicle or else a simple

glandular structure, those echinoderms that have retained

the most distinctive type of radial structures have also

as a rule, retained the most symmetrical arrangement of

the Polian vesicles. Presumably these forms, the aster-

oids and ophiuroids, have quite recently abandoned the

fixed stage, and each individual usually has four Polian

vesicles and a stone canal, one in each interradius.

Among most of the holothurians a secondary bilateral

symmetry has become superimposed over the radial type,

and it is reasonable to suppose that there was a time in

the ancestral history of Thyone when the Polian vesicles

were symmetrically and radially disposed, or else the

animal quit its fixed habits before the radial symmetry

of the vesicles was thoroughly established. In the one

case we would have a regression, a sort of backward

retracing of the steps of evolution, or, which seems more

probable, the ancestors of Thyone began a free-living

existence before the radial arrangement of the Polian

vesicles had become complete. Also the fact that the

embryology of the holothurian egg is probably much

compressed and shows no trace of a fixed stage indicates

that the corresponding ancestral stage was compara-

tively short, or, very remote. Since the modern habits of

Thyone are bilateral, and since it is altogether improb-

able that such habits would produce the present arrange-

ment of Polian vesicles, the position of these organs must

be due to ancestral influence.

Now the Polian vesicles are capable of contracting and

expanding and their function when they are well devel-

oped is to act as accessory reservoirs of the water-vas-

cular fluid. Muscle and connective tissue in the wall of

the vesicle furnish the means to do this work. Of course,

if the ampulle are well developed there is little or no

need of Polian vesicles, as is the case in Asterias. But,

though the size and number of these vesicles is function-

ally correlated with the general development of the

water-vascular system, especially of the oral tentacles,

No. 569] REGENERATION 305

and hence shows great variability in the different species

of holothurians, this does not in any way explain the

great excess of these vesicles on the left side of Thyone

briareus. In regeneration, probably through the influence

of functional correlation, there is a tendency for the old

tissue to reproduce the exact number and arrangement

of the lost vesicles, but it may reproduce a somewhat

more radial (ancestral) arrangement.

Enough has been given in this paper to show the need

of a more extensive and intensive reexamination of the

Polian vesicles. This would give a better idea of their

morphological and functional significance. The follow-

ing summary and conclusions are based on the work

described:

1. Evisceration in Thyone includes the following or- `

gans: Esophagus, stomach, intestine, calcareous ring,

nerve ring, tentacles, ring canal, Polian vesicles, stone

canal with madreporite, and the retractor muscles of the

esophagus.

2. The method used to produce evisceration was to

allow Thyone to stand in stagnant water until it became

foul. This was followed by treatment with running water

containing much oxygen. Alternating these processes

produced as high as 65 per cent. of self-mutilated indi-

viduals.

3. The structural accident theory of Pearse is inade-

quate to explain all of the conditions arising in the proc-

ess of autotomy. At times the skin appears to dissolve

away with little or no pressure present, and retractors

frequently break off by local constrictions instead of by

longitudinal pull.

4. The parts eviscerated are at first highly irritable,

and may be kept alive for some time. The part remain-

ing is less responsive, but reacts to touch, to lack of

oxygen, and probably to other stimuli.

5. Regeneration of all lost organs may occur, but it

takes place only when all parts concerned in evisceration

are completely expelled. Otherwise the animal dies.

306 THE AMERICAN NATURALIST [VoL. XLVIII

6. During the process of regeneration the behavior

gradually becomes more responsive and finally is like the

normal individual. This appears to be correlated with

the growth of a new nervous system.

7. Thyone is functionally a bilateral animal, but the

most conspicuous individual differences involve struc-

tures that have.a radial arrangement.

8. The Polian vesicles vary greatly in number, size

and location. There is a strong tendency for these to

occur on the left side, and this arrangement is undoubt-

edly due to ancestral conditions, for the present bilateral

habits of Thyone could probably have no influence in

producing this asymmetry.

9. The retractor muscles in a single radius consist of

single or multiple strands, and this variation is closely

correlated with a similar variation in the number of

Polian vesicles. No explanation is forthcoming for this

peculiar plasticity of the retractor muscles, but the sug-

gestion is made that it may be functionally correlated

with the development of the water-vascular system.

10. It was found from the study of a number of speci-

mens that individual peculiarities of structure tend to be

reproduced in the process of regeneration. In this proc-

ess it would appear that individual variations tend to

predominate over generalized ancestral influence.

11. Autotomy enables Thyone to survive for a consid-

erable period on a smaller than normal supply of oxygen.

Nevertheless, the conditions which give rise to self-muti-

lation are seemingly in all cases pathological.

12. The conditions in Thyone afford some evidence for

believing that when this animal abandoned the fixed stage

the Polian vesicles conformed more or less to the radial

type. This is opposed to the statement of Lang that in

all cases where a multiple number is now present ‘‘there

was originally only one vesicle.” It is believed that the

present arrangement of Polian vesicles in Thyone can be

best accounted for on the theory of phylogenetic influ-

ence. That, in general, those vesicles have retained their

most complete radial arrangement in those species of

No. 569] REGENERATION 307

echinoderms which have maintained to a high: degree

the functional activity of the water-vascular system.

REFERENCES

Bather, F. A., and Goodrich, E. S.

3 A Treatise on Zaldi. Ed. by E. Ray Lankester. Pt. 3. The

Echinodermata. London

Clark, A. H.

09. The Affinities of the Echinoidea. Am. Nart., Vol. XLIII, pp.

682-686,

Clark, H. L.

79 The Synaptas of the New England Coast. Bull. U. S. Fish Com.,

1899, pp. 21-31.

700. The peng ey of Porto Rico. Bull. U. S. Fish Com., 1900,

PpP. —263.

02. The Tandon of the Woods Hole Region. Bull. U. S. Fish

Com., 1902.

Cuenot.

9] rongy morphologiques sur les Echinodermes. Archiv. de Biol.,

PAR

Gerould, J. hs

96. Anatomy and Histology of Caudina arenata. Proc. Boston Soc.

at, Hist., Vol. 27, pp. 8-74.

Grave, C,

03. On the Occurrence among Echinoderms of Larve with Cilia ar-

ranged in Transverse Rings, with a Suggestion as to their Sig-

nificance. Biol. Bull., Vol. V, pp. 169-186.

05. The Tentacle Reflex in a ee tk Cucumaria pulcherima.

Johns Hopkins Univ. Cire., Vol. 24, pp. 504-50

Henri, V.

03. Etudes des contractions rhythmiques des vaisseaux et du poumon

ee chez les Holothuries. C. R. Soc. Biol., Paris, T. 55, pp.

4-1316.

Johnson, R. A and Hall, R. W.

00. Variation and Regeneration and Synapta inherens. Science,

N. S., 1900, p. 178.

Lang, A

’96. Text-book of Comparative Anatomy. London, Vol. II.

Ludwig, H

96. Echinodermen. Bronn’s Klassen u. Ord. des Tierreichs. Bd. II,

Abt. 3, Buch 1

Morgan, T. H.

’01. Regeneration. The Macmillan Co., 1901.

A. D

81. Zoology for High Schools and Colleges, 1881.

Pearse, A. S.

08. Observations on the Behavior of Thyone briareus (Leseur). Biol.

Bull., Vol. XV, pp. 259-286.

09. Antotomy in Holothurians. Biol. Bull., Vol. XVIII, pp. 42-49.

SHORTER ARTICLES AND DISCUSSION

TERMS RELATING TO GENERIC TYPES

In the field of biological taxonomy an important reform is in

progress. The change is from fhe method of concepts to the

method of types, in order that names may be applied with

greater precision and permanence. Under the method of types

we no longer think of the technical name of a plant or an animal

as attaching primarily to a concept embodied in a description or

definition, but as relating to the first representative of the group

that became known to science. In determining the application

of a specific name we go back to the original specimen or type

on which the description was based. The original description

has become secondary to the original specimen. In like manner

generic names are treated as relating primarily to groups of

species, with the original species as the generic type.*

Without waiting to appreciate the fundamental nature of the

change from concepts to types, many systematic workers took it

for granted that generic types were to be determined by elimina-

tions in much the same way that generic concepts had been

treated, by gradual subdivision, restriction and removal of com-

ponent groups. The general results of elimination were the

same as under the method of concepts. The applications of

many of the older generic names did not become definitely fixed,

but remained dependent upon varying individual opinions of

the validity of the work of later authors. It often happened

that after elimination was accomplished only the doubtful or

unidentifiable species remained to serve as generic types. Grad-

ually it became apparent that the practise of elimination was

inconsistent with the method of types, and could not insure

stability in the application of names. Recourse was then had,

especially by zoologists, to the arbitrary designation of generic

1 Cook, O. F., 1898, ‘‘ The Method of Types,’’ Science, N. S., 8: 513; also

1900, ‘‘The Method of Types in Botanical serene ama Science, N. S.,

12: 475, and 1902, ‘‘Types and Synonyms,’’ Science, N. S., 15: 646.

Swingle, Walter T., 1913, ‘*Types of Spon in E F Taxonomy,’’

Science, N, S., 37: 864.

308

No.569] SHORTER ARTICLES AND DISCUSSION 309

types, the apparent object being to preserve the results of elimi-

nation, even though the theory had to be abandoned. Probably

it is only a question of time until the results of elimination will

be discarded, as well as the theory, and replaced by the actual,

historical types.

A plan for determining the historical types of genera was

adopted in 1907 in the American Code of Botanical Nomencla-

ture, and other applications of the method of types are being

rocognized by zoologists. Specialists in many groups are en-

gaged in the study of generic types, and the need of a special

terminology to facilitate work of this kind is becoming appar-

ent. Thus in Bulletin 83 of the U. S. National Museum, ‘‘Type

Species of the Genera of Ichneumon Flies,” by Henry L. Vie-

reck, two new terms, ‘‘isogenotypic’’ and ‘‘monobasie,’’ are em-

ployed in treating of the application of generic names to type

species. The paper is of interest, not only to students of this

group of insects, but also as an example of the tasks that con-

front all taxonomists who appreciate the need of basing their

work upon types. The distinctions to which the special terms

refer are undoubtedly useful, and the possibilities of express-

ing them in more convenient form are worthy of consideration.

The word ‘‘isogenotypic,’’ is used with reference to cases

where two or more generic names have been applied to the same

type species. For this purpose a new term is not needed unless

zoologists are unwilling to borrow from botanical nomenclature

a more convenient method of treating the same class of cases.

The botanical code provides a classification of synonyms, and

applies the word ‘‘typonym’’ to a name that has to be rejected

because an earlier valid name was proposed for the same type.

The formation and use of typonym are in accord with a familiar

analogy . As a preoccupied name becomes a homonym, it is easy

to remember that the use of a preoccupied type results in a

2A different combination might have been expected, such as ‘‘autogeno-

typie’? or ‘*deuterogenotypic,’’ since isogenotypic suggests the notion

of equall¥ good. types or of equal numbers of types, instead of con-

veying the idea of one and the same type, or of a second use of the same

ype. Genera have been termed ‘‘isotypical’’? when they were described

from more than one species, but all truly congeneric, on the assumption that

such species would have equal standing as types. A still older use of the

word ‘‘isotype’’ had reference to equal representation of a genus by similar

or corresponding species in different geographical regions or geologic periods.

See Schuchert, Charles, 1905, U. S. National Museum Bulletin 53, Pt. 1: 16.

310 THE AMERICAN NATURALIST (Vou. XLVIII

typonym. A name based on a different type species, but con-

generic with the type of an older genus, is termed a metonym.

A name rejected for lack of an identified type is a hyponym,

and one rejected for linguistic reasons, a caconym. All rejected

names fall readily into these five classes.

The other new term, ‘‘monobasic,’’ is used by Mr. Viereck to

indicate genera with only one species at the original place of

publication. In botanical literature the word ‘‘monotypic’’ is

often employed in this sense, though also applied to genera that

consist of only one species. If previous use disqualifies mono-

typic, the same objection lies against monobasic. In addition to

an older chemical meaning, the same word was employed several

years ago in a biological sense, to describe a condition of descent

in simple lines.* Apart from being preoccupied, the word mono-

basic has a misleading implication, since under the method of

types each generic name is referred to a single type species.

The idea of a genus being based on many types is discarde

with the method of concepts. Appreciation of this incongruity

may explain why no such term as ‘‘symbasic’’ or ‘‘polybasic’’

is used in contrast with monobasic, to indicate genera that were

first proposed in connection with more than one species.

Evidently there is need of a simple and consistent terminology

for indicating relations between generic names and type species.

The normal relation under the method of types is the designa-

tion of the type species at the original place of publication of

the genus. Genera provided with types by original designation

may be described as orthotypic, or normal-typed. With ortho-

typic genera there is no occasion to raise the question of how

many species were included at the original place of publication.

8 Cook, O. F., and Swingle, W. T., 1905, ‘‘ Evolution of Cellular Struc-

tures,’’ Bull. 81, Bureau of Plant telist stry, U. S. Department of Agricul-

tur re, p. 20. Plants or animals with specialized habits of asexual reproduc-

secon

tionary defines monobasis as follows: ‘‘The derivation of a stock from a

single parentage by inbreeding, or by propagation of buds or cuttings;

opposed to symbasis.’’ ee us the danger of ambiguity in using monobasis

for nomenclatorial purposes is greater than in using monotypic, though it

must be admitted that ma use of the word monotypic in two senses may

sometimes result in confusion. Genera that were monotypic in the strictly

nomenclatorial sense of being established in connection with one species may

not be monotypic in the more general taxonomic sense of including only one

species,

No. 569] SHORTER ARTICLES AND DISCUSSION 311

Genera that are not orthotypic fall into the two classes al-

ready considered, those with a single species at the original place

of publication, and those with two or more species. It is now

generally agreed that when only one species was mentioned this

should be accepted as the type. Such genera may be called

haplotypic, or single-typed. When two or more species were in-

cluded in the original treatment of a genus, and no type was

designated, we have the problem of subsequent determination of

the type, resulting in what may be termed a logotypic genus,

that is, a genus with a rationally selected type species. The

object of selection is to determine the historical type of the

genus. Names must have definite applications, and historical

applications of generic names can be made definite by ascertain-

ing the historical types. The recognition of a new generic group

is usually based on one leading or dominant species, with the

others added as associate members.

In many cases the generic type is intimated by the original

author in dividing the genus into subgenera or sections, in illus-

trating one of the species or citing illustrations published in

earlier works, in naming the genus with particular reference to

one of the species, in recording economic uses, or in giving geo-

graphical or other indications of greater familiarity with one of

the species. If the application of these or other historical cri-

teria leaves more than one species eligible for selection, the first

of the eligible species should be taken as logotype. In this way

it is possible to develop a consistent system of type selection that

will commend itself as reasonable and give the same results in

the hands of different students.*

Simply taking the first species under a generic name as the type would

viet establish more of the generic names in their historical places than

the method of elimination, which accepts the last of the original species left

in the genus as the type. Either of these methods of selecting types would

avoided by taking the historical considerations more directly into account,

as in the American Code si peaa cal Nomenclature. Probably a more sat-

isfactory system for associating generic names wit their historical types

could be developed by ait study of the problem. A policy of refusing

to revive generic names that were not directly associated with binomial spe-

mitigated priority. In proposing lists of ‘‘nomina utique conserv vanda

advance of any provision for the definite application of names, European

botanists have demonstrated one more way to put the cart before the horse.

312 THE AMERICAN NATURALIST [Vou. XLVII

In addition to the three ways of associating generic names

with their type species, there are many cases where generic

names have been applied to groups that do not include the type,

or any of the original species. Formal assignments of errone-

ous types also occur when generic names are not traced back to

their original places of publication, or when ineligible species

are designated as types. In dealing with the synonymy of

genera previously treated under names that belong to other

groups it will be convenient to have a distinctive term for this

class of cases. Such misplaced names, applied to groups that do

not contain the true type, may be indicated as pseudotypic, or

false-typed.®

It should be expected that more critical analysis of taxonomic

problems would lead to more definite distinctions and more pre-

cise terms. The older terminology was developed to facilitate

the study of names, whereas it is now apparent that provision

must be made for the study of types as another formal branch

of biological taxonomy. Nomenclature has a history of three

hundred years while systematic typology is only beginning. To

gain further insight into these typological problems is obviously

more important than to attempt premature applications of par-

tial solutions. It may take fifty or a hundred years to transfer

Failure to regulate the application of names is the fundamental defect of the

Paris and Vienna codes, and is hardly to be cured without thorough re-

cas

sae the palm genus Martinezia, as treated by Kunth, Martius, and

many later writers as relating to Martinezia caryotefolia and its immediate

relatives, was pseudotypic, for this species does not appear to be congeneric

with any of the five species originally referred to Martinezia by Ruiz and

Pavon. Hence it has been proposed to a ee this pseudotypie use of Mar-

tinezia by a new generic name, Tilm (See Bull. Torrey Bot. Club, 28:

65.) The five original species of Maries belong to three natural groups,

now recognized as distinct families, the first two species to the Cocacex, the

third species to the Acristacee and the others to the e nosed The

third species, M. ensiformis, should be taken as logotype of Martinezia be-

cause the figures used to illustrate the generic characters evidently orii

a member of the family Acristaceæ. Another reason for excluding the

cocoid species from consideration as type is that the ey are mentioned as

deviating from the ‘‘essential characters of the genus,’’ in connection with

the original description. The rule of the Vienna code, to the effect that the

name of a subdivided genus should go with the majority cf the species, would

carry the name Martinezia over to the family Chamedoreacew. The making

of such a rule shows that many European botanists were still working under

the method of concepts, and were not copra to think of generic names

as inseparably connected with type spec!

No. 569] SHORTER ARTICLES AND DISCUSSION 313

the whole structure of biological taxonomy to the new founda-

tion of types. To suppose that any permanent advantage can

be gained by elaborating defective methods under forms of legis-

lative enactments or judicial decisions is to show a limited ap-

preciation of the nature of the subject and of its historical de-

velopment. As long as legislation and interpretation are based

on inadequate study, they can represent, at most, only a tem-

porary consensus of opinion, for it is of the very nature of

science to condemn and throw aside any doctrine or method that

has proven inadequate or fallacious.

TERMS RELATING TO SYNONYMS

The following classes of synonyms were recognized in 1907,

in the American Code of Botanical Nomenclature :*

Homonym.—A name rejected because of an earlier applica-

tion of the same name to another genus.

Typonym.—A name rejected because an older name was based

on the same type.

Metonym.—A name rejected because an older valid name was

based on another species of the same genus.

Hyponym.—A name not associated with a type."

6 Bulletin of the Torrey Botanical Club, 34: 167, 1907.

7 Much confusion would be avoided by a consistent policy of withhold-

ing recognition of generie names that have not been associated with type

species. Thus the name Acoeloraphe, proposed by Wendland in 1879 in an

analytical key to genera of fan-palms (Bot. Zeitung, 37: 147), was not as-

sociated with a type, though evidently relating to a species mentioned in the

Same paper as ‘‘ Brahea serrulata.’’ This Florida palm differs from the

Mexican type of Brahea in the leaf characters assigned to Acoeloraphe in

rop

by Hooker f. in 1883 for ‘‘ Sabal sormin Rh. o ee (Genera Plantarum,

3: 926). All subsequent writers have accepted Hooker’s name, and Acoel-

oraphe should remain under Serenoa as a hyponym. Nothing has tended so

strongly to bring the principle of priority into disrepute as the incontinent

revival of abortive names, to replace properly agrees names in current

use. No species was referred to Acoeloraphe until 1907, when Beccari

(Webbia, 2: 107 ) applied the name Acoeloraphe Gigni to a Cuban mem-

ber of a genus that had been described in 1902 under the name Paurotis,

a Bahaman species, Pawrotis androsana, being the type (Mem. Torrey Bot.

Club, 12: 21). This transfer of the name Acoeloraphae to the genus

Paurotis was followed by Sargent in 1911 (Trees and Shrubs, 2: 117), but

Beceari’s genus Acoeloraphe is a metonym of Paurotis, and is also pseudo-

314 THE AMERICAN NATURALIST [Vou. XLVIII

TERMS RELATING TO TYPE SPECIES

Orthotype.—Type by original designation. A species desig-

nated as type in connection with an original publication of a

generic name. A genus whose type was-formally designated at

the original place of publication is orthotypic.

Haplotype.—Type by single reference. A single species re-

ferred to a genus at the original place of publication, and on

this account accepted as the type. A genus proposed with refer-

ence to a single species is haplotypic.

Logotype.—Type by subsequent determination. The histori-

cal type of a genus, selected from two or more original species.

A genus whose type is selected from two or more original species

is logotypie.

Pseudotype.—Erroneous indication of type. A species erro-

neously indicated as the type of a genus. A genus treated on

the basis of an erroneous type, or so as to exclude the true type,

is pseudotypie.

O. F. Cook

BUREAU OF PLANT INDUSTRY,

U. S. DEPARTMENT OF AGRICULTURE,

March 13, 1914

typic, because of the original application of the name to Serenoa. Two

species of Paurotis are supposed to exist in Florida, one that is identi-

fied with the Cuban P. wrightii (Grisebach & Wendland) and a local species,

P. arborescens (Sargent).

NOTES AND LITERATURE

LINKAGE IN THE SILKWORM MOTH

ONE of the most striking recent developments in the study of

genetics is the discovery of linkage in many of those forms which

were supposedly thoroughly worked out. The most recent ex-

ample is a very interesting paper by Y. Tanaka,’ entitled

“‘Gametic Coupling and Repulsion in Silkworms.’’ The data

presented in this paper demonstrate the existence in the silk-

worm moth of a group of four pairs of linked genes. Following

Tanaka’s nomenclature we may designate these genes as follows:

N, which differentiates the larval color pattern known as

‘‘normal’’ from that called ‘‘plain’’; S, occurring in larve

having the ‘‘striped’’ pattern, and epistatie to N; M, the differ-

entiator for the ‘‘moricaud’’ larval pattern, also epistatic to

Y, the gene which differentiates caterpillars with yellow

blood and yellow cocoons from the recessive whites. Of the six

possible combinations of these genes, taken two at a time, all

but NM and SM were made, and all showed linkage. F,

‘“‘eoupling’’ tests, i. e., from matings where both dominants

entered the cross from ‘the same P, individual, were made for

SY and for MY. In each case there occurred cross-overs, or new

combinations of the characters, in such proportions as to lead

Tanaka to suppose the ratio of parental combinations to cross-

overs among the gametes to be about as 7:1. ‘‘Repulsion”’

(where one dominant entered from each P, individual)

results were obtained for NS and for NY. In neither case did

any double recessives (cross-overs) appear, though over 3,000

caterpillars were obtained in the case of NY, and 224 in the case

of NS. From these data Tanaka concludes that the repulsion was

complete in these two cases. It has, however, been pointed out by

Morgan? that such results will be obtained if the linkage is com-

plete in one sex only. In Drosophila such ‘‘repulsion’’ crosses

never produce double recessives in F,, and it has been shown

that this is due to complete linkage in the male, crossing over

being frequent in the female between some pairs of genes. In

order to test this possibility it is necessary to mate doubly hetero-

zygous individuals to double recessives, when the gametic ratio

is obtained directly and without the complications present in

most F, results. It so happens that Tanaka reports two such

crosses, one for each sex, though he does not recognize their im-

1 Jour. Coll. Agr., Tohoku Imper. Univ., Sapporo, Japan, V, 1913,

2 Science, N. S., XXXVI, 1912.

315

316 THE AMERICAN NATURALIST [Vowu. XLVII

portance in this connection. When a male heterozygous for S

and for Y, one dominant having been derived from each parent

(SysY), was mated to a doubly recessive (sysy) female, there

were produced 63 Sy and 65 sY—no cross-overs. A female

heterozygous also for S and for Y, but having them ‘‘coupled’’

(SYsy), was mated to a male sysy, and produced 215 SY and

188 sy—again no cross-overs. Yet that crossing over may occur

between these two pairs of genes is shown by the fact that the

‘‘coupling’’ F, results indicated a gametic ratio of about

:1:1:7. We are, therefore, still left in the dark as to whether

crossing over occurs in only one sex, or in both. But it is certain

that the strength of linkage in this case is not always the same—

a point of great interest and importance. Similar cases have been

reported by Baur? in Antirrhinum, by Punnett* in the sweet pea,

and by the writer® in Drosophila.

Tanaka refers to his case as differing from previously reported

eases of linkage in animals in that the sex differentiator is not

one of the genes involved, and in that the linkage is sometimes

only partial. However, he refers several times to a paper by

Morgan® in which it is clearly shown that three of the sex-linked

genes in Drosophila also show partial linkage to each other, inde-

pendently of their sex-linkage. Punnett,’ in referring to the

same paper, has said, ‘‘Morgan’s experiments with Drosophila

suggest coupling of some kind between factors for eye color and

shape of wing, though both of these factors may show sex-limited

inheritance in other families.’’ A study of the data referred to,

or of any of the similar data on Drosophila since published, will

show that these genes always show sex-linkage, and that at the

same time they always show linkage to each other when both can

be followed in the analysis. The two phenomena are not mutu-

ally exclusive, but both are always present.

Both Tanaka (in a footnote) and Punnett refer to the latter’s

ease in rabbits as the first example of linkage in animals not

involving sex. If the linkage between sex-linked genes is, for

some strange reason, not considered to belong in this category,

there are still at least two cases which antedate Punnett’s slightly.

A few months before Punnett’s paper appeared I had suggested?

the possibility of linkage in mice. It now seems rather probable

that the relation in both mice and rabbits may really be that of

3 Zeits, f. ind. Abst.-u. Vererb.-Lehre., VI, 1912.

4 Jour. Genet., III, 1913.

5 Science, N. S., XX XVII, 1913.

6 Jour. Exp. Zool., XI, 1911.

7 Jour. Genet., II, 1912 (Nov.).

$ AMER. NAT., XLVI, 1912 (June).

No. 569] NOTES AND LITERATURE 317

triple allelomorphism. For this reason I am inclined to assign

priority to Morgan and Lynch,’ whose paper on linkage of genes

in Drosophila which are not sex-linked appeared after my own

paper and before Punnett’s.

CoLUMBIA UNIVERSITY A. H. STURTEVANT

NABOURS’S BREEDING EXPERIMENTS WITH

GRASSHOPPERS

IN a recent paper, Nabours (’14) describes breeding experi-

ments that he has been carrying on for some years with grouse

locusts of the genus Paratettiz. His work is of special interest

in showing that in a wild species there exists a number of distinct

types that show alternative inheritance of a particular kind. His

paper may be summarized as follows:

1. Nine distinct, true breeding forms of Paratettix were col-

lected ‘‘in nature.’’ These ‘‘species’’ (as Nabours is inclined

to consider them) ‘‘are mainly distinguished by their striking

color patterns. ”?

2. When an individual of one of these species is mated to one

of a different species the hybrid character of the offspring is

apparent at once, in that ‘‘all the characters of each parent are

represented in the F, hybrid.’ In other words, the hybrid is in

a certain sense an intermediate, and ‘‘the terms dominant and

recessive’’ are probably not ‘‘applicable at all. This point,

while of little theoretic importance, has a practical value in that

the zygotic constitution of any hybrid can be recognized without

further breeding tests.

3. With one exception, each color pattern factor was found to

behave as an allelomorph to any other color pattern factor.

_ 4. The various lengths of the wings and pronotum are appar-

ently not inherited, as such but are determined by environmental

factors, especially such as tend to prolong or to shorten the length

of larval life.

It appears that Nabours confuses the relation of the facts men-

tioned under 3, and that he supposes this to be the ordinary

behavior of ‘‘mendelizing characters,” for he says:

The essential result of these experiments has been the extension of

this principle [Mendelian inheritance] to a considerable number of

types of a phylogenetically low group of ametabolous insects.

To be sure, he recognizes that other workers in genetics have

‘an attitude quite different from his, and he takes some little pains

to make clear his own point of view. To quote again (p. 142):

® Biol. Bull., XXIII, 1912 (Aug.).

318 THE AMERICAN NATURALIST [Vou. XLVIII

The existence of unit characters in the De Vriesian sense does not

appear to have been as clearly demonstrated as that of alternative in-

heritance . . . and the interpretations are at great variance. Thus, one

group of authors [reference made to Bateson, Doncaster, and Tower]

recognize characters in organisms that ean be replaced by other char-

acters when the proper crosses are made, . . . while on the other side

there are those [references to Whitman and Montgomery] who believe

that the organism as a whole is the only unit and that there are no

actual unit characters.

Again he says (p. 169) :

No character of one parent species is ever replaced in the F, hybrid

by any character of the other parent. All the characters of each parent

are represented in the F, hybrid. It follows then that these grass-

hoppers do not exhibit characters which by crossing can be replaced by

other different characters; the whole pattern appears to be the only

unit.

There is no real conflict between Whitman’s idea and the

accounts given by students of Mendelism, for the latter realize

that far-reaching somatic effects may result from a single factor,

and the composite character of the hybrid is not an uncommon

osourrenda. Nabours identite a particular pattern with the

‘‘organism as a whole,” but since his evidence relates here to

color patterns only, nothing is gained by the introduction of such

a vague phrase as the ‘‘organism as a whole.’’ Specifically he

shows that the hereditary differences between any two types can

be explained on the assumption of a single differential for each

ase.

With reference to the antithesis presented by Nabours, it must

be recognized that the modern literature of Mendelian heredity

affords innumerable instances where two or more characters

entering from one parent and their allelomorphs from the other,

reappear in the F, generation in new combinations.

If we assume with Nabours that each of the eight color patterns

are represented by a characteristic condition of the ‘‘germinal

material,’? we may use his terms A, B, C, D, E, F, H or I to

symbolize this ‘‘germinal material’? for the various color

patterns. As Nabours uses the terms, an individual homozygous

for A is represented simply by A, and a hybrid between A and B

by AB. In ordinary usage, the homozygous form would be

represented as AA and its germ cells by A. This is a minor

matter. Ordinary usage has the advantage of being more

consistent.

According to Nabours, then, A mated to B gives AB; B mated

to F gives BF; C mated to E gives CE, ete. In gametogenesis

these factors segregate, so that, for example, BA gives germ cells

No. 569] NOTES AND LITERATURE 319

A and B; BF gives B and F, ete. In other words, he treats the

matter as if he were dealing with a system of multiple allelo-

morphs, though he nowhere specifically calls them such. From

this point of view there are eight distinct allelomorphs con-

cerned with color pattern any two of which may constitute a

pair; in any zygote two allelomorphs (perhaps alike, perhaps

unlike) will be present, and in any gamete only one of the eight

will normally occur.

With one exception of which I will treat later, all of Nabours’s

results can be explained by this hypothesis. This sort of explana-

tion is not new. (Shull (’11), de Meijere (710), Sturtevant

(713) and others have used it to explain results obtained in

Lychnis, Papilio, rabbits and other forms, and it will almost un-

doubtedly be shown to apply satisfactorily in still other cases.

The exception just mentioned occurred in the cross which

Nabours describes at the bottom of page 156 (e). Here a male

of the constitution CE was mated to a female of the constitution

BI. On Nabours’s theory, the gametes of the male should carry

C or E, but not both, and the gametes of the female should carry

B or I, but not both. The union of the two kinds of sperms with

the two kinds of eggs should give four classes of offspring, and

these were in fact obtained; viz., 12 BC, 11 BE,7 CI,10 EI. But

there appeared also one individual BEI. Nabours ’s explanation

of the case is that perhaps the BI ‘‘female parent gave at least

one gamete containing the factors for the patterns of both her

parents and that this double character gamete was fertilized by

one of the E gametes which came from the CE male.’’* Let us

see whether this is the most probable interpretation.

As Sturtevant has pointed out, for any case to which the idea

of multiple allelomorphism is applicable, an equally valid ex-

planation may be found in ‘“‘complete linkage’’ of the factors

concerned. To decide in any case between the two explanations

would be impossible.

If, however, linkage were not complete, a ‘‘eross-over’’ class or

‘‘recombination’’ class might occur, and this would suffice to rule

out the explanation based on multiple allelomorphs.

Such a ‘“‘eross-over’’ class perhaps is furnished by the BEI

individual, The demonstration of this may be given by the use

of symbols, as follows:

Let us assume that A is the allelomorph of a, B that of b, C of

c, D of d, F of f, I of i, ete., making eight pairs of allelomorphs

altogether. Assume that each gamete of any individual carries

1 This explanation is essentially similar to that advanced by Bridges (713)

to explain certain peculiar results in Drosophila. Bridges assumed that in

gametogenesis the two X-chromosomes of a white-eyed female failed to segre-

gate (in Bridge’s terminology, non-disjunction occurred), and passed over

together into one gamete.

320 THE AMERICAN NATURALIST [Vou. XLVIII

one allelomorph of each pair, and that the eight factors thus

present in a gamete form a linked group, tending to segregate as

a unit in gametogenesis. Thus Nabours’s form A would give

gametes of the form Abcdefhi. AB would give gametes of only

two forms, one corresponding to A and the other to B, viz.,

Abcdefhi and aBcdefhi. Two other forms are possible, formed

by the exchange of A with B, and of a with b, but these will not

occur if linkage is complete. In dealing with the hybrid AB in

practise the factors cdefhi would not be put into the formule,

as they are alike in all gametes.

These rules would apply similarly to all other species and

hybrids. Therefore in the case in which the BEI individual

occurred, we would represent the male parent, which Nabours

designated CE, by bCei—bcEi, and its gametes by bCei and bcEi.

The female parent, which Nabours designates BI, we would

represent by Bcei—bceI, and its gametes would be Bcei and

bcel if linkage were complete. If linkage were not complete there

would occasionally be formed gametes bcei and Bcel. One of

these latter (Bcel) was probably formed and fertilized by a

sperm of the type bcEi, thus giving rise to the BEI individual.

No gametes corresponding to bcei appear to have been fertilized,

though of course we do not yet know what the appearance would

be of an individual so formed.

This matter would be easy to test, and it is to be hoped the

cross may be repeated. If then BEI forms should appear again

and in these when mated to other forms the factors B and I should

be found to stay together to the same extent as they before sepa-

rated, it would show that close linkage, rather than multiple alle-

lomorphism explains this particular instance.

It may be, too, that both linkage and multiple allelomorphism

play a part in the production of these phenomena. In any case

it seems that the test is at hand, and not difficult to perform,

excepting in so far as there are practical difficulties connected

with the rearing of the grasshoppers in sufficient numbers to

cover the point.

LITERATURE CITED.

Bridges, ©. B. 1913. Non-disjunction of the Sex Chromosomes of Droso-

phila. Jour. Exp. Zool., Vol. 15.

de Meijere, J. ©. H. 1910. Ueber Jacobsons Ziichtungsversuche bezüglich

Polymorphismus von Papilio Memnon L. 9, ete. Zts. ind. Abst.-

Vererb.-Lehre, Vol. 3

Nabours, R. K. 1914. Studies of Inheritance and Evolution in Orthoptera.

I. Jour. Genet., Vol. 3

Shull, G. H. 1911. heversihte Sex Mutants in Lychnis dioica. Bot. Gaz.,

Vol. LII.

Sturtevant, A. H. 1913. The Himalayan Rabbit Case with some Consid-

erations on Multiple Allelomorphs. Am. Nat., Vol. XLVII.

COLUMBIA UNIVERSITY Joun S. DEXTER

VOL. XLVIII, NO. 570 JUNE, 1914

bet

HI.

: ares and Evolution. By X -

THE

AMERICAN

NATURALIST

A MONTHLY JOURNAL

Devoted to the Advancement of the Biological Sciences with

Special Reference to the Factors of Evolution

CONTENTS

Page

pies nen by Hybridization and Mutation. “Professor — JOHN H. ii

Rereaiey of Bristles in the Common gar ges ie Phen Study í of Factors

verning Distribution. PHINEAS W. 339

Pagsilogent eS ana CASIN peace in Alfalfa Breeding. GEO. <r

~ ~ — 369

Shorter Articles and Discussion: Nabours’s : Greninibiak Multiple Allelo-

EAT ae Tankage and Misleading ‘ Terminologies i in Geneties Profemor Ww. a

THE SCIENCE PRESS

LANCASTER, PA. GARRISON, N. ¥.

NEW YORK: SUB-STATION 84

The American

Naturalist

MSS. intended for publication and books, etc., intended for review should be

‘sent a oe of THE AMERICAN NATURALIST, Garrison- -on-Hudson, New York.

or u

icles containin

arie

s of research work beari

pae ~ Selaci evolution are sopoclaly welcome, and will be given preference

in pas

reprints of Sprong are supplied to authors free of charge.

saan ndre

Fariba reprints “will be supplied a

ubscriptions an

subscription price is four dollars a

Canadian postage twenty-five ce

forty cents.

advertisements should be sent to the publishers. The

ear.

nts additional.

The advertising rates are Four Dollars for a page

oreign postage is fifty cents and

The charge for single copies is

THE SCIENCE PRESS

NEW YORK: Sub-Station 84

Lancaster, Pa.

Entered as second-class matter, April 2, 1908, at

Co.

Garrison, N. Y.

the Post Office at Lancaster, Pa., under the Act of

mgress of March 3, 1879.

FOR SALE

ARCTIC, ICELAND and GREENLAND

BIRDS’ SKINS

Well Prepared Low Prices

Particulars of

G. DINESEN, Bird Collector

Husavik, North Iceland, Via Leidle, England

WANTED TO PURCHASE

a set of BIRDS OF AMERICA by J. J. Audubon,

7 or 8 volumes, please report, stating cash price, stat-

ing condition, binding and dates of volumes.

F. C. HARRIS,

Box 2244 Boston, Massachusetts

TO OOLOGISTS

and ENTOFMOLOGISTS

W. F. H. ROSENBERG

Importer of Exotic Zoological Specimens

57, Haverstock Hill, London, N. W., England

Begs to announce the publication of a new Price

List (No. 20) of Mammals, including over 400 species

from various parts of the world.

These will be mailed ti on application, as well

as any of the following lis

Birdskins (over 5000 an . Birds’ Fega (o ( na

1,100 species); Reptiles, Batrachians an

(over 400 species); Exotic Lepidoptera fore a 8, 000

species

Largest stock in the world of specimens in all

branches of Zoology

All Museums and Amateurs

should write for these Lists

All specimens sent on approval.

Please state which lists are required and give

name of this periodical.

Photographic Films Developed

5 cents per Spool, Prints 3 cents up, Post

Card: ts.

8x10 Enlargements 25 cents mounted.

Send for Samples and Price List.

FRANK R. BARBEAU

89 West Bridge Street Oswego, N. Y-

The University of Chicago

Mitchel! Tower Chicago, Hin —

THE

AMERICAN NATURALIST

VoL. XLVIII - June, 1914 No. 570

SPECIES-BUILDING BY HYBRIDIZATION AND

MUTATION

PROFESSOR JOHN H. GEROULD

DARTMOUTH COLLEGE

THe mystery that has surrounded the origin of new

species in the incipient stages of their evolution has lately

been penetrated and cleared away to a large extent by the

light of studies in Mendelian inheritance and the attend-

ant idea of mutation. Species building is no longer a

hypothetical process based on the preservation of minute,

useful, fortuitous variations, but it is a process open to

observation and experimental control. Its raw materials

are variations that are usually not minute, useftl or for-

tuitous, but clean-cut unit characters, tending to vary

only in certain limited, well-defined directions depending

upon the chemical peculiarities and physical structure of

the particular form of protoplasm, and, in the vast plural-

ity of cases, nonuseful.

The fields of systematic zoology and botany, illumi-

nated by the new science, genetics, are emerging from the

mists of formalism, and invite biologists of the broadest

type to exploration. The geneticist turns to systematics

for many of the materials with which to solve the prob-

lems of organic evolution. The systematist sees that in

order to keep abreast of the times he must stand ready

to rebuild his pigeonholes and test with experiment that

which he puts into them.

Every one occupied with zoology or botany realizes

that there are no adequate criteria by which this or that

assemblage of individuals is or is not to be regarded as a

321

322 THE AMERICAN NATURALIST [Vou. XLVIII

distinct species. Arbitrary rules for species making, de-

signed to restrict the activities of the more vigorous

“ splitters ’’? have been indeed laid down by experienced

and conservative systematists. The final test, however,

so far as any exists, is acknowledged to be whether a

group breeds approximately true to its kind and is ap-

‘proximately sterile with other closely related stock, and

yet in how few cases have both or either of these criteria

been actually applied by the describer of species!

As a matter of fact no stock that has been bred on a

vast scale, so far as I am aware, breeds absolutely true to

specific characters. In Morgan’s Drosophila! and De

Vries’s Œnothera, numerous mutants appear, probably

through the absence of certain chemical elements, or by

unusual combinations of elements, in the chromatin of the

germ plasm. That this phenomenon has not been shown

for many other species is due, in all probability, to lack

of close attention to all the individuals in a huge proces-

-sion of stock in the process of breeding. Any insect bred

as extensively as Drosophila ampelophila, the pomace fly,

has been would probably show as many mutants; some

would show more. Colias eurytheme, the ‘‘ orange sul-

phur ”’ or alfalfa butterfly, is such an example. Though

this butterfly can not be bred on a scale comparable with

Drosophila, every thousand individuals yield many dis-

continuous variations: red eyes instead of green, tongue

uncoiled instead of wound in close flat spiral when at rest,

one antenna shorter than the other, the absence of certain

spots from the wings, gynandromorphism, caterpillars

with two longitudinal rows of large black dorso-lateral

spots or white dorso-lateral stripes upon a dorsal surface

usually unmarked, caterpillars with one proleg less upon

one side than the other. This is a partial list of points at

which the descendants of three females of Colias eury-

theme failed in a single summer to breed true to the char-

acteristics of the species, though bred under uniform

normal conditions. The fact that these discontinuous

1 Science, N. S.; Vol. XX XIII, Nos. 847, 849, pp. 496-499, 534-537, 1911.

No. 570] SPECIES-BUILDING 323

variations appear under uniform external. conditions

leads one to be very skeptical toward most of the past

experimental work supposed to show the effects of the

environment upon insects in modifying the germ cells.

Any one wishing to try an experiment on the production

of variations by the influence of the environment, or upon

the inheritance of acquired characteristics, should deny

himself absolutely this privilege until he shall have bred

under normal conditions at least a thousand individuals

of the stock that he will subsequently employ.

That species necessarily breed true to the specific char-

acters ascribed to them by their inventors is an unveri-

fied dogma. At best the reporter picks out stray individ-

uals here and there from a vast procession of which he

can only see glimpses, and, trusting to the credulity of the `

public in the established ideas about these matters, he

creates upon paper a new species. Doubtless the unit

characters of ‘‘ specific ’’ grade in the stock of some spe-

cies are more generally constant or homozygous than

those of certain others, but it is reasonable to suppose

that, owing to dominance the heterozygous? condition re-

garding certain characters is frequently masked and un-

noticed in apparently pure strains of wild stock. If the

heterozygote respecting a certain character be compara-

tively rare, or if it be a heterozygote based on several

interacting factors, like redness in the kernel of Nilsson-

Ehle’s wheat, it may cross again and again with

the homozygous dominant, or with another heterozygote

of similar nature to itself, without the appearance in the

population of the recessive. That specific and varietal

characters do exist in heterozygous condition in wild

stock of ‘‘ pure’’ species, unmasked by dominance and

easily detected, I have found to be the case in Colias at

Several points. The color pattern as a whole apparently

fluctuates in variation, but these variations in detail are

2 The mixed Saia condition, D(R), producing germ cells D and E

in equal num

3 Act. sai Tavid, 1909.

324 THE AMERICAN NATURALIST [Vor. XLVIII

strictly a matter of inheritance. Its ‘‘ fluctuation ’’ is not

due to a difference in environmental conditions surround-

ing different individuals, but evidently to the condition of

the germ plasm. The parents of any brood may be

heterozygous or homozygous for the determiners of color

pattern. If they come from a strain homozygous in this

respect and are alike in appearance, the offspring will

resemble the parents closely and show a narrow range of

variation, but if unlike and derived each from unlike

parents, a wide range of inherited ‘‘fluctuation’’ occurs.

Such is often the case in the inheritance of a melanic

tendency so often attributed to the action of the environ-

ment, and of spots used in the diagnosis of species as, for

example, the conspicuous spot in the middle of the under

side of the hind wing. This is commonly double in Colias

philodice and C. eurytheme, consisting of a chief and an

accessory spot, single in C. paleno, an arctic cireumpolar

species, but it varies enormously. In ewrytheme and

philodice the accessory spot may be absent; in paleno, in

rare cases, it may be present. I have bred large families

of C. eurytheme in which both the chief and accessory

spots were, like those of the parents, almost uniformly

large and nearly equal in size. In other families, from

parents in which the accessory spot is nearly or quite

lacking, the offspring show a similar reduction. In

C. philodice I have found it possible by selection to estab-

lish a race devoid of the row of submarginal red-brown

spots of the under side of the wings. Thus, by selection,

strains, nearly or perhaps quite homozygous for definite

points of color pattern, may be established, derived from

a population which in the main is in an extremely hetero-

zygous condition. Yet species are named and distin-

guished on the basis of these features. :

Another example of heterozygous condition of a char-

acter within a wild species is the white pigment in the

ground color of the ‘‘ albino ” female of Colias, both in

the yellow species, philodice and the orange species,

eurytheme. The white female is regularly heterozygous

No. 570] SPECIES-BUILDING iS

for this sex-limited character. Her daughters are white

or colored (yellow or orange, as the case may be) in equal

numbers. Still another interesting heterozygous feature,

though not of ‘‘ specific ’’ grade, was seen last summer in

a pure strain of Colias eurytheme. A female appeared

that could not upon stimulation coil up her tongue.

Mated with a normal male, this abnormality was in-

herited in various degrees by half her offspring (37

uncoiled and 28 coiled). One of her daughters, abnormal

in this respect and mated with a normal of a different

strain, transmitted the abnormality to about 16 per cent.

of her offspring (29:151), showing that the possessor of

this abnormality is regularly heterozygous in respect

to it.

Whether @nothera lamarckiana is or is not a complex

hybrid produced from two American species, is it not

certain that, like other wild and cultivated stock, it does

possess characters for which it is heterozygous, and that

the watcher for mutants frequently seizes upon rare com-

binations of recessive features as a part of his elementary

species?

But to breed true is only a secondary criterion of spe-

cies. Inbred strains of domestic animals and plants do

that to a certain degree. Varieties and races to a certain

extent may do the same. The real criterion (and the one

least often practically used by the systematist) is fertility

within the group and sterility with other closely related

groups. Here dogma holds sway among writers on or-

ganic evolution as well as among systematists, for we are

told by those who have been accustomed since childhood

to the idea of the objective reality of species that hybridi-

zation of species, that is, genuine species in good and

regular standing before the scientific public, has played

very little part in the origin of new species. This atti-

tude was entirely logical in view of the accepted ultimate

definition of a species. If the individuals of one species

are actually sterile with members of another, hybridiza-

326 THE AMERICAN NATURALIST [Vou. XLVIII

tion of species can not play an important part in the manu-

facture of new wild strains. But only in comparatively

rare instances have attempts been made experimentally

to mate Linnean species. The dogma of the objective

reality and uniform value of the species unit has diverted

us from seriously attacking this problem. Just as in the

nineteenth century the fixed idea of the immutability of

species blocked the progress of the doctrine of evolution,

so this dogma now stands in our way, and obstructs the

possibility of vision. We need now fully to recognize the

fact, which most biologists are ready to admit, that the

term species is applied to most heterogeneous groups of

individuals, groups of every conceivable size, based on

differences that are most diverse in number and impor-

tance, often separated from allied groups entirely by the

arbitrary judgment of the describer, and depending ulti-

mately upon his personal temperament. These groups,

as already stated, have been tested in comparatively few

instances by the only reputable criterion that can be

_ applied in the separation of closely allied groups, that of

sterility or fertility inter se.

To one who tries to divest himself of the accepted ideas

regarding species and is on the watch for evidence of

hybridization among unlike strains that we are accus-

tomed to call species, new cases of such hybridization

frequently come to light. Especially is this true among

the insects. In regions where the faunal areas of two

‘ vood ’’ species overlap or are contiguous, such crossing

not infrequently occurs.

A most interesting case is that of the four species of

the coccinellid beetle Adalia that occur in the same region

in Colorado, as worked out by Palmer. These four forms

with clean-cut differences in color and color pattern had

been named and described by different authors as dis-

tinct species, yet three of them were found to be inter-

breeding with complete fertility but still respectively

maintaining their identity, forming a regular Mendelian

4 Annals Entom. Soc. America, IV, 3, September, 1911.

No. 570] SPECIES-BUILDING 327

epistatic series: a red-brown spotless form, melanopleura,

dominant at one end of the series, then annectans, a red-

brown, spotted type, and finally the recessive, melanic,

red-spotted humeralis with a color pattern different from

that of annectans or of Coloradensis, another red-brown,

spotted type of that locality. ‘‘ But ’’ says the upholder

of the present idea of species, ‘‘ here we have a single

polymorphic species, not three or four different species.

The breeding experiments show that the describers of

these forms were wrong in ascribing systematic rank to

mere color varieties.’’ It goes, of course, almost without

saying that the makers of these species did not before

naming their beetles, breed them to determine whether

they would breed true to type and were infertile

inter se. Indeed, in how few cases has this been done!

Even the larval stages of most known beetles are imper-

fectly unknown, much less the possible genetic relation-

ship of one type to another, as determined by breeding

them to maturity. Blaisdell’ describes the case of two

Californian Coccinellide which are found in winter in

small groups under the bark of eucalyptus trees. ‘‘ Usu-

ally there was one Olla plagiata with each of the groups

[of O. abdominalis], irrespective of whether they were

made up of two or more individuals.’’ The same author,

by selection of specimens of abdominalis representing

different types of color pattern, describes its range of

variation, but adds that his studies throw no light on the

relationship of the two species. Had he bred certain in-

dividuals of O. abdominalis together, it is not at all un-

likely, in view of his observation of the regular occur-

rence of a few plagiata in every group of abdominalis,

that the former interbreeds with the latter and may be a

simple recessive in respect to it. Miss Palmer’s work

on the allied Adalia certainly suggests this as a possi-

bility.

Another remarkable case is that of the nine true-breed-

ing species of grouse-locust, Paratettix, recently de-

5 Entom. News, Vol. 24, No. 9, November, 1913.

328 ; THE AMERICAN NATURALIST [Vou. XLVIII

scribed by Nabours.* These nine color types, or species,

freely interbreed. The color pattern of the resulting F,

hybrid in each case is a mosaic combination of those of

the two parents. The latter in subsequent inbreeding

may be extracted intact, each having been transmitted as

a distinct unit, without dominance.

In Lepidoptera, an order in which polymorphism is

notoriously common, hybridization between species has

been frequently observed. Standfuss’ devotes eight

octavo pages of his excellent ‘‘ Handbuch ’’ simply to the

enumeration of examples of such hybridization between

palearctic species of moths and butterflies, and acknowl-

edges that he mentions only a fragment of all such cases

on record or preserved in collections. This list would be

greatly extended if American species were included.

Seven different hybrid combinations within the genus

Colias in the palearctic region have been noted by Stand-

fuss.

Colias philodice, the clouded sulphur or clover TRE E

of the eastern and central United States, readily crosses

with C. eurytheme, the orange sulphur or alfalfa butterfly

of the western and central states. The territory of philo-

dice, according to Scudder extends like a wedge westward

from the Atlantic into the faunal area of eurytheme.

Overlapping thus occurs in the Mississippi Valley, though

philodice does not extend as far southward as the Gulf

States, Texas, Louisiana and Mississippi, in which

ewrytheme is found.

These two species are fairly sharply distinguished by

the difference in the ground color, which in eurytheme is

orange, in philodice sulphur yellow. The middle spot of

the upper side of the hind wing is brilliant orange in

eurytheme, pale orange or yellow in philodice. The dark

border of the hind wing of the female is wider in eury-

theme than in philodice and broken with a row of large

yellow spots.

6 Journal of Genetics, Vol. 3, No. 3, February, 191

T ‘t Handbuch d. paliiarktischen A es 7? 1896, p. 51-53.

No. 570] SPECIES-BUILDING 329

It has long been known that these two species hybridize

in the Mississippi Valley, where both occur. By extended

experiments during the past summer and previous au-

tumn with eurytheme stock sent to me from Arizona

through the kindness of Messrs. V. L. Wildermuth and

R. N. Wilson and with philodice from New Hampshire, I

have found that the two species mate together readily, and

produce vigorous offspring. The species-hybrid males

were then mated with ewrytheme females, and more than

half of the pairs (viz., four out of seven) were fertile.

Mated together, however, the species-hybrids showed

much sterility. Out of ten such matings, nine were in-

fertile. From the tenth pair, nineteen adult butterflies

were produced.

Orange in this cross is distinctly dominant over no

orange, or yellow, but the color of the heterozygote is a

pale orange overlying yellow, and is by no means as bril-

liant as the almost fiery orange of the large, summer

seasonal variety, the typical ‘‘ ewrytheme.’’ In broods

emerging the last week in August and the first three

weeks of September, when intense color may be expected,

the heterozygote is pale orange, corresponding approxi-

mately to the variety known as keewadin, whereas those

raised in the greenhouse and emerging early in December,

resemble the small orange-yellow winter type known as

ariadne. Keewaydin, according to Wright,’ occurs at all

_ Seasons in California, though probably more abundantly

in spring and autumn. Hence he regards this as the

typical variety, rather than ‘‘eurytheme.’’ It is inter-

mediate, however, in size and intensity of color.

In general, therefore, there is an incomplete dominance

of orange, the color of the heterozygote corresponding

either to that of the intermediate or to that of the winter,

seasonal variety of euwrytheme, depending upon the time of

the year when, and the environmental condition under

which, the cross is made. The wide, spotted margin of the

hind wing in the female euwrytheme, moreover, when pres-

8 ‘‘ Butterflies of the West Coast of the United States,’’ p. 119.

330 THE AMERICAN NATURALIST [Vou. XLVIII

ent in marked degree, is dominant over the narrower

margin in philodice. This dominance of the orange mani-

fests itself quite as distinctly if the albino female of

eurytheme, instead of the orange female, is bred to the

yellow philodice male. The daughters of such a family

in one case (0, 1913) were 36 white, 35 orange; the sons,

numbering 72, were, of course, all orange. The white

species-hybrid (F) is identical in appearance with the

albino eurytheme, the female color pattern of the latter

(wide marginal bands) being dominant, and the orange

middle spot both in pure bred albino ewrytheme and in

the albino hybrid being usually paler than i in their orange

sisters.

The second hybrid generation inbred (F,) shows a well

marked segregation of the sulphur-yellow color of philo-

dice, as a simple Mendelian recessive. Three out of the

sixteen colored (non-albino) individuals of the brood ob-

tained in December, 1913, are definite recessives of clear

sulphur yellow, with pale yellow middle spots on the hind

wing. The most highly colored individuals are four that

correspond in hue to pale examples of the light orange-

yellow winter variety, ariadne. There is no return, at

least in this winter brood (enclosed in a greenhouse in

New Hampshire in December), to the brilliant orange of

the grandparental eurytheme stock. Nor do they even

return to the suffused light orange (intermediate) tint

of the heterozygous father (keewadin type), for the

ground color of all individuals of this brood (F,) is

yellow, either finshed or spotted, except in three indi-

viduals, with orange. -

An interesting case of probable hybridization in the

allied genus Meganostoma, or dog’s head butterfly, is re-

corded by Wright? between the Californian M. eurydice

and M. cesonia, common throughout the southern states.

The two species are remarkably different in color and

have different food plants. The male of eurydice differs

from that of cesonia in having a violet luster and lacking

9 Loc. cit., p. 116.

No. 570] SPECIES-BUILDING i 331

the black border upon the hind wings possessed by

cesonia; in the female, eurydice is clear yellow with no

dark border, while in cesonia the female has a wide

border similar to that of its male, though less well marked

on the hind wings. The probable hybrid called amorphe

is a female, intermediate in color between the typical

cesonia and eurydice. That is, the border of cesonia

crossed with no border (if my interpretation is correct)

is incompletely dominant. Wright says:

At one time I was of the opinion that Amorphe was a hybrid between

Eurytheme and Caesonia ... but of late years, as no male Amorphe

is known, I have concluded that Amorphe is simply a dimorphic female

[of eurydice].

Possibly it is both, an example of dimorphism produced

either by immediate hybridization, or by a mutation re-

sulting from some previous hybridization. That a male

appears to be lacking in this case would not be an argu-

ment against the possibility of hybridization, for by such

crossing the sex ratio is frequently upset, the product

being of one sex only. But it appears to be possible that

the male of this cross is that described as M. bernardino,

a variety of ewrydice found in the mountains of the same

region where amorphe also occurs. It is an interesting |

combination of the male coloration of both species, having

the violet hue of ewrydice that is lacking in cesonia and

having the dark border of the hind wings of cesonia lack-

ing in eurydice. Its female is described as being smaller

than that of eurydice, but otherwise practically identical

with it. This case, as Wright has suggested, is a most

inviting subject for further study, and, judging by me

he says of the'sexual instincts of the eurydice male—“ a

wooer .. . energetic and persistent, not hesitating to

ignore all rules of propriety, of species and of genera ”’

—not difficult of experimental management.

The genus Basilarchia, the admiral butterflies, is well

known for the hybridization of its very unlike species,

B. arthemis the “ banded purple ’’ of the northern states,

332 THE AMERICAN NATURALIST (Vou. XLVII

B. astyanax the ‘‘ red-spotted purple ’’ of the southeast-

ern states. The hybrid species, B. proserpina, occurs

in a zone in which their two faunal areas overlap.

In this same group is the common ‘‘ viceroy ’’ B. archip-

pus, the range of which roughly covers that of both the

other species and extends further westward, touching the

Pacific coast in Washington (Scudder). The experiments

of Edwards, and especially of Field, have shown that

these three well-differentiated pure species occupying

contiguous, or in respect to archippus overlapping,

territory are in some cases at least mutually fertile.

B. arthemis and astyanax regularly interbreed in the

narrow zone where proserpina occurs. Proserpina, the

hybrid, usually shows the general dominance of the

astyanax characters (lack of white band).

From eggs laid by a wild female proserpina Edwards!’

secured three arthemis, one proserpina. Field‘! raised

from a similar lot of eggs nine proserpina, seven arthe-

mis. Presumably in each case the male parent was the

recessive arthemis, and hence equal numbers of the two

types would be expected. Field has also succeeded in

crossing a 2 astyanax with a g£ arthemis, and a @ viceroy,

archippus, with a g arthemis, the latter pair producing

nine males intermediate in color. Specimens of an ap-

parent hybrid, intermediate in color between astyanax

and archippus, have also occasionally been captured.

The complete overlapping of the faunal area of archip-

pus upon those of the two other species indicates that,’

though crossing sometimes occurs, the resulting hybrids

are probably usually sterile, though this matter has not

yet been thoroughly investigated. Proserpina, however,

is a fertile and extraordinarily variable hybrid. In view

of its great variability it appears, by the way, not impos-

sible that archippus, the red-brown ‘‘ mimic ”’ of the mon-

arch, Anosia plexippus, may have arisen as a mutation

from the hybrid proserpina, though the wide-spread

10 Canadian Entomologist, Vol. IX, 1877.

11 Psyche, Vol. XVII, No. 3, 1910.

No. 570] SPECIES-BUILDING 333

range of archippus at present and our ignorance of the

state of the Basilarchia stock at the time of the origin of

the ‘‘mimic’’ make any such specific historical guess

hazardous. It may, however, some time be possible by ex-

perimental breeding to extract from this red-spotted pur-

ple hybrid a red-brown type similar to archippus. If the

Basilarchia stock were as easily bred as Drosophila, one

might be very confident of accomplishing this. In any

event, the theory of the origin of mimicry by natural

selection is, in the opinion of the writer, entirely super-

fluous, though this celebrated monarch-viceroy case

should be exhaustively studied by experimental methods,

to determine whether natural selection now operates in

any degree in the matter. .

Examples of clusters of interbreeding types may be

drawn in large numbers from various classes of animals

and plants. Bateson?? has recently called attention to the

interesting case of the two American flickers described

by Allen, the eastern Colaptes auratus and the western

and Mexican C. cafer, which hybridize in the zone in

which their faunal areas overlap, the American grackles,

the golden-winged and blue-winged warblers and their

hybrids, Lawrence’s and Brewster’s warblers, and others.

In reference to the common purple grackle, which

Chapman regards as a hybrid between the Florida

grackle and the bronzed grackle, Ridgeway'® says:

My own opinion in the matter exactly coincides with Mr. Chapman’s

but since so many forms now ranked as sub-species are similarly in-

volved I prefer, at present, to leave the matter in abeyance.

This significant statement from a master of ornithologi-

cal taxonomy indicates that hybridization among Ameri-

can birds is a promising subject for investigation.

Of the occasional mutual fertility of unlike strains dif-

ferent enough to be classed as unquestionable species.

12‘< Problems of Geneties,’? 1913, Chap. VII.

~ Bull. American Mus. Nat. Hist., Vol. IV, 1892.

14 Ibid.

15 “* Birds of North and Middle America,’’ Part 2, p. 219, 1902.

334 THE AMERICAN NATURALIST [Vou. XLVIII

there also can be no doubt. ‘‘ We can only escape the

conclusion that some species are fully fertile when

crossed,’’ wrote Darwin,'® ‘‘by determining to designate

as varieties all the forms that are quite fertile,’’ and he

added that some plants exposed to unnatural conditions

are so modified ‘‘ that they are much more fertile when

crossed by a distinct species than when fertilized by their

own pollen. |

The rareness of these crosses between unlike strains or

species and the partial sterility of the offspring are not

obstacles in the way of regarding occasional hybridiza-

tion as one of the chief sources of mutation and hence

eventually of new species, for, as my preliminary experi-

ments in hybridizing species of Colias have already

shown, there may exist within a strain of species-hybrids

certain individuals that are fertile, though the most of

their brothers and sisters, mated, respectively, in a similar

way, are sterile. Nature probably makes more random

experiments in hybridization than we imagine; many fail;

some succeed; and in especially favorable stock like

Colias, judging from the numbers of closely allied but

different types (species) occurring in the same localities

in western Asia or in northwestern United States and

British America, probably many succeed.

In seeking to determine how mutation, whether the re-

sult of hybridization or of possible climatic influences,

- acts in the production of new species, it is possible from

cases already at hand to suggest possible steps in the evo-

lution of distinct, mutually infertile, types from one com-

paratively simple polymorphic species.

The well-known dimorphic European currant moth,

Abraxas grossulariata, in which the light-colored (reces-

sive) variety, lacticolor, is found in nature only in the

female sex, will serve as an example of an elementary

condition. Lacticolor males, as Doncaster’? has shown,

16‘ Animals and Plants under Domestication,’’ Vol. II, Chap. 19, p. 179.

17 t‘ Report of the Evolution Committee,’’ 4, 1908.

No. 570] SPECIES-BUILDING 335

may readily be bred. When one of these males is mated

with a lacticolor female, there is produced in captivity a

pure lacticolor strain. If lacticolor males and females

should be segregated and allowed to breed together until

they have become as abundant as the typical form, this

case would then resemble that of the Colorado lady beetles

of the genus Adalia, described above, in that it would con-

sist of different types maintaining their identity while

freely interbreeding with complete fertility. The

Abraxas complex differs from the Adalia species-cluster,

however, in the occurrence of sex-linkage in the inheri-

tance of the lacticolor variety, whereas in Adalia the

factors for the different color patterns apparently are

distributed in the gametogenesis of a heterozygous indi-

vidual without sex-linkage, freely and at random.

A more advanced stage in evolution is that represented

by the Basilarchia species-cluster, in which partial steril-

ity between the viceroy and the two purple species, over

the faunal areas of which its own overlaps, and the differ-

ence in geographical distribution between the banded

purple and red-spotted purple, keep the three elements

apart.

By easy stages we may in imagination pass on to

groups composed of closely allied species which sterility

and local segregation completely separate from one

another, groups that probably have arisen from a poly-

morphic species that has broken up into its constituent

parts, and thus given rise to new elementary species.

The dimorphism of Colias differs from that of Abraxas

in that the color of the rarer type of female can not be

transferred in the ordinary course of breeding, without

further mutation, to the male. It is a sex-limited char-

acter, like the female color pattern in Colias, (i. e., a wide

dark border broken with spots) and not sex-linked like

the variety lacticolor of Abraxas.

The white female of Colias is regularly heterozygous

for color. She produces as many white daughters as

336 THE AMERICAN NATURALIST [Vou. XLVIII

yellow, or orange, as the case may be. Evidently, in order

to extract a pure white race from C. philodice or C. eury-

theme, it will be necessary by a mutation to obtain first a

homozygous white female, and then by a further mutation

a homozygous white male. White males are known in

nature as rare aberrations, but, whether they are homo-

zygous or heterozygous for color, it is impossible to say.

Among the two thousand offspring of heterozygous white

females of philodice and eurytheme that I have bred

since 1908, there has been not a single white male. The

sons of a white female, though some are capable of trans-

mitting the white, are always yellow or orange. I have

lately, however, raised a large brood in which all the

females were white. This was a ‘‘ back cross ’’ between

a white female of the orange ewrytheme and a male spe-

cies-hybrid (son of a white mother). Precisely similar

matings, however, gave both white and colored female

offspring in equal numbers; hence in the production of

this brood there was probably a mutation. From such

stock as this the extraction of a pure white race from

Colias at some time may possibly be accomplished.

In this connection it is interesting to note that we have

the testimony of a good observer, the late Mr. W. G.

Wright," who made the study of Californian butterflies

his life work, to the effect that the white variety of Colias

eurytheme ‘‘is now quite common, though twenty-five

years ago it was a great rarity, and it was accounted a

feat to secure one of them, and if the present rate of in-

crease of the blond form shall go on, in a few hundred

years the normal orange-colored female will be extinct

and unknown.’’ If this is a fact, and not an illusion due

to a general increase in the population of eurytheme

owing to an increase in the cultivation of the food plant,

alfalfa, in that region, it may be the result of possible

mutations, whereby homozygous white females may have

been introduced into the population. It will be of inter-

17 Loe. oit., p. 117.

No. 570] SPECIES-BUILDING 337

est to determine whether such true-breeding white

females actually occur in California.

Evolution in Colias is usually regarded, on the other

hand, as tending towards suppression of the white stock

rather than its further extension, inasmuch as Pieris and

other allied genera are white. It seems to be a reason-

able hypothesis that, by progressive mutations in Colias

affecting first the male then the female,1’ white has be-

come yellow; yellow, orange; orange, red, or a fiery

orange ;'° or yellow may be transmuted into black, as in

an aberration of the male in C. philodice. By retrogres-

sive or degressive mutations, accordingly, we may hope

to isolate from C. philodice or C. eurytheme a pure white

race.

SUMMARY AND CONCLUSIONS

The erroneous idea that Linnæan species are homo-

geneous, well-defined groups of equal importance has

done much to retard progress in the experimental study

of evolution. The limits of a species are often arbitrary,

depending ultimately upon the temperament of the des-

criber, and frequently based upon ignorance of the near-

est allies of the individuals described, living in other

parts of the world.

The most definite criteria of species, viz., that ‘‘ spe-

cific? characters are constant, and that hybrids of

Linnæan species are infertile inter se, are only approxi-

mately correct. Characteristics of species sometimes

occur in heterozygous condition. Hybrids of Linnæan

species, as has long been known, are often fertile. These

matters, owing to traditional, unwarranted respect for

described species, have received comparatively little in-

vestigation.

Examples of hybridization in Adalia, Colias, Meganos-

toma, Basilarchia and Paratettix among insects, in Cv-

laptes, Quiscalus, and Helminthophila among birds are

cited.

18 In C. Rimes of South America, for example, the female is yellow, but

in the male the fore wings are orange.

19 As in the Asiatic eogene.

338 THE AMERICAN NATURALIST [Vou. XLVIII

Occasional fertile crossing of unlike strains that rarely

interbreed is a probable source of mutations and new

types.

A suggestion is made that a comparatively simple poly-

morphic species (like Abraxas grossulariata) may break

up into a cluster of mutually fertile elementary species

(e. g., Adalia in Colorado). Further differentiation, in-

volving partial sterility, may be illustrated by the Basi-

larchia species-cluster. This may be followed by the

establishment, and isolation through complete sterility,

of distinct types, or species in the strict sense of the term.

Evolution of color in the yellow and orange butterflies

of the genus Colias involves white, which exists to-day in

heterozygous condition in certain females. If the an-

cestors of Colias were white, as in Pierids generally, we

have only to imagine a mutation in the male-producing

germ cells of the original white females, by virtue of

which white pigment was replaced by, or transmuted into,

yellow. This would make all the males yellow, leaving

all the females white, which is true of certain arctic

species to-day.

A similar mutation affecting the germ cells of these

white females, but introducing the factor for yellow into

only half of them, would produce the heterozygous condi-

tion found in C. philodice and C. eurytheme. Pure yel-

low strains may readily be bred from such mixed stock,

and hence, probably, it has come about that four fifths or

nine tenths of the females of C. philodice in eastern

United States are pure yellow.

Progressive mutations from yellow to orange and fiery

orange, affecting first the male, then the female, have

probably occurred in Colias in many part of the world,

especially in warmer climates. Climatic conditions deter-

mine the amount of orange pigment in the cross between

the orange eurytheme and the yellow philodice. This

hybrid is larger and contains more orange when raised in

summer than when bred in late fall and winter. C. philo-

dice in this cross is a Mendelian recessive.

HEREDITY OF BRISTLES IN THE COMMON

GREENBOTTLE FLY, LUCILIA SERICATA

MEIG. A STUDY OF FACTORS GOVERN-

ING DISTRIBUTION?

PHINEAS W. WHITING

Bussey INSTITUTION

In a previous paper? I have given data showing that

variation in the number of posterior dorso-central and

acrostichal bristles of the common greenbottle fly, Lucilia

sericata Meig., is determined by hereditary factors.

Since the publication of that paper further evidence,

bearing upon the nature of the hereditary factors in-

volved, has been obtained.

Two general conclusions from the work may be stated

as follows:

1. Reduction in bristles tends to affect the males more

than the females, while additional bristles are found more

often in the females.

2. Distribution as well as number of bristles is heredi-

tary.

On account of very high mortality in these flies it has

been impossible to make selections as might seem desir-

able. The results, however, furnish considerable evi-

dence for the foregoing conclusions, and throw light, I

believe, on the nature of factors governing distribution,

Such as spotting factors, for example.

Fig. 1 shows the mesonotum of Lucilia sericata with

chetotaxy normal. The bristles considered in my work

are those lettered A, B, C, the post-acrostichals, and 4’,

B’, C’, the post-dorso-centrals.

1 From the Entomological Laboratory of the Bussey Institution, Harvard

ict todd No. 77.

2 Whiting, P. W., ‘‘Observations on the Chetotaxy of Calliphorine,’’

Annals of the Entomological Society of America, VI, 2

339

340 THE AMERICAN NATURALIST [Vou. XLVIII

It is evident that these bristles form a group of twelve

in four rows of three each.

This arrangement is recorded as 3, 3, 3, 3, the separation into rows

being denoted by commas.

When one or two of the anterior bristles of a row are omitted, the row

is denoted by 2 or 1, respectively.

In order to denote the omission of the second or third bristle when

those anterior to it are not omitted, the normal positions of the bristles

are recorded as a, b, ec, from anterior to posterior. Thus a row lacking

the second bristle would be called ac.

Addition of a supernumerary bristle into a row is denoted by ! in-

serted in the proper position between or in front of the letters denoting

the normal bristles. Thus addition of a bristle in front of a row wo d

be expressed by ealling the row !abe.

Insertion of a supernumerary bristle between the normal rows is

denoted by parentheses enclosing a, b, or c, acording to the position

of the bristle from anterior to posterior. Thus a definition as 3, (a),

3, 3, 3, would denote the addition of a bristle between the first left post-

~- dorso-central and the first left post-acrostichal.

Additional bristles are usually smaller than the normal, but range all

the way from microchaetæ to the size of the normal macrochete.

small bristle is denoted by italics.

The progeny of a few wild females have been bred and

counted since my previous paper.* These have been

FACTORS GOVERNING DISTRIBUTION 341

No. 570]

‘gg ‘SoyeuIes ur Suryoul sopyjstaq Jo r9quinyy

"Ler, ‘sopeurey 1840

‘S'6G Sorem ur Suryoul SSQ Jo roqunNy “BPL‘T ‘SOE T840,

TET oli (| 801) Fe) 6 | ote | ezz | 1% | 9% | st | s9 | 980'T| G90'T| e BOB Og To STRIOL,

g ? 0 t |0|T |e |9t) 9 | got jose | & | @ L oe de ae ee "soe + @uppoey

| | Pap g pue ‘puosss g ‘sy 4)

| | | | seyonsoIoe CT

| | | Suppovy sory OT

cS 91 0 0 1188 |8 £ SI | 6LET| 9I Er 9 | Ig 329 | T9 eoo o o eS

è |e |è |e jelelsiels| 2 ajele e] e]?

pe | pe | 48 fue

SOT §[¥1}190-0S10q “Bold *pəsvərvur + “pasveroaq yemsz0N

u posvoesouy, UF segonsoaoy 10L SIIMON

PEPPY SINSHA ;

SOTA c pasvesoeg ,, U} Supyow'y SISH | SISH SULAMOYS SƏFA JO loquny

UMN[OD }SIPI UT PIPIOIIH SƏMA WOIJ AudZ01g

‘g'o JO Əsveərəəp e se pəzunoə sT Əsrq pəzıs-jews Y “OT SB

pojuNod sr Əysuq B JO suəsqe Tezo} uoronpər 0} Surpel Or S}UNOD ƏY} UT “porqut jou əsuəy ‘rəy pəpnpur Ájuo suorerouoS qsari

CIM NAV], STIVWAT WOUA ATUQ SAITA AO ayoowy

I Wiavi

342 THE AMERICAN NATURALIST [Vou. XLVIII

averaged with those recorded previously and the results

given in Table I.

From this table it appears that progeny of normal

mothers show a certain degree of variation in the direc-

tion both of loss and of acquisition of bristles; progeny

of reduced mothers tend more toward reduction; and

progeny of mothers bearing additional bristles tend more

toward the addition of bristles. It is also evident that in

3 It is thought desirable to da on record a detailed account of these fam-

ne as they furnish in themselves a few points of interest. This pgp" is

given below with _ exertion of the progeny of 1913-A, discussed in

Sie part of this

1913-B, L. EFU oak ab, ab, 3, taken at Bussey Institution, May 6,

1913, gave

11 gg = 3, 3, 3, 3

1 g=3, 38, 3, 2.

3668, 3 3: ¥: 17 99==8, 8, 8, 3.

= 3, 3, ae ==

1 ¢ ? , 3. l ; 2, 3, 3.

1913-C, L. sericata 9 = 3, 3, 2, 3, taken at Bussey Institution, May 6,

1913, gave

49 j= 3, 3, 3, 3. 57 99 =3, 3, 3, 3.

1 g=3, 3, abe, 3. i os labe, abe 3, 3.

do =83, albe, albe, 3. 1 9==38, a/be, i

= !

case I attribute the additional bristles to the combination of fac-

tors introduced by the male. An example of this sort in which a reduced

ur e produces a abnormal predominantly by addition is very un-

ual. There are, however, occasionally flies with extra bristles in reduced

a, a fact which may be explained by recombinations of factors or by

mutation.

1913-F, L. sericata 9? = 3, 3, 3, 3, taken at Bussey Institution, March 19,

1913, gave

24 dd = 3, 3, 3, 3. 1999 =3,'3, 3, 3.

from a mating of these were produced

92 df = 3, 3, 3, 3. 3399 =>

1 g=3, 3, albe, 3. 1 sae a ; 3.

1913-D, L. cæsar 9 = 3, 2, 2, 3 (the rapiodgad normal for pa species),

taken at Bussey Institution, May 5, 1913 gav

55 dg = 3, 2, 2, 3. usai 2, 2, 3

4 gg = 3, 1, 2, 3. 1 Q= ac, 2, 2, 3.

1 g==3, 2,b

2433 =3, 1,13

1 d=3, 2 1

The flies of this mating are not averaged with the others, as it is possible

that this species may be different in its variability from L. sericata. It is

noteworthy, however, that here also reduction favors the male more than

the female. '

No. 570] FACTORS GOVERNING DISTRIBUTION 343

general reduction tends considerably to favor the males,

while addition favors the females to a slight extent.

In my previous paper (p. 264) is given in detail a

record of the progeny of a female L. sericata (1912-c)

lacking both of the first and the right second post-acros-

tichal (3, 2, 1, 3). These were inbred to the third gen-

eration, in all cases brother being mated with sister in an

attempt to analyze the stock as thoroughly as possible

and to reduce heterozygosis of factors. Here again, due

4 Mr. Harold D. Fish has kindly furnished me the following note:

‘‘The importance of mating sisters with brothers for a long series of

generations in the experiments aimed to detect Mendelizing units of inherit-

ance and analyze groups of them, quite generally seems to have been over-

looked. As first shown by Castle (’03), random mating of the individuals

of successive generations beyond F, tends to produce in each generation a

population with the sam e per ag of E and heterozygosis as is

gous for one faktor of a beak tats pair, 25 per cent. homozygous

for the other factor, and 50 per cent. heterozygous for both. Such a system

of random matings often has been confused with the more restricted. system

of motini sisters with brothers.

AE ip peny that if A and B are an allelomorphic pair the F, zygotes,

resulting from a mating of AA with BB, will be

Further, if pas are all females and are mated in all possible ways with the

same number and kinds of males, one sixteenth of the matings will be 4A

with AA, and one sixteenth will be BB with BB. One eighth of the TAA

then, will be homozygous and produce only homozygous young, which, be-

cause of the restricted system of mating only sisters with brothers, ay pro-

duce, in turn, only homozygous matings. The remaining m

This would mean that the reoht ai of heterozygous matings between in äi-

viduals of the F, ge eneration would be (7/8)"*. “Accordingly one would ex-

pect an automatic increase in y a The expectation is justified al-

Though the figures are misleadin

De. Raymond Pearl first published the figures exactly expressing the

344 THE AMERICAN NATURALIST [Vou. XLVIII

to high mortality, selection as might have been desired

has been impossible.

A detailed account of this strain is given in Table II.

In recording any mating of this strain the letter c denot-

ing the entire strain, is followed by F,, F., ete., denoting

the generation from which the mated flies hive been

chosen. This symbol is then followed by a, b, or c, denot-

ing the first, second, or third mating, respectively, of the

generation indicated. Thus mating cF,b is the second

mating of the second inbred generation of strain 1912—c.

This method of recording matings has been followed

throughout my work.

Several points of interest are to be noted in this strain

but it is thought best to present the remaining data on

reduced strains before proceeding to a discussion of this

matter.

Strict inbreeding has been followed in the strain re-.

corded below. In no case have there been either cousin-

matings or outcrossings.

1913-4, L. sericata 9 =3, ac, ac, 3, taken at Bussey Institution, Forest

Hills, Mass., May 6, 1913, gave

96 JJ = 3, 3, 3, 3. Bers we 3, 3, 3.

1 gg =3, ac, ac, 3 1 Yo 3, ac, 3, 3.

2 dg = 3, ae, 3, 3. S99==5, 2,3; pi

2 dd = 3, 3, ac, 3

1 an 2, ac, 3

4 $$ =3, 2, 3, 3.

pe am, 8. 3,8,

as the per cent. of the allelomorphice factors which are homozygous in the

average individual of that generation. Because Dr. Pearl in his October

met referred rae to a paper by Dr. E. M. East ( ie, on ‘‘ Hetero-

zygosis in Evolution and Plant Breeding.’’ I gave Dr. East my figure ex-

Bele pa per pated of homozygosis in successive coma resulting

from matings of sisters with brothers. Dr. Pearl’s correction followed a

matics of random matings in each generation to a case where sisters always

had been mated with brothers. The percentages, as computed, were pub-

lished by Dr. Pearl for the following generations: P,—100 per cent., F,—0

per cent., F,—50 per cent., F,—50 per cent., F,—62.5 per cent., F,—68.25

per cent., F,—75 per cent., F;—79.687 per cent., F,—83.594 per cent, Fy—

86.719 per cent., F~—89,258 per cent. Previous to giving these figures to

Dr. East I computed the number of generations megri to reduce heterozy-

gosis to less than one half of one per cent. and found this condition first

realized in the F., generation, which is 99.553 or pe: — e The

importance of these figures in work of this nature is quite obvious.’

FACTORS GOVERNING DISTRIBUTION 345

No. 570]

ot |_ 81 Ot E 1 Oy AF 2] Sty O18 4: TE 1 OL | St 7 Bt) oe Fale Sea Tats -oL

Lá SI 0 O01107 T0040 £61 t Il I 0 eg | CRN Ssh flags ama [BulIOU= Q79 Wor] ty

(pg) &% OPO aie) Ost HSE 4 a 9 9 OLT £91 Stars eet ge So[BUley ewou g

(PZ) S'e : X eyeur puou

8 G. IOT F 0 = DJI UIO} ty

0 T De t Prr Lodra zet | o I 6 Mo eer a et aed ewou = 91,79 WOI t

0 j 0 \PaZslo;/oleloloilr | 28 0 Z g Z CG. We “+ aud [wuLLOU= qiyo woa ty

OO 0 0;0;/0/0/0;oj0/] at | o 0 0 0 L oo on aed Buou = D II WOY t

0 0 0 001010000 | g 0 0 0 0 a 5 Gee S Ss EIZ ‘g) = 19y}owu wory y

é 2 é 2 ô| ejó] jól] é 2 é © fe] ©

pe pe] at {uo Ta

SOTTT S[@1jU90-osaoqy |— Soe OCS “501g |, pesvesouy pasear q [vULioN aia

<< POSBOIOUT,, UL s[ByoIjso10y IL NEW

Aenea pas i oa A E Sa ta S dee a: ; JO 19vIvyp pur suopesauan

sera « PeseaI00q » UL Jurgowa səapsug səpsug She ag SƏFA jo po adan

UUIN[OD ISI UL paps0deyy vy sSurEyy moiy AueSo01g

GIGI ‘8 ISAÐNV ‘Nossog ‘moog apVaUVvy THL LY NAAVL ‘(g ‘T ‘g “¢) & 01021438 “7 IZIGI NIVULS A0 axoomy

Il Widvi

346 THE AMERICAN NATURALIST [Vou. XLVIII

from AF,a= ioc ac, 3, 3.

== 5, 2, 8, 8:

g 1,

3 Sg = 3, ac, 2, 3

5 dd ama, 2, ac, 3

Log 8, 2,2, 3

5 dg =3, 2, 3, 3

238 8, 2, 3

BRED RED HH OD Ww

+0 +C+0+0

+0+0+0+0+0+0

Gs ea oa GS OD

+0+0+0+0+0+0

Hee

MS Sl! RS ORR MR SOE a a R e

.- ~ Sa `~

Pair segregated July 12; larve

July 25

99 ae S

6 a5 == 3, 3; 8, 3 7 2=3, 2, ae, 3

42 == 8; ót; BC, 10 4223, 2, 2,3.

9 18 == 3, ne, 8,5 7 728, 8,8, 3

16 1 3, 3, ac, 3 5 19, 8,2, 8

5 2 =ne ae, 2,9 1 O== 3; ac, 873

from AF,.b =

d = 3, ae, 2, 3 j

Q =3, abe, 2, 3.

Pair segregated July 12; larvae July 25.

3c. n dd

16 45 = 3, 3, 3, 4 1==8, 3, 2,

34 9 = 3, ac, ac, 3 0 1==38, abe, ac, 3

8 0==3, ac, 5, 3. 0 $==3, 3, abe, 3

7 Li 88): dey 8. 0 3=3, abe, abe, 3

8 == Ot, oy Os 1 0==3, b, 3, 3.

9 i= 3, 2) At, 3. 1 0=3, ać, abe, 3. ‘

3 OSS. Fy p B. 0 1==8, abe, 3, 3.

2 1==8, 2, 3, 3.

= g=3 3, from AB. P 14;

Ag == 8, a6, 8, rom AF,a. ps) ee August 14;

from APa = 19 = 3, ac, ac, 3, larve August

do D 33 e

6 32 = 3, 3, 3, 3 4 T= 9, 2, 3; 3.

24 11 ==8, ac, ae, 3 1 p= B, 3, 2, 3.

4 20== 3, ae, 3, 3. 1 0=8, ac, 1, 3.

6 $==38, 8, ac, 3. 0 T= 8, ae, 3, 9.

9 "6-29, ae, 2, 3: 1 0= 3, ac, bee, 3

6 3, 2, ac, 3. 0 1228, 2, 2, 8,

3 =n. 2, 2, Sy 0 1= 3, 3, abb

from AF,b = and ie 2, ac, 3, from AF,a. Pair segregated August

13; larve August 2

dd Go n

0 18 =3, 3, 3, 3 13 2==5,'2,' ae, 3

10 == 3, ac, ac, 3 2==5, 3, 3.3.

4 5=23, ac, 8, 3 1 72B, 2,8, 3.

2 == 3, 3, ae, 1 4==3, 8, 3,3

from AF = g — Eea 2, 2, 3, from AF,a. Pair segregated August

13; larvæ August

No. 570] FACTORS GOVERNING DISTRIBUTION 347

Go. oy do fs

6 15 = 3, 3, 3, 3. 2 15 = 3, 2, 3, 3

10 7 = 3, ac, ac, 3 3 8 = 3, 3, 2, 3

1 7 =3, ae, 3, 3 0 1 = ac; h 2, ac

10 12==3, 3, ae, 1 0=83, 2, ace,

8 3 == 3, ac, 2, 2 1==3, ace, ac, 3

t fr as a 1 0= 3, ace, 3, 3

5 4=

from AF d= Jg aka yes c, ac, 3, from AF,a, Pair segregated August

13; nels died August 18 mes another with same chetotaxy put in; larve

Septem ber

22 od. T7

0 == 3, 3, 3, 3. 1 be= S, ac; 2, 3.

1 2==3, ae, 3, 3. 1 0 = 3, 3, aabe, 3.

The tá it 1018 2. eooriei in tabular form is given in Table III.

We are now in a position to consider the nature of re-

duction of bristles in Lucilia sericata.

It is evident from Table I (record of first generation

flies), that reduction and addition of bristles are both

hereditary. It is further evident from Table ITI, (inbred

strain), that reduction yields readily to selection. This

effect may be expressed by making the number of bristles’

lost the numerator of a fraction of which the denominator

is the number of bristles normal. We then have a ratio

for each generation of 1913—A as follows:

OES

fı. 3892 7 0.006 = .010, F; - fen = 0.093 + .003,

99 532.5

Fy - 1788 ~ 0.055 = .004, Fi: 5100 > ure + .003.

It may be readily seen by glancing at these figures that

selection has a very rapid effect. It also appears that as

we pass from F, to F, the effect of selection gradually

diminishes. This may be expressed by dividing the above

decimals for each generation by that of the preceding

generation.

ba thant Pe a a o «= 1,69,

K oom >e aOR

Fy _ 0.104 _

es 0.093 11k,

The reason for this decrease in the effect of selection in

the later generations is that as the selection advances the

majority of the flies become reduced in two bristles only.

[ Vou. XLVIII

THE AMERICAN NATURALIST

348

€ L | € |0|0)@|9892| Zh \seT 16r srj € | 9 T | T | 008| 11% hoa Oe i ae tt SOL

= | ea

0 o tok do folates 9 te Tae gee olo lolin eee ee ogy woa

| T i ‘g ‘on ‘oe ‘g= & pue (0

| | = pty wor tI

0 I y OE toloto 9 Z9 IEF 18.| OFT if $ 0 | 0 Pao IST 9 o May MoI

| ‘e ‘S's ‘€= 6 pue p

=98 V Worf rT

sees e es eope ay WOJ

0 0 0 0 41 0-} 04.0 : St 19 (0% OF | 6 Dto 0 | O |83 | 6% ST 0 | ‘98 ‘S‘g= § pue p

| | ER =qay mol rg

Z 0 T |Q@8T) E E 969 94 J98 |88 | EZT Ea 0 | 0 Ja 89 gujo "oyy utoj

| | ‘g ‘ow ‘ov ‘g= b

| | 5 tg ‘oR ‘om 2

| | =p y Wor +

ZI 0 0 0 |0/0|T {Sr | STOTT |e [08 | SZI 0 0 0 | 0 0P 22 SF 9T] L g ‘9qQ0 ‘g= 8

tg tz ‘on ‘g= yp

= y y Woy My

0 0 0 0 LO1OPTE?: of EZI 19S [PE | STZ 0 0 0 | 0 (04 OL Se 9 |g ‘ov ‘ou‘g= 6 pue,

= OY Wor tT

8 Z I 0 |0/0|0/ST-} 9°8S [2 0r RO I I T | O |9T | 9 [Eo g et eee

: tø 'o tog tpa ve)

=p y Woy FT

0 0 0 6°10; 0763.7 L e 18.| 1% 0 0 0 | pe 18i Brog ette (g ‘98 ‘ow “g)

| = Joy},OUL wory Ty

ò | 2 6 olej] ej 6 2 48/2" é 2 61/216 | P1 bl eS

e z R

syo S[eI}ua9 Aue

elba sora -0si0(] peatnacane “Box d renpeesent pesveriouy | pesveroeq | [VION iene

PZ TIBMUS |. PesBetouy ,, UF 1870. Ul pasvetouy 194081840 PUB suopean

POPPY Sos ug puer pasear q sənsug

SAHA c PISBIALQ ,, UF BUPATI SONSHA ButMoysg sərf JO 1əquny

umn[oD ISAT UL pəpiooay supe wos Luafosg

SIGI ‘9 AVIV “SSY ‘STITT LSau0,7 ‘NOILALIGSNT AASSAJ AHL LV Nauvy, (E ‘ov ‘ov ‘g) $ 01048 `T ‘P-EIGT NIVULG 40 auoomy

II wiav

No. 570] FACTORS GOVERNING DISTRIBUTION 349

Rarely does a fly occur lacking more than two. In the few

cases in which three or more bristles are lacking, the

absence of the third acrostichals or of the dorso-centrals

is as frequent as the absence of first and second acros-

tichals. Why this should be is difficult to understand, as

it would be expected that both first and both second post-

acrostichals might frequently be lacking in the same fly,

especially as flies asymmetrical for the loss of these

bristles are common.

A further point of interest lies in the fact that not only

is number of bristles a hereditary matter, but their dis-

tribution is also hereditary.. Thus from Table I (first-

generation flies) we see that in general the first post-

acrostichals tend to be reduced more than the second.

This may be expressed as a fraction:

First post-acrostichals lacking n 40.5 _ 1.19.

Second post-acrostichals lacking 34

It is possible that this tendency to reduce the first post-

acrostichal more than the second is evidence of relation-

ship to L. cesar Linn., in which the absence of the former

and the presence of the latter is the normal condition.

Strain 1913—4 (Table III), however, gives

Second post-acrostichals lacking 750

Considering the reduction in the first post-acrostichals

Separately, we may express the effect of selection — as

follows:

First post-acrostichals lacking a 329 _ 0.43.

Offspring.

Parents. lst post-acros. First post-acros. lacking.

Matings. lacking. OSEA -across. normal (2 per fly).

A 0(?) a = 0.021 + .004.

AF,a 1 io 0.087 + .011.

AF.a 0 As == 0.111 £ .002.

33 — 0.093 + .010

AF,b 2 356 i

350 THE AMERICAN NATURALIST [Vou XLVIII

AF aap Ewin +

sa 0 34g 0-185 + .014.

AF,b 2 op —=— 0.316 + .023.

s 190

AF, 4 a0. .253 + .017.

From these figures it is readily seen that reduction in

the first post-acrostichals is not entirely consistent with

the direction of selection.

Let us test the same matter for reduction in the second

post-acrostichals.

Offspring.

Parents Ist post-acros, First post-acros. lacking.

Matings lacking First post-acros. normal (2 per fly)

—8_ — ES

A 2(?) 182 0.017 + .004.

Soha as +

A¥,a 1 398 0.211 + .016.

193

AF,a 4 26 == 0.453 + 016.

147

AF .b 1.5 “a6 0.413 + .018,

AF a 4 TOD = 0.420 + .179.

69-

AF,b 2 “90° = 0.363 + .023

108

AF;¢ 0 “292 = 0.370 + .019. +

In this ease also the results are not consistent with the

direction of selection, although there is better agreement

here than in the case of the first post-acrostichals. This

is probably due to the fact that the numbers are larger.

As regards the irregularities that do occur, I consider

them as evidence of recombinations of multiple factors,

insofar as they are not due to probable error.

1912—c (Table IT) is a strain that especially tends to

lack the first post-acrostichals. Thus for the entire

strain

First post-acrostichals lacking

_ Second oe lacking 11` sll

No. 570] FACTORS GOVERNING DISTRIBUTION 351

In the 137 offspring of a single mating of this strain, cF,c,

there are 23 first post-acrostichals lacking, showing that

it is due to this mating especially that the strain is so

lacking in first post-acrostichals.

It can not as yet be said that the factors governing the

first post-acrostichals are altogether independent of those

governing the second. That a certain degree of inde-

pendence obtains is evident from a comparison of the

ratio of reduction in first to reduction in second post-

acrostichals in flies in general (Table I), with the same

ratio for strain 1913—A. In the former case we have

40.5/34 or 1.19. In the latter we have 329/750, or 0.43.

In order to establish the independence of the factors un-

derlying these two tendencies it will be necessary to

obtain, either by selection from a strain showing both

tendencies or by breeding from wild stock, two strains,

one tending to lack the first while retaining the second,

and the other tending to lack the second while retaining

the first.

A point of interest in strain 1913—A is the presence of

twelve small second post-acrostichals in the progeny of

AF.,b in which the female had one of these reduced to half

size. The progeny of AF,a in which there was total ab-,

sence of these bristles showed either presence or absence

of the same but no reduced bristles. In F, however, we

have eight reduced bristles. The occurence of these

small bristles in the progeny of certain matings is taken

_ as an indication of recombinations of multiple factors,

but the numbers are too small to establish this with cer-

tainty.

A glance at the tables shows that third post-acrosti-

chals are rarely lacking. These are normally present in

all related species, while in a few,—Cynomyia mor-

tuorum, Musca domestica, Pseudopyrellia cornicina, and

others, there is normally but one post-acrostichal, and this

is always the last.

Posterior dorso-centrals are very rarely absent. Thus

in the 2,273 flies recorded in Table I only one had a single

post-dorso-central missing. Reduction in post-acrosti-

352 THE AMERICAN NATURALIST [Vou. XLVIII

chals among these is 79.5. Among the 1,206 flies of strain

1913—A there -are but three post-dorso-centrals gone.

This latter is a highly reduced strain as regards post-

acrostichals, lacking 1,081. This great reduction in

acrostichals seems not appreciably to have affected the

dorso-centrals, a fact which argues for the independence

of the factors controlling the distribution of these two

sets of bristles.

Thus for flies recorded in Table I we have

Post-acrostichals lacking _ 79.5 — 0.03

Number of Flies oe

One post-dorso-central lacking.

For flies in strain 19134 (Table III) we have

Post-acrostichals lacking TB! 0.89

Number of Flies HATO .;

Three post-dorso-centrals lacking.

Among the 3,238 flies recorded in Tables I and III only

four post-dorso-centrals are lacking, while among the 810

flies of strain 1912—¢ (Table IT) there are 13.5 lacking.

The lack of post-acrostichals in this latter strain is 37.

There are 9.5 dorso-centrals lacking in the progeny of the

‘trio, cF,a, among which there are only seven post-acros-

tichals lacking.

Thus we see that lack of post-dorso-centrals is in no

way correlated with lack of post-acrostichals, but is evi-

dently governed by distinct factors,

VaRIATION By ADDITION oF BRISTLES

A strain of Lucilia sericata, 1913—E, showed some

interesting variations chiefly in the direction of addition

of bristles. The mother was normal (3, 3, 3, 3) , taken at

the Bussey Institution, March 19, 1913. The detailed ac-

count of the strain follows:

dd Y3 :

38 43=3, 3, 3, 3.

1 0=3, 3, able, 3.

fim BV oes eS

No. 570] FACTORS GOVERNING DISTRIBUTION 353

1 ? 2 P 3.

0 = labe, 3, 3, !abe.

from EF.a= ¢ and 9? =3, 3, 3, 3.

QF dd

318 251=3, 3, 3, 4 5= 3, albe, 3,

13 51=!abe, 3, 3, labe 0 = lalbe, 3, 3, labe

1 4 = !abe, 3, 3, 3 sl = 3, abc!, abe!

6 =3, 3, 3, labe 1 0= labe, albe, albe, la

1 =3 (a), 3, 3, 3 0 1= labe, atbe,, 3, labe

5 2=3, 3, 8 (a), 8 1 0==3, able, 3, 3

3 5 = 3, 3, albe, 3 1 1=albe, 3, 3, albe.

$ 0 = 3, albe, aes 3. 0 = labe, 3, 3, !albe.

1 0=3, 3,3; 0 13, albe, 3, la

1 0=3, 8, gei 0 1 = labe, 3, 3 (a), la

1 0 = labe, albe, 3, 3 0 1= !abe, a!be, 3, la!be

1 0=3, 3, 3 (b), 3. 0 l=albe, 3, albe, albe

1 1==3, albe, albe, 3. 1 Ibe, 3, 3, la!

1 = lalbe, !a!be, falbe, lalbe. 1 Onn 3, $, 2,

0 is oo one. 1 == 3, ac, 3, 3

0 2=lalbe, 3, 3, la 1 0=8, ac, ac, 3.

‘ 0 1= tabe, b, albe, tae. 0 1-53, abe, 3, 3.

— f = 3, 3, 3, 3.

from EFa= {8T} 3 albe, 3.

Pair segregated, July 22; larvæ July aT

191 15) a 5, 1 = oe 3, 2 labe.

25 43= labe, 3, 3, labe 2 0 = 3, 13

4 = labe, 3, 3, 3. 1 (= tabe (a), 3, 3.

0 - J=88, 8, 8, labe 1 0O=8 (a), 3, 3 (a), 3.

0 1= labe, albe, 3, labe. 1 1= labe, 3, 3 (a), !abe.

1 2 = labe, 3, albe, !abe. 0 lbe, 3,

0 2 = labe, albe, albe, labe.

from EF,b = and 9 = 3, 3, 3, 3. Pair segregated August 22.

41 57=3, 3, 3, 3. 2 0=3, albe, 3, 3.

0 =3 albe, 3, labe, 0 iz laby 3, 3, tabe.

1 028, abel, 8, 3. 0 1= labe, albe, albe, labe.

A summary of this strain is given in Table IV.

The points of interest to be noted in this table are as

follows:

There are many supernumerary bristles in the flies of

this strain.

The number of bristles added in the progeny of any

mating is very variable and has no consistent relation to

the visible character of the parents.

Addition of bristles tends very much to favor the

females, reduction still affecting the males.

Despite the high ratio of bristles added, there are

THE AMERICAN NATURALIST [Vou. XLVIII

354

163 SSI S'I Lá ESP‘ I I SPT S8 I £ LGG L FAR Maneater Rasta i So Po OE STSIOL

8 € 0 0 FOL 0 0 £ £ 0 0 2g E A EAN E pres ol A ‘ared ewou = qA seme’ Je

IIT € £ 0 I8€ I 0 €¢ 9g 0 0 OOT T61 g ‘oq t ‘g ‘g= 6 }

‘eg ‘e ‘eg ‘g=

= DIY WOT tI

OLT cL ‘0 Lá 602 0 I T6 tPF $ £ LEA BE Poa ee ee Ied ewou = Dga WOI t

Z I 0 0 L¥L 0 0 T I 0 0 94 69 ee EEFC AEUR, HERA ELS nE €‘e‘e‘e=6

‘g ‘oj, qu ‘g ‘g= 9

: = DIJA WOY ey

0 i 0 0 8 0 0 0 $ 0 0 EF Be tee Se pa ot aaa or

é © é Ps é Le é Pa é 2 ô Pad

sua

: 2 pasverouy :

F i a e » Ul sh 5) es mond pus AIR hors CBS: patents ER s3urnew JO 194981840 pur uopuIouap

POPPY INSHA | -A08 SIST

səpsug Baraoyg sət q JO saquny

UUIN[OD ISI UI pepsoosy ssaywyy WoIy Audso01g

SI6L ‘GI HOUYI “SSVI ‘STINET ISTO,T ‘NOILALIGSNI AUSSOG THL LV Navy, (¢ ‘ge ‘g ‘g) & 04091423S “7 ‘NIVALE M-QIGI AO TUOTA

AI WTa&vVib

No. 570] FACTORS GOVERNING DISTRIBUTION 355

nevertheless a few flies in the strain in which bristles are

lacking.

Bristles normally present may be lacking in individuals

having additional bristles.

GENERAL SUMMARY AND CONCLUSIONS

Taking a general summation of all the bred material of

Lucilia sericata, we find that reduction affects the males

while addition affects the females. Of the 5,367 flies bred,

2,708 are males and 2,659 are females, giving practical

equality.

Reduction in the males is 748.5 bristles, while in the

females it is only 455.5 bristles. As has been noted before

the degree of reduction in the females is increased by the

later generations of strain 1913—A, by reason of the fact

that reduction rarely goes beyond the loss of two bristles

in a single fly. Thus when most of the flies of a popula-

tion become reduced to this extent it is evident that reduc-

tion in the males would be but slightly in advance of that

in the females.

There are 210 bristles added in the males, while there

are 343 added in the females. Thus addition affects the

females more than the males. These figures for bristles

added represent number of bristles, and thus no distinc-

tion is made between bristles of large and bristles of

small size.

I wish to express my appreciation for the advice and

criticism offered me in this work by Professor W. M.

Wheeler, Messrs. H. D. Fish, S. G. Wright, and C. C.

Little. |

PHYSIOLOGICAL CORRELATIONS AND CLI-

MATIC REACTIONS IN ALFALFA

BREEDING?!

GEO. F. FREEMAN

ARIZONA AGRICULTURAL EXPERIMENT STATION

Curmatic REACTIONS

To the worker who attempts to apply the recognized

laws of heredity to the actual operations of plant improve-

ment many difficulties arise which heretofore have been

largely avoided by students of pure genetics. Color and

form characters are but little affected by the immediate

ordinary environment and hence, for the sake of simplic-

ity, are usually chosen by investigators of heredity. To

the economic breeder, however, such characters are of but

little consequence except in so far as they indicate phyletic

relationships. Of greater importance to the breeder are

those differences in yield and quality which are the re-

sults of inherited, invisible, physiological powers within

the plants, whereby each variety may respond differently

in manner or degree to the same environmental stimulus.

Those hereditary units which have to do with vegetative

vigor, heat, cold and drought resistance, time of maturity,

chemical structure, reproductive strength, etc., are as yet

but little understood. This is largely due to the difficulty

of exact experiments concerning them. This difficulty is

occasioned by the complexity of the reactions of these

hereditary forces with the external environment, and also

by the direct influence of the development of one part of

the plant upon that of some other part. The plant at ma-

turity presents the resultant of its environmental reac-

tions during development. The nature of these reactions

1 Read before the American Breeders’ Association, Columbia, S. C., Jan-

uary 26, 1913.

356

No. 570] ALFALFA BREEDING 357

is determined by the structure of the vital forces within.

These differences in vital structure may or may not be

accompanied by visible morphological differences. Such

cases of correlation have been known and used in selecting

for qualities which they were thought to indicate. The

much quoted example of the supposed correlation between

the short-haired rachilla and high brewing quality in bar-

ley is a case in point. It has been found, however, that,

whereas, in one strain or race the correlation may hold, in

another, the two characters are in no way related.

Another case of similar nature is the coupling of cob and

pericarp color in certain varieties of corn and their com-

plete independence in others. Many other examples could

be adduced to show that the coupling of two characters in

a given race of plants is no indication that these same

characters are inseparably linked in all races of the same

species. These facts have greatly reduced the value for-

merly ascribed to gametic correlations in plant breeding.

Under our present knowledge, therefore, we must depend,

for the most part, upon direct experimentation, rather

than correlations, to discover the hereditary physiological

characters of the varieties with which we are working.

Any additional light, therefore, which may be had con-

cerning the nature of such characters, together with meth-

ods for the study of the behavior of the same in their rela-

tion to each other and to their physical surroundings, will

have not only a scientific value, but will also fill a distinct

practical need.

As an illustration of such a study we may now examine

the data concerning the development, yield and chemical

composition of forty-four regional varieties of alfalfa

which were grown on the Experiment Station Farm at

Phoenix, Arizona, during the season of 1910. In the case

of this plant, which occupies the ground throughout the

year and from which six or seven crops may be harvested

during the growing period, the climatic factors include a

long series of variations coincident with the changing sea-

sons. Now, since every variety consists of its own pecul-

358 THE AMERICAN NATURALIST [Vou. XLVIII

iar complex of hereditary physiological forces, each sensi-

tive in its own manner and degree to the impinging ex-

ternal stimuli, it is not surprising that the resultant (the

gross climatic reaction) should be sharply different in the

several varietal groups.

The unequal effects upon the vegetative growth of the

different varieties brought about by the climatic changes

which occurred during the course of the summer may be

exhibited by calculating the place variation in yield. This

is best shown by correlating the first with each of the fol-

lowing cuttings throughout the season. The result is a

definite curve, beginning and ending high with a strong

sag in the middle. |

TABLE I

PLACE VARIATION IN YIELD

Cuttings 1 and 2 1 and 3 land 4

Correlation. + .75+ .04 + .68 + .05 + .33+ .09 + 36 oe 09 + 58 .07.

These figures indicate the presence of some disturbing

factor which reached its maximum intensity during the

fourth and fifth cuttings, and to which certain plots were

more sensitive than others. The average period through

which the growth of these two crops extended was June 22

to August 27. The fact that these dates include the hottest

portion of the summer strongly suggests temperature as

the disturbing factor.

The mean maximum temperature, mean minimum rela-

tive humidity and the correlation between yield and water

supplied are given in the following table:

TABLE II

TEMPERATURE, RELATIVE HUMIDITY AND WaTER SUPPLY

Mean Max- Mean Mini Correlation

Cutting) Dates pagan yo meas Periods imum T Tempera “ikelative pr rae a

Supply

1 From March 23 to April 23 82.8 27.00

z From April 23 to May 23 sii +.10

3 | From May 23 to June 22 103.6 20.40 + .05 + .10

4 | From June 22 to July 23 104.8 : +-.40 + .09

5 | From July 23 to August 27 104.4 30.00 + .21+.10

6 From August 27 to October 5 102.0 95.18 ra + .10

No. 570] ALFALFA BREEDING 359

That the relative humidity had little to do with yield is

shown by the fact that the highest averages for this factor

occurred on the first and fifth cuttings which were the

highest and lowest in yield, respectively.

_ Although it was intended to give each plot approxi-

mately the same amount of water for each cutting, uneven-

ness in the slope made this impossible. The average

amount of water applied to each cutting was 6.28 inches

with an average standard deviation of 1.54 inches. Now,

taking cognizance of this variation in the water supply,

we find that its effect upon the yield was only appreciable

in the fourth and fifth cuttings. Records were not made

of the water supplied to the first cutting, but after that

time they are complete. By reference to Table II it will

be observed that these correlations in the second, third

and sixth cuttings are so small as to be negligible, but in

the fourth and fifth cuttings they are sufficiently large to

indicate that this factor was of some importance in gov-

erning the yields. These results may be interpreted as

meaning that approximately 6.28 inches of water were

ample for each cutting during the cooler weather of spring

and fall. That too much was not given at these seasons,

however, is shown by the absence of large minus correla-

tions. Factors other than water supply, therefore, gov-

erned the yields during these periods. Hot, dry weather

came on during the growth of the third cutting, but the

amount of water supplied plus the winter and spring sur-

plus left in the soil was ample to mature the crop. With

the continued high demand for water during the hot

weather of July and August, the surplus having been

exhausted and the summer rains helping but little, six and

one fourth inches was not sufficient. There was, therefore,

marked suffering for water, which was reflected in the

yields of those plots that received slightly more or less of

irrigation than the others.

It would seem, therefore, that high temperature and a

slight deficiency of water were the disturbing factors in

360 THE AMERICAN NATURALIST [Vov. XLVIII

the relative yields of the varieties tested, and that certain

ones were more sensitive than others to these influences.

If we turn to the relation between stand and yield, we

shall again find a strong disturbance of the normal corre-

lation as shown in the following:

TABLE III

CORRELATION BETWEEN STAND AND YIELD

Cutting Ist 2d 3d

Corrolntion 6366s + .78 + .04 + .55 + .07 + 47 + .08

Cutting 4th 5th i 6th

WOrrelianion ee + 54+ .07 + .10 + .10 + .70 + .05,

The exceptionally low coefficient of the fifth cutting was

due to the low yields on the part of plots which had good

stands but were relatively inactive during the hot

weather and partial water famine which occurred at this

period. On the other hand, certain plots through their

resistance to heat and consequent activity at this period,

overcame to a large extent their handicap of poor stands,

and nearly obliterated the usual plus correlation between

stand and yield.

The data thus studied en masse indicate at least two

physiological groups which are unequally sensitive to the

climatic changes which occur in the course of a growing

season, and whose reactions were sufficiently strong to

change almost completely the order of the productivity of

the plots. In order to test this conclusion let us turn to

the individual plots and endeavor to discover and classify

the physiological varieties indicated above,

If, now, we arrange the forty-four regional strains

according to their morphological characters and geo-

graphical origin, we shall have five more or less distinct

groups as follows: Mediterranean, Peruvian, European,

American and Turkestan. The behavior of these varietal

groups through the course of six cuttings during the sum-

mer of 1910 substantiates the conclusions already drawn

and illustrates the sharp differences in climatice reactions

which may be observed in the several varieties of a single

species.

No. 570] ALFALFA BREEDING 361

Morphologically the Mediterranean and Peruvian al-

falfas are so distinct in type that any one at all familiar

with the different kinds of alfalfa would recognize them

at a glance, whether a whole field or a single plant be

observed. The presence of yellow or greenish blue flow-

ers also determines a variety to be of northern origin with

mixtures of falcata characters, which usually carry with

them resistance to cold and drought. Otherwise, the

Turkestan, American and European types are so nearly

alike that only an expert would recognize them in mass

culture. The individual variations within these three

types intergrade to such a degree that one could scarcely

assume to judge, from the observation of a single plant,

the type prevailing in the field from which it originated.

The three types, however, differ markedly in their phys-

iological reactions as we shall presently see. The distinc-

tions, in this regard, as exhibited on our plots, are not

nearly so marked between the American and Turkestan

alfalfas as between these two types, on the one hand, and

the European, on the other. However, in northern cli-

mates where winter resistance enters as a potent factor,

the Turkestan alfalfa exhibits greater hardiness than the

American form, and, therefore, is able to maintain a more

perfect stand through seasons of extreme frost.

When grown under Arizona conditions, the average

yields of each of these five type groups present seasonal

curves at once striking in their diversity and contrasts.

These differences are exhibited more easily by plotting

the average of all the plots as a straight line, and the aver-

age of the different groups as percentages of the total

average above and below the ‘general average line.

In observing Fig.1, we are first impressed with a marked

similarity in the performance of the European and Medi-

terranean alfalfas, on the one hand, and the American and

Turkestan on the other, and also with the striking differ-

ences exhibited between the two groups.. Although the

average yield of the European plots greatly exceed that of

the Mediterranean plots, the shapes of their respective

362 THE AMERICAN NATURALIST [Vow XLVIII

curves are almost exactly alike, the greatest relative yield

of each being in the heated part of the summer after the

beginning of the water famine. In like manner, the

American and Turkestan varieties made similar relative

yield curves, that for the Turkestan being slightly above

the curve for the American strains. Here, however, the

CUTTING / 2 3 P. f é

$ JF

a Jt

S J+

> J

gar

5 at

wu 2#

[S]

@ S

a 20

AVE. YIELD 27/5 LBS. 27/6 LBS 72 LBS

a N

= & 5.6

PERCENT BELOW AVERAGE

RELATIVE YIELD OF REGIONAL VARIETIES BASED ON THE AVERAGE OF ALL PLOTS

AS 100 PER CENT.

curves bend strongly downward in mid and late summer,

as if these types were much less resistant to the accumula-

tive effects of drought and heat. In fact, it would seem

that during the hot period included within the fifth cutting

No. 570] ALFALFA BREEDING 363

(July and August), the American and Turkestan varieties

were comparatively inactive, yielding only about eight

hundred pounds of dry hay per acre, as against more than

| ton and a half each on the first cutting. The relative

~ield curve for the Peruvian type stands separate and dis-

tinct from the others. Although here, as with other varie-

ties, the yield declines with the advance of the season, the

persistence and vigor with which this strain resisted the

summer heat and drought caused it to gain rapidly on the

other varieties in relative yield throughout the season

until the very last cutting, when there was a slight decline.

Disregarding the shape of the curves we may now notice

the total yield for the season. In this respect the different

regional varieties take the following relative order: Peru-

vian, European, Turkestan, American and Mediterranean.

_ It is here noticeable that, though the European and Medi-

terranean varieties have similar seasonal yield curves,

they are not contiguous in the arrangement based on total

yields. This is a result of a marked difference in the

stand maintained by the two varieties which averaged

ninety-two per cent. for the former and seventy-four per

cent. for the latter. In their ability to maintain stand, the

Peruvian, European, Turkestan and American varieties

were about equal, averaging 92, 92, 93 and 94 per cent., re-

spectively. The lack of stand on the part of the Medi-

terranean alfalfas was not due to the poor quality of the

original seed, for all of these plots once had perfect stands.

This behavior is also in accordance with the records of

other fields of Mediterranean alfalfa in the southwest,

which have come under the observation of the writer. The

explanation of the weakness of the Mediterranean and

corresponding strength of the otherwise similarly reacting

European alfalfa in maintaining stand under Arizona con-

ditions is a subject for further careful physiological study.

The recognition, analysis, and calibration of these dif-

ferences of the physiological reactions of varieties are

thus seen to become a first essential in the study of cli-

matic adaptation, and form the basis for rational pro-

cedure in the choice of varieties and in selective breeding

for the improvement of the same.

364 THE AMERICAN NATURALIST [Vou. XLVIII

CORRELATIONS

In the improvement of varieties of plants, quality is

often as important as quantity of yield. This is especially

true in a forage crop, such as alfalfa. Since nitrogen, next

to fat, is the most expensive of the necessary food constit-

uents, it may be taken as the measure of quality. Com-

merical buyers judge alfalfa hay by its purity, odor, color

and percentage of leaves retained in curing and baling.

The value of the leaves lies in their relatively high nitro-

gen content and the consequent increased food value which

they impart to the hay. Expressed quantitatively, the cor-

relations between the nitrogen content of the hay and the

percentage of leaves for the six cuttings were as follows:

TABLE IV

CORRELATION BETWEEN NITROGEN CONTENT OF HAY AND PER CENT. OF LEAVES

Cutting 1st 2a 3d

Correlation yess so nae + 46+ .08 + 61+ .06 + .72 + .05

Cutting 4th 5th 6th

Correlation.«..i ve.cecierss + .68 + .05 + .61 + .06 + .52 + .07.

That the final value of the hay is markedly dependent upon

the composition as well as the percentage of leaves is

shown by the following high and fairly uniform correla-

tion between the nitrogen content of the hay and the nitro-

gen content of the leaves:

TABLE V

CORRELATION BETWEEN NITROGEN CONTENT OF Hay AND NITROGEN CONTENT

F LEAVE

Cutting Ist 2d 3d

Cörrolation o. oic cise + .69 + .05 + .73 + .05 + .42 + .08

Cutting 4th 5th 6th

Correlation, 6. ssccednacss + .67 + .06 + .85 + .03 + .74 + .05.

If, now, we have shown that the quality of the hay de-

pends primarily upon the percentage and composition of

the leaves, we may proceed to investigate those factors

which indirectly modify the feeding value by influencing

the amount or character of these organs.

The factors most profoundly affecting the percentage

of leaves were yield, height and stage of maturity at

which the cutting was made. Local or varietal forces were

No. 570] ALFALFA BREEDING 365

sufficiently constant to hold the place variation of this

character to the plus side of the equation for four out of

five determinations made, as is seen in the following

table:

TABLE VI

AON VARIATION IN PERCENTAGE OF _—

Cortelntion. aa 10. 10 + 23 4 10 + 36. +10 + res 08 — he 10.

These correlations, however, are low and seem to indicate

that the natural varietal traits were being overcome and

obscured by other variable factors.

Contrary to expectation, the stand had little to do with

the percentage of leaves, as the following low and incon-

stant correlations show.

TABLE VII

CORRELATION BETWEEN THE PERCENTAGE OF LEAVES AND STAND

Cutting 1st 2d 3d

Carton saras irri — .14+.10 — .02 + .10 + .03 + .10

Cutting 4th 5th 6th

Uorrelation ciso ccs +.10+.10 + .07 + .10 + .24 + .10.

On the other hand, the relation between height and

yield and percentage of leaves was constant and marked,

except in the last two cuttings.

TABLE VIII

CORRELATION BETWEEN PERCENTAGE OF LEAVES AND HEIGHT AND YIELD

Cutti 1st 2A 3a

Yield” Pecks cae saw eh oe lie — 41 + .08 — .60 + .07 —.15+.10

Har oa a — .48 + .08 — .62 + .06 — 68 + .05

Cutti 4th 5th 6th

Yield” bev esa wet oles dae — .40 + .09 + .20+.10 + .30 + .09,

OIRO ce ae — 55 + 07 + .09 + .10 + .19 + .10,

The sudden change from minus to plus in these correla-

tions should be noted. The average heights of the first

four cuttings were 32, 30, 28 and 27 inches, respectively.

The average height of the fifth and sixth, were 15 and 12

inches. This would suggest that at or below 15 inches the

mutual shading of the stems is not sufficient to cause an

appreciable shedding of the lower leaves. Up to this

point, moreover, growth usually takes place by an increase

366 THE AMERICAN NATURALIST [Vowu. XLVIII

in the number of nodes, each with its accompanying leaves

and side branches. Above fifteen inches, however, the

principal growth in height consists in a lengthening of the

internodes and, consequently, a relatively greater produc-

tion of stem as compared with leaf tissue. In this phys-

iological correlation lies the core of the difficulty in breed-

ing at once for quality and quantity. The act of high pro-

duction within itself cuts down the quality of the product

by reducing the ratio between the leaves and the stems.

This difficulty, moreover, occurs in the composition as

well as the percentage of the leaves. The correlation ex-

isting between the nitrogen content of the leaves and the

number of days required to mature a cutting is shown in

the following table:

TABLE IX

CORRELATION BETWEEN THE NITROGEN CONTENT OF HAY AND THE PERIOD

REQUIRED FOR MATURITY

Cutting 1st 2d 3d

Correlation: -ofo. 6 sence bes — .33 + .09 — .30 + .09 — 27 + .09

Cutting 4th 5th 6th

Gorritin. vee — 52+ .07 — .50 + .08 — 17 + .10.

Quickly maturing varieties thus have leaves richer in

nitrogen than those which require a greater length of time

for completion of growth. When, however, we take the

average number of days required throughout the season

to mature a cutting for each plot and compare this with

the total seasonal yield we find a correlation of -+ .43.

Thus we are again confronted by a minus correlation be-

tween quality and yield which must be overcome if we

would make progress simultaneously in both lines.

As further examples of antagonistic correlations, a few

instances may be taken from the data furnished by forty-

three plots of pure races of alfalfa grown during the sum-

mer of 1910. The correlation between height and percent-

age of leaves was again constant and marked. The results

here paralleled those found for the regional varieties.

Whereas yield was uniformly correlated positively with

both stooling capacity (av. No. stems per plant) and

height, it is interesting to note that there was also a uni-

No. 570] ALFALFA BREEDING 367

TABLE X

CORRELATION IN PURE RACES

Cuttings

Correlation Between

July August September | October

Green weight and average |

mber stems + .75 + .04/+ .42 + .08/+ .62 + .06 + .50 + .08

Green weight and average |

height + .01 + 10/4 .44 + .08/+ .22 + .10 + .33 + .09

Average height and number of |

ste — .29 + .09 — .19 + .10/— .82 + .09|— .21 + .10

Average height and per cent. |

(NOW naea e eaa — .39 + .09|— .15 + .10/— .55 + .07 — .51 + .08

form minus correlation existing between them. We thus

have two factors both making for yield, but seemingly

(probably physiologically) antagonistic to each other. In

breeding for high yielding strains we are here again called

upon to overcome by selection an antagonistic physiolo-

gical correlation.

This brings us to the following final conclusion which

the writer wishes to emphasize:

In economic plant breeding one frequently encounters

physiologically negative correlations such as those, in

alfalfa, between height and stooling capacity, height and

percentage of leaves, and between yield and quality. In

seeking improvement, therefore, the breeder must recog-

nize and make use of these facts in the interpretation of

results obtained, and also search for races which violate

such naturally antagonistic correlations to the greatest

possible extent.

GENERAL CONCLUSIONS

That the complex of allelomorphs, which we call a va-

riety, may be definite as both to ultimate composition and

organization is not here questioned. When, however, we

consider that visible characters are only the expression of

the reactions of the vital forces of the plant with the en-

vironment, we can realize that the variety, as we see it, is

not a definite thing, but is a result of two independent

classes of factors. Change either and the result corre-

spondingly changes.

368 THE AMERICAN NATURALIST [Vou. XLVIII

We are therefore to look upon the variety as a delicately

organized chemical compound. The various factors of

climate and soil may be compared to different physical

influences to which the original compound may be sub-

jected. As the chemist would expect reactions varying in

accordance with the physical stimuli used, so will the plant

react in agreement with the different environmental com-

binations. The extent to which this will change the nature

and appearance of plants is often far reaching. Cook,

working with cotton, has found that certain cultural condi-

tions at an early stage of growth will make profound dif-

ferences in the method of branching which determines the

whole subsequent development of the plant and affects

materially its economic value. Cultural and climatic reac-

tions often lead to error among those who assume them to

be mutative changes induced by the new conditions. That

these reactions may bring to light sub-races with heredi-

tary tendencies not hitherto called into expression and

which, by selection, may be secured as pure races, is the

probable explanation of many cases of supposed direct

climatic adaptation.

Thus, realizing the true nature of a variety, we can

draw further upon the analogy of the chemist who investi-

gates an unknown substance by testing its reactions with

a large number of known reagents. In like manner the

breeder can only understand the true nature of the hered-

itary vital forces within a plant after he has tested and

calibrated its reactions against a variety of soil and cli-

matic factors. These reactions are of interest to the

farmer only in so far as they affect the economic value of

the variety as grown in his own locality but to the breeder

and student of heredity their importance is fundamental.

This is so because they enable him to classify, coordinate

and interpret the experimental results that he obtains.

This ability finally must form the basis of all rational pro-

cedure, whether one be engaged in the study of pure gen-

etics or in the operations of practical plant improvement.

TAXONOMY AND EVOLUTION

By X,

“Some passages in this book, if taken alone and read hastily, may

appear to discourage systematic Zoology. This is far from my inten-

tion. No one can study the great naturalists of the seventeenth and

eighteenth centuries without feeling how seriously their work is impaired

by the defeetive systems of the time. It is not systematic but aimless

work that I deprecate—work that springs from no real curiosity in

Nature and attempts to answer no scientific iad "—T. C. Miall,

“Natural History of Aquatic Insects,” Preface, p. i.

INTRODUCTION

Linn&zvus bestowing Latin names upon animals and plants

was simply tripping gaily across the back of a half submerged

Behemoth and mistaking it for dry land. Now the beast is

careering around, and in spite of zoological congresses and inter-

national rules nobody quite knows what to do with him. No

doubt when some zoological ezar arises and issues his fiat a uni-

form system of nomenclature will be adopted and things will

begin to straighten themselves out. This can only be a matter

of time—the past can not be altered. On systematists to-day

necessarily devolves the dull, difficult and important duty of

going through the descriptive work of the early naturalists and

emending it; so that Spallanzani’s derisive sobriquet of

‘nomenclature naturalists’’ was a little unjust, even in his time.

Whatever opinions may be held upon the genius of Linneus,

in justice to him it should be said that it was not until his ex-

ample had been followed by a crowd of other workers eager to

attain to immortality by way of the back door he had left open

that the fat was really in the fire.

Well knowing the confusion into which systematic work in

zoology was brought by the early naturalists, modern systemat-

ists in our opinion will be the authors of a similar confusion in

the future if some of the slipshod methods of modern syste-

matics are not corrected. Moreover, a confused nomenclature

is not the least of the evils which second-rate systematic work

brings in its train.

: 369

370 THE AMERICAN NATURALIST [Vou. XLVIII

Systematists with a proud curl of the lip may tell us that the

work is not done now as it once was. Indeed, to those who are

not able to project themselves into the future it may seem in-

credible that the systematists of a later date will be able to find

much room for complaint in the elaborate descriptions and care-

ful figures of modern descriptive writers. For the moment, how-

ever, it suffices us to point the parable by remarking that in

1780 Spallanzani was able to refer to the ‘‘beautiful figures’’

and ‘‘careful descriptions’’ of a systematic worker on frogs.

We, of course, know without seeing them that the figures were

not beautiful nor the description, careful—any way in the sense

of being complete. We have therefore to reflect whether the

zoologists of a future generation will find the work of to-day any

freer of faults than that of the past centuries. vik

SYSTEMATIC Work. GENERAL CONSIDERATIONS

It is necessary to insist at once that systematic work is not

merely a question of nomenclature, names and novelties. Sys-

tematists have only themselves to thank if such a narrow con-

ception of their provinee is very widely spread, especially

among morphologists and anatomists, who. are ready to belittle

the value of the systematists’ work. But science is measurement

and zoology—if you like—is description, and it is impossible to

dispense with the systematists’ descriptive work. But we think

it possible to dispense with a good ‘deal of stuff after this

fashion :

Metopidium high, mab raktinernin rather long, acute, arcuate and

curved at the tips. Pronotum roughly punctured at the bottom of fine

furrows. Color dark-ochreous. Posterior horn uniformly cylindrical,

undulating or sinuous without rugosities. Underside, seutellum and

legs sordid-ochreous.

The phrase ‘‘sordid ochreous”? comes ready to hand and

ry it unnecessary for us to go in search of a suitable com-

men

This is the 30th memoir” writes a PPPE ‘on the

Zonitidæ which I have published in this journal, describing in

all about 560 new species.’’ We feel inclined to put our hands

resolutely on his shoulders and inquire if he ever saw a cteno-

phor swimming in thé sea or watched the progress of an Asterias

towards its prey.

No: 570] “°° TAXONOMY. AND. EVOLUTION. STE:

- No one'can look unmoved upon the Hymenopteran or Helicoid

Specialist with head bent over a drawer full of shells or dried

insects on`pins. It is not that we resent concentration or enthu-

siasm or even specialization, but the systematist has lost touch

with his own science of zoology.

' Zoology, a cornucopia of marvels, lies at his elbow full to over-

flowing, but he is unmindful of it. It-is as if a man should use

the Parthenon only ‘as a convenient place on which to strike a

match for his pipe.

The divorce between systematic work and the rest of zoology

is the more regrettable because it is practically complete. It is,

we admit, expedient that zoology should be divided up into

anatomy, morphology’ and so.on. But such a division is allow-

able only when it is expedient, while for intellectual purposes

such a division is and has always been a danger. To obtain

facts one must be an analyst, to consider them one must be a

synthesist. _ Between the two there is all the difference between

a hodman and a natural philosopher.

- But our contention is that not even the plea of practical ex-

Pia ior can justify the extreme state of specialization into

which: systematie zoology has fallen, making itself manifest in

the concatenation of such purely artificial characters as that

‘the third joint of the antenna is longer than the second, that

the mesoscutellum ‘is ovate and the color pink with blue spots.’’

All this simply makes one yawn, though there is this much to be

said in favor! of this stamp of systematist, that nothing bores

him so much as the recitation of one of his own diagnoses or

being introduced to the systematist of another group.

Systematic work is a withered branch of the biological tree

which’ there is still hope of rejuvenating. Treviranus long ago

remarked that if we once regarded systematic. work as a part of

biology and nomenclature as a means to an end rather than as

an end in itself, both might take their places in science. Let us

take every precaution against systematic work becoming one of

those unproductive and artificial pursuits which spring up like

mushrooms around centers of splendid endeavor and high

achievement. After Shakespeare came his commentators. Shall

it be said that after biology came the systematists?

^. We assume that the principal object of systematic work is to

discover the phylo-genetic. classification of animals, for which it

is surely necessary that every animal as it passes through the

372 THE AMERICAN NATURALIST [Vow. XLVIII

systematists’ hands should be, as far as possible, thoroughly ex-

amined and described, no dependence being placed upon a few

superficial characters usually selected from the external parts?

That the systematist should concern himself, as he does, with

the external parts, leaving the anatomy to other workers, we

consider is as bad for the systematist himself as it is bad for the

science; for himself, he is doing work which can only keep his

soul alive with difficulty—superficial clerical work which can be

‘prompted by no real curiosity and attempts to answer no

scientific questions,’’ and the results of the work itself is often

invalidated by the arrival of the destroying angel in the person

of the anatomist. For a superficial description often means a

wrong classification ; whence it follows that any zoo-geographical

deductions therefrom are invalidated; while a careless descrip-

tion usually ignores the possibilities of variation and shows no

evidence of pains having been taken to make identification easy.

Systematic work, then, is concerned with classification, geo-

graphical distribution, variation and identification, and there

would be no need for this paper, if it were more generally re-

alized that one thorough examination and description of the

whole animal assists those branches of the inquiry more than

twenty loose and superficial ones.

Of course systematic workers are not the only zoologists who

over-publish ; yet they especially might cultivate a little of the

salutary reticence of C. L. Nitsch and Alfred Newton, who, with

no discredit to themselves, wrote and published little, yet it must

be admitted by those with an eye on the extravagant output of

others, to the advantage of zoology. The words ‘‘res non-verba’’

were the motto of Delle Chiaje, who, like Nitzsch, on his death

left behind many important discoveries unpublished and only

indicated in his drawings.

CLASSIFICATION IN GENERAL

The coming of Evolution meant for systematic workers that

no system of classification would henceforth be considered as a

serious contribution to science, which was not constructed on

phylogenetic lines. It meant the final overthrow of such ideas

as Agassiz held, that the divisions of the animal kingdom were

instituted by the Divine Intelligence as categories of his mode of

thought—of such fantastic systems as those, of Rafinesque and

Swainson and such strictly artificial ones as the arbitrary ar-

No. 570] TAXONOMY AND EVOLUTION. 373

rangements of convenience which should be now used only in

those groups where, and for as long as, our knowledge of the

anatomy is so slight that some sort of temporary device for

sorting out genera and species has to be adopted.

The ideal system is now phylogenetic, i. e., it aims at recon-

structing in a genealogical tree the actual lines of descent.

Only those who have attempted the reconstruction of phylo-

genetic trees understand the intrinsic difficulties of the work.

There can be no doubt that the coming of Evolution has put

before the systematist a very difficult task, As to whether the

methods usually employed by him are adequate to the demands

placed upon them we are frankly sceptical.

Fortunately for the systematist the main lines of classifica-

tion in most groups are given him ready made by the morphol-

ogists who have laid down the foundations trusting to the ‘‘sys-

tematist’’ to fill in the details. Such classifications—the main

phyla, classes and orders are of permanent value, because they

are founded upon a combination of characters of tried worth

judiciously selected after a careful survey of extensive embry-

ological and anatomical data.

SINGLE CHARACTER CLASSIFICATION

On the other hand the minor systems—the families, genera

and species—the realm of the ‘‘systematist’’—too frequently

consist of haphazard combination of a few characters selected

because of their convenience in not entailing any anatomical

work, or selected on account of the ignorance existing of any

other—particularly internal—important characters. Ignorance

of their morphology has been the main reason for the difficulty

in classifying the Coleoptera. Entomologists are especially

prone to give their whole attention to what is visible without

the aid of dissection. In the Polyzoa the majority of forms are

only known by their external appearance and their classification

is proportionally unsatisfactory. In the Mollusca reliance is

placed on the shell; in mammals the skull and the skin, in birds

the plumage are the articles of faith.

Single character classification or diagnosis by one or two

characters, as zoological history shows, has proved inadequate—

that it is unphilosophical is patent to all.

Such single character classification even when practised by

374 THE AMERICAN NATURALIST [Vou. XLVIII

the great morphologists, men who, being acquainted with the

whole of the anatomy of the forms they were classifying, de-

liberately selected one or two characters after a survey of the

whole—was rarely a success. Huxley set out unabashed to

classify birds by their palate, and Agassiz fish by their scales—

systems which have now shared the fate of most others which

set out to erect a classification on the modifications of a single

organ alone. Alfred Newton said that there was no part of a

bird’s organization that by a proper study would not help to

settle the great question of its affinities.

The systematist who deals with the minor subdivisions of the

animal kingdom—families and genera—shou e as much a

morphologist as the one who deals with the larger—the phyla

and classes.

DESCRIPTION

We have pointed out above that the adequacy of a system of

classification depends in great measure upon the thoroughness

of the description of the species and genera. Classification in

all groups has progressed in just proportion to the more exact

examination of the species considered in the classification. |

The history of zoological research brings out this fact very

clearly, beginning with the work of Linnzus, the originator of

the superficial diagnosis, passing on through Cuvier, who appre-

ciated the value of anatomical knowledge, to Von Baer, who

emphasized the importance of embryology.

_ It was not a “‘systematist’’ as we know him who first correctly

classified Lepas—the econchologists blindly accepted it as a

Molluse. It was not a ‘‘systematist’’ who first established Peri-

patus as an Arthropod, for the first describer of that animal

regarded it as a slug!

How rare it is to find in a description of a new species any-

thing more than an indication of the external parts. It is a

peculiarly arbitrary limit to a man’s curiosity that restricts

his enquiry to the superficial aspect of an animal. A natural

philosopher ought never to be satisfied with the external ap-

pearance of things. The wisdom of the ancients bids us ‘‘be-

ware of what things appear ’’; and the method of our modern

science is one of close and detailed observations. In scattering

names broadcast with liberal largesse upon species, varieties and

No. 570] TAXONOMY AND EVOLUTION. 375

genera, systematists have sometimes dropped into some curious

errors. Teratological specimens have been described as new

species and most zoologists have heard of the man who de-

scribed as a new species the longicorn beetle, the head of which

having fallen off, had been fixed on upside down. His examina-

tion of a new species makes so slight an impression on his mind

that sometimes the same worker has described the same form

twice under different names.

The descriptive papers on Mollusca usually consist of short

descriptions of the shells, even written in a dead language. This

is conchology. Conchologists confine themselves to the pat-

terns and shapes of shells—nature’s medallions—numismatics !

Much of this work—along with similar productions in entomol-

ogy and carcinology—we regard as positively fiagitious.

Sir Ray Lankester in the article ‘‘Zoology’’ in the Encyclo-

pedia Britannica (ed. XI.) remarks that museum. naturalists

must give attention to the inside as well as to the outside of

animals and that to-day no one considers a study of an animal’s

form of any value which does not include internal structure,

histology and embryology in its scope. Agassiz, too in his

famous ‘‘Essay on classification’’ wrote that ‘‘the mere indi-

cation of a species is a poor addition to our knowledge when

compared with such monographs as Lyonnet’s Cossus, Bojanus’

‘Turtle’ Strauss Durekheim’s Melolontha and Owen’s Nauti-

lus.” i

‘“‘But,’’ it will immediately be asked in chorus, ‘‘do you

seriously suggest that a monographie volume should be devoted

to every new species?’’

This is a leading question which brings us to the crux of the

whole matter, and ean not be answered in simple ‘‘Yea’’ or

Ney?’ .

THE PROVISIONAL DIAGNOSIS

The amount. of analytical study that may be given to any one

animal form in any one stage of its development is infinite.

The result is that in describing a new species for the purposes

of exact phylogenetic classification there must be a limit beyond

which it is unnecessary to go. Such a limit can not be otherwise

than arbitrarily selected according to the best judgment of the

systematic worker as to how much analysis is required to place

his new species, although at present, miserabile dictu, relatively

376 THE AMERICAN NATURALIST [Vov. XLVIII

very few animals have been thoroughly explored, yet in the dis-

tant future, in the millennium, it can not be doubted that every

genus, even every species will have been examined in toto in every

stage of its development and life-history as thoroughly as our

instruments and eyesight will allow, and perhaps a whole vol-

ume or several volumes will be devoted to every animal form.

At present, however, it is a waste of ink to consider a future so

far away. A more pressing duty is to consider how far modern

methods of superficial diagnosis fulfil the obligations placed

upon systematists not to give an exhaustive analysis of animal

forms, but to give sufficient data to meet the searching demands

of phylogenetic classification.

We are aware of the fact that the convinced and determined

systematist does not maintain that the method of superficial

diagnosis does meet or is intended to meet the demands we have

been indicating. If he reads as far as this and does not throw

aside this paper in contempt, he is ready with eager forefinger

and glib apology to convict us of begging the question that sys-

tematic zoology can be ever anything, or should be ever any-

thing more than we have sai

It is often argued that the srperiéial diagnosis of the syste-

matic worker is simply a provisional diagnosis awaiting the con-

firmation of the anatomist. A plausible defence of the provi-

sional diagnosis is advanced by many workers in perfect good

faith which it is now necessary to anticipate and examine.

This argument defends the provisional diagnosis on two

grounds: (1) The advertisement theory; (2) the recognition

mark theory.

The supporters of these theories admit that the provisional

diagnosis in no way settles either an animal’s systematic posi-

tion or its validity as a species. But it is alleged to be of value

and should be encouraged because it advertises the existence of

a presumptive new form which would otherwise remain un-

known and overlooked in the store rooms of the museum and

laboratory, and because in giving an account of the external

parts, at all events, the systematist is describing those features

by which we are more or less easily able by a superficial exami-

nation to recognize summarily the form when it turns up again.

The first part of our answer amounts to a recapitulation of what

has been previously stated in general, viz., that systematies have

lost touch with the rest of the science. The output of systematic

No. 570] TAXONOMY AND EVOLUTION. 377

work and the output of anatomical and morphological work

nowadays move along completely different channels. The work

turned out by the systematic worker is scarcely, if ever, con-

ceived in the light of modern biological theory, is rarely couch

in terms of modern biology and rarely indicates a problem to be

solved or a question to be answered. It proposes distinctions

the anatomist sweeps away and hazards affinities the morphol-

ogist laughs at. It performs work that has to be done over

again, and instead of giving the morphologist what it claims to

give him—a sketch map of the country he is to traverse—all it

does is to bewilder him with a Will-of-the-Wisp’s lantern, an

intolerable multitude of slipshod and untrustworthy directions

that he has come instinctively to suspect. We can not too often

ask the question, why should the work be done twice? Surely

it is time that something were done to stop this tremendous rush

for publishing provisional diagnoses that more time could be

devoted to the systematic study of animal forms, obtaining

thereby sound phylogenetic classification, sound deductions in

geographical distribution, valid species and a less confused

nomenclature.

Thus the systematist’s protest that at least he ‘‘advertises’’ pre-

sumptive new forms we can reply that he may do so, but that for

any purpose other than a dull census of the animal kingdom with

a very generous ‘‘-+’’ to it, me is a positive Benedick of zoolo-

gists, for ‘‘nobody miris him.’

The upholders of the provisional diagnosis will say that at

any rate they are giving us a description of the external parts

and are increasing our knowledge by so much. True, but by so

inconsiderable an amount that when the anatomist comes along

with his scalpel he so quickly disposes of the external parts

merely by the use of his eyes that it is a matter of indifference

whether the former have been described or not. Moreover, the

great majority of the tens of thousands of descriptions that are

issuing from the press are of animals so closely related to pre-

viously described species that such descriptions really amount

to little more than a recitation of their distinguishing characters.

It is certainly useful to know that Caccabis rufa is to be dis-

tinguished from Perdix cinerea by its red legs and that the

Leporide can be discriminated by the character of their upper

incisors. But the question may well be asked, what is the use

of being able to distinguish one species from another without

378 THE AMERICAN NATURALIST [Vou. XLVIII

being able to record at the same time anything about its bionom-

ies or anatomy which would give the distinction its real value.

A great deal is known about the partridges and hares, hence the

distinctions alluded to above are useful as an easy way of

quickly identifying them. But so long as nothing is known

about either of two species that are distinguished we are none

the worse off, if both remain indistinguishable.

Finally we would point out that of all people the systematist

should know that at present of the forms he advertises and

describes so copiously and summarily only a fractional part is,

or can be, dealt with by the laboratory worker. We are speaking

now of the anatomy pure and simple of new species and genera.

The laboratory worker proceeds slowly, is fewer in numbers

and has other problems—embryology (descriptive and experi-

mental), heredity, physiology (descriptive and experimental)

and morphology to attend to besides purely descriptive anatomy.

And yet anatomy—the very corner stone of the temple of

zoology—has to be restricted in output because none of the sys-

tematists will learn how to use a scalpel or look down a dissect-

ing-microscope—feats in themselves perfectly easy and calling

for no special training or faculties. -

Possibly the upholders of the provisional diagnosis will main-

tain that by publishing his account of the difference between

closely allied forms the systematist is providing the biologist

with a stimulus to discover how much deeper such differences go.

But surely it is a strange perversion of a man’s natural instinct

of curiosity that enables the systematist to rest content with

advertising problems instead of endeavoring to equip himself

for the task of undertaking them himself, who is eminently

suited to the work and whose — daily brings him into

close contact with them.

Finally we would point out that the enormous mass of species

which have been created upon superficial diagnosis so far have

remained unincorporated for the most part in the structure

it is designed to build up, viz., a clear comprehension of the

phylogeny of the lesser divisions of the animal kingdom. It is as

though a man were to set about building a house by making a

vast quantity of bad bricks and then to leave them scattered

about his site in the hopes that some one would come along and

make a house of them. Surely it is an economy of effort for the

systematist to take up the bricks and build himself, what time

No.570] `. TAXONOMY AND EVOLUTION: 379

the T and morphologist are engaged upon their own

- special tasks.

THE COMPARATIVE VALUE OF INTERNAL AND EXTERNAL PARTS

Briefiy reviewing the discussion as far as we have carried it,

it will be seen that we are asking for sound phylogenetic classi-

fication of the smaller groups as well as of the larger ones, based

not upon single characters, but upon the whole of the characters

regarded collectively, for more careful and more thorough mor-

phological methods in description and for. the discontinuation

of the provisional diagnosis. In view of the desirability of work-

ing up sounder schemes of classification from the enormous, un-

wieldy and superficially known mass of genera and species sys-

tematists can be rendering little service by continuing to turn

out indiscriminate provisional diagnoses.

It remains now to discuss in greater detail the proposal we

bring forward in the place of the provisional diagnosis.

The commonly accepted opinion is that while for the classifi-

cation of families and orders the internal parts must be taken

into consideration, for that of species and genera a summary of

the external parts is all that is required. On account of the

labor and difficulty sometimes involved in dissection we are too

ready to assume that the internal parts in genera and species

present a dismal monotomy of character which it would be

profitless to investigate for systematic purposes.

If it is admitted that internal characters are of value among

the higher divisions of the animal kingdom, can the systematist

tell us at what precise point in the downward scale they cease to

have value, and at which reference need only be made to the

external parts? Even supposing for a moment that there is

such a limit, we are strongly of opinion that it does not come

before the genera.

A genus is of different value in different groups but as a rule

it presents so much difference in external form from other

genera as to warrant the inference that internal differences of a

like extent will be found if sought for. At the present moment

a genus is a perfectly arbitrary collection of species. We ven-

ture to prophesy that with more elaborate descriptions inter-

generic relationships will be more carefully defined and genera

will become less heterogeneous and more natural. But this is

by the way.

380 THE AMERICAN NATURALIST — [Vou. XLVIII

A priori it seems improbable that less variety will be found

among the various internal systems of organs than in the in-

tegumentary or exoskeletal parts. But an argument may be

put forward that the external parts in immediate contact with

the environmental forces would be the first to register change in

the modification of a species. The internal parts as stanchions

and bulwarks remain firm to give characters to orders and fami-

lies, while change makes assault without and gives characters

for species. For example, among the Asteroids it is said that

the internal organization is so uniform that the only method of

classification is to take the different ways in which the demands

of the external environment have been met. :

But generally speaking a species depends for its survival not

simply upon the external front it presents to its environment.

An animal’s form cannot arbitrarily be divided into external

and internal parts. It is an integral whole, and variation and

selection may occur anywhere, while the correlation of variation

is a text-book commonplace. As opposed to correlative variation

there is the law of the independent variation of parts. Not only

may variation occurring in one part cause a variation to take

place in another, but variation may take place independently in

some areas and be limited in another, so that in deciding upon

the comparative value of the internal and external parts in any

group consideration must be given to both these laws. In the

Asteroids, we assume that anatomists have taken the matter in

hand and found that the external parts vary as a rule independ-

ently of the internal which remain constant. But in how few

groups has such a precaution been taken! Is it not rather the

general rule simply to assume that the internal parts lack varia-

tion and are of no value systematically, as, for instance, in the

Lepidoptera, where the Lepidopterists expect that a classification

based upon the wing-markings or upon wing-neuration can ex-

press the true relationship of the various units?

Even in those groups where systematists have dissected and

found the internal parts valueless it still remains necessary, in

view of the law of independent and unexpected variation of

parts for them, to apply the scalpel to every new form

It is impossible to deny that the external parts are often of

extreme systematic importance—they are exposed to the light

and develop color patterns (although color is usually an unsafe

guide if taken alone), and the external parts of such forms as

No. 570] TAXONOMY AND EVOLUTION. 381

Arthropods and Molluses being hard provide systematists with

a sculpture on which it is easy to detect minute differences in

pattern. On the other hand we would remind the conchologist

that the external parts are by their very positions most liable

to exhibit lesions and weathering, and certainly in the case of

Mollusca where the dependence of the exoskeleton upon a spe-

cific article of diet (viz., lime salts) is very close, to register

“*fluctuating variation” according to the constitution of the

medium or of the food ingested.

But here again if a more common practise were made in dis-

secting by systematists, variations would be found even in closely

allied species making the descriptions complete and often

even necessitating the erection of new genera. One of the

writers was dissecting an ordinary species when he discovered

that the epipharynx was so entirely different in form and struc-

ture from the usual type for the genus that, had it been an

external character it would long ago have been formed into a

new genus.

Karel Thon? has demonstrated how in Holothyride a single

internal structure is at variance with the other indications of

genetic affinity. A great many similar instances will be immedi-

ately called to mind by those who practise dissection.

Again, if systematists are convinced of the taxonomic value

of hard parts how comes it that they. need to be reminded that

there are hard parts in the internal anatomy as well which they

so frequently and habitually leave unnoticed? The endoskele-

ton of Arthropods, gastric mills, pharyngeal ossicles and carti-

laginous supports are all systems which might be profitably

studied by the entomologist and carcinologist, while the con-

chologist generally proceeds as though the radula and jaw were

part of the ‘‘mush,’’ as he so inelegantly terms the viscera.

GEOGRAPHICAL DISTRIBUTION

The advent of the morphologist into the particular sphere of

systematics or the metamorphosis of the systematist into a mor-

phologist (it matters not how we put this desirable event) will

result in the annexation not only of classification, but also of

questions of geographical distribution by anatomy and morphol-

How many pretty theories in geographical distribution

1 Zool. Iahb., Bd, XXIII, Syst., pp. 720-21.

382 THE. AMERICAN NATURALIST [Vou. XLVHE

have éollapsed because they were built on the sands ofan in-

correct classification? The similarity between the faunas of

South America and Madagascar is ‘supported. by many- facts,

but the value of Solénodon in Cuba and Centetes in Madagascar

has been lesseried by the recognition that the two genera re-

semble each other by convergence, and should now be paatna

in different families.

The Dendrobatine also are’ rébuekdersd by Dr. Gadow as. an

unnatural group, the two divisions—South American and Mas-

earerie—having, according to him, lost their teeth independently.

Again, Dr. Gadow refers to the Ratite as a heterogeneous as-

semblage of birds which is ‘‘absolutely’ worthless’? for -the

zoogeographer. There are scores of such artificial groupings—

the work of the a vette ii rahi wa APOT

astray.

The result is ENI ayetematid wolk: as at aol fiend is

of very little use to us in the study of geographical distribution.

It is hopeless nowadays for a zoologist to sit down with “a

list of species and their range ‘and trusting implicitly in sys-

tematic work to make maps of-distribution’ and, as he so often

does, to draw deductions therefrom, for the validity of such de-

ductions must ultimately depend upon the anatomical and mor-

phological data. Moreover the study of geographical distribu-

tion is developing new methods of tackling its problems.

` We do not consider it necessary tó touch on the other remedies

that might be applied with a view to redeeming zoological taxo-

nomy from its present artificial state A to oe it into line

with the rest of biology

Such remedies—for fetannd testing the validity “of species

by genetic experiment and the intensive study of variation—

have been advocated many times before,? although with little

success. We believe, however, that the reforms in descriptive

zoology we have advocated above are the more urgent.

' 2Cf. E. B. Poulton, ‘‘Essays on Evolution,’? 2. ‘*What.is a-Species?’’ `

and K. Jordan, ‘‘ Novitates Zoologice,’’ 3, 189

SHORTER ARTICLES AND DISCUSSION

NABOURS’S GRASSHOPPERS, MULTIPLE ALLELO-— |

MORPHISM, LINKAGE AND MISLEADING

TERMINOLOGIES IN GENETICS

In a review of Nabours’s breeding experiments with grass-

hoppers,’ Mr. Dexter makes a distinction between an interpreta-

tion of Nabours’s and his own, where I fail to see a difference ex-

cept in terminology. This is so typical of much recent Mendelian

work that I am tempted to call attention to it.

Nabours describes a cross between a female with characters BI

and a male with characters CE and comments on the production

of an individual with characters BEI. He says, as quoted, that

the ‘‘female parent gave at least one gamete containing the fac-

tors for the patterns of both her parents (B and J) and that this

double character gamete was fertilized by one of the E gametes

which came from the CE male.”

Dexter prefers to call the supposed exceptional BI gamete of

Nabours Bcel, and the supposed E sperm which fertilized it bcEi,

stating that Nabours’s terminology would involve multiple allelo-

morphism, his own linkage. (Nabours uses, I think, neither ex-

pression.) Now what is the difference between the two interpre-

tations? Is it anything but verbal? Is there anything significant

in the small letters which Dexter has added to Nabours’s form-

ule? If so, what is their significance? Do they mean any more

than the extra zeros in the expression 1.000 as compared with 1.0?

Dexter proposes an experimental test, that the cross be re-

peated. “If then BEI forms should appear again and in these

when mated to other forms the factors B and I should be found

to stay together to the same extent as they before separated, it

would show that close linkage, rather than multiple allelomorph-

ism explains this particular instance.” How would it show it?

If we take Nabours’s assumption that B and J have exceptionally

gone into a single gamete and formed with E a zygote BIE,

would it be counter to his assumption that they should subse-

quently hang together and that gametes should arise BI and E,

respectively? Would adding a few small letters to the formule

1 Am. Nar., May, 1914.

383

384 THE AMERICAN NATURALIST [Vou. XLVIII

alter the case, changing it from multiple allelomorphism to link-

age’? It seems to me that this is one more case in which a fallac-

ious conclusion is reached in consequence of using small letters

for absent characters in Mendelian formule. Professor James

Wilson has pointed out others.

W. E. CASTLE

BUSSEY INSTITUTION,

ForEsT HILLS, MASS.,

May 6, 1914

VOL. XLVIII, NO. 571 JULY, 1914

THE

AMERICAN

NATURALIST

A MONTHLY JOURNAL

Devoted to the Advancement of the Biological Sciences with

Special Reference to the Factors of Evolution

CONTENTS

Page

I. Pattern Development in Mammals and Birds. GLOVER M. ALLEN - — 385

II. Internal Relations of Terrestrial Associations. ARTHUR G. VESTAL- - 413

Til. Shorter Articles and Discussion: Another Hypothesis to Account for Dr.

Swingle’s Experiments with Citrus. A. C. and A, L, HAGEDOORN

THE SCIENCE PRESS

LANCASTER, PA. GARRISON, N. ¥.

NEW YORK: SUB-STATION 84

The American

Naturalist

intended for panais and books, etc., intended for review should be

MSS

sent to the Editor of THE AMER

Short

CAN NATURALIST,

articles containing summaries of research work bear

Garrison- -on-Hu rs New York,

on the

problems of = evolution are especially welcome, and will be given iprefeteiae

in publication

ndrea reprints of egika are supplied to authors free of charge.

ne hu

Further reprints will be supplied a

ubscri paa and advertisement should be sent to the publishers.

a year. Foreign postage

Canadian postage twenty- -five cabs additional.

subscription price is four dollar

forty cents.

The

s fifty cents and

The orge for single copies is

The advertising rates are Four Dollars for a pa

THE SCIENCE PRESS

Lancaster, Pa.

Garrison, N. Y.

NEW YORK: Sub-Station 84

Entered as second-class matter, April 2, 1908, at the Post Office at Lancaster, Pa., under the Act of

Congress of March 3, 1879.

FOR SALE

ARCTIC, ICELAND and GREENLAND

BIRDS’ SKIN

Well os cen Prices

rticu

G. DINESEN. ota jt

Husavik, North Iceland, Via Leidle, England

WANTED TO PURCHASE

a set of BIRDS OF AMERICA by J. J. Audubon,

7 or 8 volumes, please report, stating cash price, stat-

ing condition, binding and dates of volumes.

F. C. HARRIS,

Box 2244 Boston, Massachusetts

For Sale Entire

An important collection of Indian Birds’

Eggs, containing approximately 3,300

specimens of about 620 species, and in-

cluding among other varieties the Green-

ish Willow Warbler (Acanthopneuste

viridanus), Red-browed Finch (Calla-

pygius), Bearded Vulture (Gypaétus

barbatus), numerous Cuckoos with their

hosts, etc., ete.

Particulars and full list may be had

from

W. F. H. ROSENBERG

57 Haverstock Hill London, N. W.

Photographic Films Developed

5 cents per Spool, Prints 3 cents up, Post

Cards 5 cents.

8x10 Enlargements 25 cents mounted.

or Samples and Price List.

FRANK R. BARBEAU

89 West Bridge Street Oswego, N. Y-

eae E

The iray of Chicago

e set eng Saw

à during t Erga quarters of the

academic th

4 The cee Bs evades colleges, the

A graduat ools, and the [aeg

ua!

sional schools provide courses 10

er Qu

2d Term July 23-Aus- 28

pas cannes will be

pon application.

The University of Chicago

(Mitchell Tower Chicago, Minois =

THE

AMERICAN NATURALIST

Vor. XLVIII July, 1914 No. 571

PATTERN DEVELOPMENT IN MAMMALS

AND BIRDS

GLOVER M. ALLEN,

Boston SOCIETY or NATURAL HISTORY

THe particular coloring of mammals and birds is pro-

duced by two factors—pigmentation and the physical

structure of the hair or feathers. Both are often present

together. In certain mammals, for example the golden

mole (Chrysochloris) and the European Galemys, a

beautifully iridescent sheen is produced by the reflection

of light rays having a certain angle of incidence upon the

hairs which themselves contain pigment of a character-

istic color. In the duckbill (Ornithorhynchus) the same

thing is found. The peculiarity of feather structure that

causes iridescence is largely developed in certain families

of birds, as the hummingbirds and the pigeons (see

Strong, 1904, for an account of the feather structure).

It is not my purpose to discuss the use of this irides-

cence to the bird, beyond stating my belief that it is in

part at least for sexual display, as no one can doubt who

watches the male street pigeon strutting before his mate.

With amorous coos and lowered head, he confronts her

and, swelling out his throat feathers, turns about and

about, so that the light is reflected from his neck and

throat in a sparkle of rainbow hues. It has also been

Suggested (Thayer, 1909) that iridescence may be a

strong factor in concealment, since from the variety of

the colors produced the bird is more difficult to resolve

from its many tinted environment amid foliage and

flowers.

385

386 THE AMERICAN NATURALIST [Vou. XLVIII

With many birds the characteristic coloration may not

be at all that of its pigment. Thus the blue of the male

indigo bird (Passerina cyanea) is due solely to the phys-

ical structure of its feathers which though pigmented

with brown, appear blue by reflected light. If, however,

a blue feather be immersed in oil and viewed under a

microscope by transmitted light, it is seen to be brown-

pigmented. The physical feather-structure of the adult

male is thus in this species a secondary sexual character

chiefly developed during the breeding period.

The important point at present is, however, that the

color effects just described are none the less due to pig-

ment, quite apart from the fact that the apparent color

of the pigmented area may be different from the actual

color of the pigment (except that iridescence may some-

times be faintly seen in an unpigmented feather).

The use of pigmentation to its possessor is a matter

still under discussion and investigation. In many cases

it is doubtless the result of purely physical causes and

it is quite without the power of the animal to make use

of its coloration for outward effect. Thus the beautiful

colors inside the shells of some molluscs are never appar-

ent from an exterior view, and are supposed by some to

be in part a waste product, the result of metabolism

within the organism.

The present discussion has to do only with the external

pigmentation of the hair and feathers, respectively, in

mammals and birds.

The simplest cases of coloration are those in which the

body or its covering is everywhere of the same hue, or

nearly so—as in the elephant, the wild buffalo, or the

house mouse in which the hairy covering (or hide in the

elephant) is of a nearly uniform tone everywhere. So

too, the crow, the apteryx, and the nestlings of many

birds whose parents show a more highly differentiated

style of markings. Such mammals and birds, so far as

the development of pattern is concerned, I would con-

sider unspecialized, yet it does not follow that in this

No. 571] PATTERN DEVELOPMENT 387

respect they are also primitive, though in most cases I

venture to think this may be true. The uniformity of

plumage is probably a derived condition in such a species

as the Cuban blackbird (Holoquiscalus assimilis) in which

the duller colored females have yet a yellow patch at the

bend of the wing, a style of marking widespread among

allied forms. The adult males, however, have lost this’

and are wholly black. Gadow as well as Keeler (1893)

conclude that among related species in which there is a

tendency to differentiation of the coloring the end result

of the stages through which the species may pass is the

production of a wholly black bird. In general a wholly

black condition is no doubt to be considered as a derived

rather than a primitive state among birds whereas a uni-

formly dull plumage of a brownish or grayish tone is

probably in most cases primitive. Among mammals the

same is probably also true, for in both the black condi-

tion indicates either an excessive production of the black

over other associated pigments, or a loss of the power to

produce the latter, whereas the neutral gray or brownish

coloring is due to a more even mixture of such pigments.

As pointed out by Professor W. E. Castle, the ‘‘ticked”?

pattern of the hairs of mammals is probably primitive,

and it is certainly very widespread. It is well illustrated,

for example, by the house mouse (Mus musculus) or the

wild guinea-pig (Cavia), in which three separate pig-

ments occur as granules in the individual hairs—yellow,

chocolate, and black. These three in their normal mix-

- ture produce a neutral gray tint—mouse color—and an

examination of this type of coat usually shows that some

hairs are wholly black, others dark at base barred with

black and yellowish near the tip.

There are two ways in which patterns may be developed

from a uniformly tinted covering of hair or feathers: (1)

by a local change in the relation of the associated pig-

ments so that in certain areas only one or two sorts are

produced instead of three, or only one; (2) by a failure to

388 THE AMERICAN NATURALIST [Vou. XLVIII

develop pigment at all in certain places, so that a white

or unpigmented area is produced.

It is not rare among mammals to find that one or more

of the characteristic sorts of pigments are not produced

in certain individuals and probably the factor or factors

for these are lost altogether from the somatic and sex

cells alike. Such variations may be perpetuated through

inbreeding and so no doubt have arisen sundry domestic

color varieties of animals and plants. For example, in

the course of experiments with color varieties of the

house mouse (carried on some years since with Professor

W. E. Castle) we found that the chocolate-colored mice

which we bred as extracted recessives from black mice,

contained only chocolate pigment in their hair, whereas

in the black parents both black and chocolate pigments

were present, but the black masked a chocolate pigment.

Moreover, the chocolate mice always bred true to that

color, but if bred back to the black parents, gave black

young or both black and chocolate in Mendelian propor-

tions, according to the nature of the matings. The inter-

esting point here is that the chocolate mouse once pro-

duced, through the loss of its black-and-gray-pigment-

potentiality, can transmit no other pigment character but

the chocolate. What causes the occasional production of

an individual in which one or more of the characteristic

sorts of pigment is absolutely lacking is still unexplained.

Nevertheless it is of frequent occurrence not only among

domesticated species, in which the natural conditions of

life are so greatly modified, but also in species in a state

of nature.

A skunk normally marked, but chocolate instead of

black, a raccoon likewise of normal pattern but the pig-

mented areas yellow, are merely examples of the drop-

ping out of the factor for black pigment from the normal

combination of the two. Such specimens are of occa-

sional occurrence, and examples are in the museum of the

Boston Society of Natural History. Similarly are pro-

duced red woodchucks or muskrats, or wholly yellow field

No. 571] PATTERN DEVELOPMENT 389

mice (Microtus). Melanism commonly results through

an excess of black pigment which may mask a second

pigment. Thus the black hairs of ‘the black variety of

fancy mouse commonly contain a considerable amount of

chocolate pigment as well, and so of the hairs of the

black-appearing skunk. A black mouse thus does not

contain the yellow pigment, while the chocolate pigment

is largely masked in general view by the black. In other

cases it may be that black pigment alone is present.

It is probable that many cases of dichromatism among

animals are explicable as similar cases in which one or

other of the pigments normally present becomes to a

greater or less degree inactive. Thus red forms of certain

blackish or dull-colored bats (e. g., the small Molossus of

Cuba) are apparently the result of the dropping out of

the factor for black pigment or its great reduction. The

red and gray phases of the screech owl (Otus asio) are

probably also explicable as a similar phenomenon.

It is only when this inactivity of one or more of the

pigment factors occurs locally on the body that a definite

color pattern is produced, in which neighboring areas of

the body are of contrasting hues. As an example may be

cited the variegated guinea-pigs, whose monotone ances-

tors are still abundant in a wild state in South America.

Professor Castle, through his studies of these patterns

in guinea-pigs, first suggested to me in 1903 that there

were definite areas of the body which, though contiguous,

are independent of each other in their pigment-producing

capacity. In this suggestion lies the key to the chief

investigation of this paper, namely, the defining of these

areas, and a study of their behavior in the development of

pattern by the second of the two methods previously

given—that is, through the failure of pigment to develop,

so that white or colorless areas result. This condition of

partial albinism is not uncommon among animals which

in their normal condition are completely pigmented. In

domestic species it is very general and in them tends to be

preserved. It also occurs normally in the shape of defi-

390 THE AMERICAN NATURALIST [Vou. XLVI

nite white markings in the patterns of many mammals

and birds. Magazines of natural history abound with

instances of total or of partial albinism among mammals

and birds, either of domesticated or of wild species.

Some writers have even recognized the fact that such

white markings tend to occur in certain parts of the

body, as at the tip of the tail or on the forehead. Darwin

speaks of the white forehead spot or star, and the white

feet so common among horses, and implies that such

markings must be of some significance. His statement

on hearsay that white-marked horses are more suscep-

tible to poisoning from noxious herbs is, however, un-

corroborated. In 1882, W. H. Brewer gathered a number

of statistics as to the presence of white marks in horses

and cows, but reached no conclusion. He could find no

necessary correlation between the presence or absence

of white spots in forehead and feet, though it appeared

that white marks might be more frequent on one side of

the body than the other. But the tentative conclusion

that such animals habitually reclined on the side showing

the more white, is begging the question.

As briefly stated in my paper of 1904, the important

thing is not that white tends to appear at certain places,

but the converse, that pigment production is more intense

at certain definite centers on the body and the occurrence

of white or pigmentless areas is due to the restriction of

pigment formation at the periphery of these centers, so

that white occurs at their extremities or as breaks be-

tween contiguous color patches.

In mammals and birds these centers are typically five

on each side of the body, and a median one on the fore-

head. They appear to be homologous in both groups,

though in different species they show varying degrees of

modification in their behavior and development. When

a reduction of the pigment areas occurs, the appearance

is as it were a shrinking of the particular color patch

toward its definite center. The reduction may vary to

any degree, from that condition in which the break

No. 571] PATTERN DEVELOPMENT 391

between two adjacent patches is merely indicated by a

white streak to that in which it is reduced to a small spot

of pigment, or to zero, when the entire patch drops out,

leaving a white area. These patches are wholly independ-

ent of each other in the extent to which they may be

developed, so that a particular patch may be quite want-

ing on one side of the body, while its fellow of the oppo-

site side is completely developed. Nevertheless, there is

often a marked tendency to bilateral symmetry in such

reduction. From a study of partial albinos in which the

pigment reduction is considerable, the location of the

ultimate centers of these patches becomes possible as well

as the determination of their normal extent. I have

studied several domesticated species in which white

marks are common, with the results briefly detailed below.

When all the centers are fully developed the animal is

completely pigmented; when none is developed, it is a

total albino. Between these extremes may be found every

conceivable degree of development. In an ideal case in

which each center is slightly reduced so as to be circum-

scribed by white, the animal would have a dark coronal

or crown patch and a series of five patches on each side

separated by a median dorsal and a median ventral

stripe. The anteriormost of the lateral patches center at

the base of each ear, and each in its greatest development

covers the side of the head from muzzle to behind the

ear. These I have called the aural or ear patches; the

next posterior are the two neck or nuchal patches each

of which pigments its proper side of the neck, and extends

from behind the ear to the shoulder and anterior edge of

the foreleg. When much reduced the patch, as it were,

contracts to a small area on each side of the neck, varying

slightly in its location among different species. Posterior

to these come the scapular or shoulder patches one on

each side of the body. Each pigments the shoulder area

and foreleg, except (usually) the front edge of the upper

part of that member. This patch shows interesting slight

variations in the extent over which it spreads in different

392 THE AMERICAN NATURALIST (Von. XLVIII

species. Centering nearly at the lower part of the back

are the pleural or side patches, each of which pigments

the area from the shoulder to the lumbar region and ante-

rior part of the hind leg of either side. Last of all, the

two sacral or rump patches, each of which on its respect-

ive side pigments the buttocks and tail. In most species

these two patches are so closely associated that they tend

to remain fused dorsomedially, so as to give the appear-

ance, when reduced, of a single median patch at the base

of the tail. Their frequent bilaterality, however, indi-

cates the dual origin of such median patches. Each of

the lateral patches in its complete development extends

from the mid-dorsal to the mid-ventral line or those of

opposite sides may overlap slightly. Reduction usually

first appears mid-ventrally.

It is probable that the retinas should also be considered

as an additional pair of patches, since morphologically

the eye is of dermal origin, and there is sometimes seen

a tendency to the formation of a small cireumorbital

patch, which appears to break from the ear patch when

this is largely reduced. A

Pocock (1907) has pointed out that in black-and-tan

dogs the tan appears about the muzzle, along the sides

and on the limbs, while the blacker portions are more

dorsal. It may be added that in tricolor hounds, in which

the several primary patches are reduced, these are often

tan color at their several peripheries and black centrally.

In both cases, the explanation is simply that pigment

formation is less intense the farther away from the pri-

mary centers.

The reason of the division of the body surface into

these independent areas of pigmentation does not here

concern me. It is no doubt the result of physiological

causes, and it is rather suggestive that the several patches

correspond externally to important nerve centers or

groups of nerves. Thus the eye pigment corresponds to

the optic nerve, the aural patch to the auditory nerve, so

that these two great external sense organs of the head

have each their corresponding pigment patch. The neck

No. 571] PATTERN DEVELOPMENT 393

patch corresponds with the group of cervical nerves, the

shoulder patch with the brachial plexus, the side patch

with the nerves of the trunk, and the rump patch with the

sacral plexus. It may be further suggested that the

median crown patch of the head corresponds to the pineal

eye, a suggestion that is strengthened by the fact that it

is more or less obsolete in mammals, just as the pineal

gland is vestigial, whereas. in birds, which are more

reptilian in structure, the patch is usually well defined.

At all events it is a median unpaired structure, as are the

pineal and the interparietal bone.

Turning now to a more detailed consideration of these

pigment patches in sundry species of animals, we may

first examine a series of diagrams (Figs. 1-15) of the

j L

Wil

g q 10 t dd

Fics. 1-7. DIAGRAMS ILLUSTRATING PIGMENTATION IN THE Domestic Doc.

Fics. 8-15. DIAGRAMS ILLUSTRATING PIGMENTATION IN THE DOMESTIC Dog.

394 THE AMERICAN NATURALIST [Vou. XLVIII

domestic dog, all of which are carefully drawn from

photographs or from living animals, and are selected

from a great number to show various conditions in the

reduction of the pigment patches. In these and the other

diagrams the black portions represent pigmented areas,

irrespective of the actual colors.

For convenience I have called the white stripes demark-

ing these chief or primary, patches, ‘‘primary breaks,”

since they are the first indications of a decrease in pig-

mentation such that two adjoining patches no longer

meet. Secondary or further breaks result in a general

disintegration of these primary pigment patches, and are

apparently more irregular in nature, though often they

follow certain fairly well defined lines. The first of the

primary breaks generally occur as white patches on the

chest or belly, about in the median line. These are not

shown in the dagrams, but in most cases should be under-

stood as present. In Fig. 1 the pigment areas show a

beginning in reduction. The two aural patches have

become separated and their failure to spread to the

normal limit in the median line has resulted in a white

nose stripe. A short transverse white marking indicates

a separation of the neck patch at its anterior edge from

the ear patch. Elsewhere the various patches are contig-

uous; but the extremities of the limbs and tail are pig-

mentless, as if pigment had failed to spread to the tips of

these members in its reduction. In Fig. 2 the same

primary break between the ear patches is present, and in

dogs it is one of the first and most frequent to appear.

The same shrinkage of pigment from the extremities 1s

also seen. The neck patch of the left-hand side, however,

has completely dropped out, and its fellow of the right-

hand side is reduced posteriorly so that it fails to reach

the shoulder patch. Thus a white collar is formed. It

is also interesting to see that at its anterior end a distinct

constriction is present where the neck patch joins the ear

patch of the right side. Fig. 3 shows a somewhat similar

condition but the neck patch of the right side as well as

No. 571] PATTERN DEVELOPMENT 395

the ear patch is missing, while those of the left side are

fully developed. In Fig. 4 both neck patches are missing,

so that a white collar is formed. In dogs the neck patch

is usually the first to drop out altogether, so that a white-

collared dog is of very frequent occurrence. In fox

hounds this patch is shown unusually well, either wholly

or partly separated from neighboring patches. The sep-

aration of the ear patches, wholly or partially, so as to

produce a white blaze or line in the middle of the forehead

is about as frequent. In Figs. 8 and 12 a single neck spot

only (as it happens, in one on the right, in the other on

the left side) is still present but so slightly developed as

to be only a small island of pigment wholly separate from

the neighboring patches.

The crown spot is so often present in dogs as a little

oval island; always on the top of the head about in line

with the anterior bases of the ears (Fig. 4) that I am

convinced it is a primary patch. It is common in bull

dogs and bull terriers, and in other breeds is often seen

but is so commonly not indicated at all, that it seems

probable it is becoming lost, and its area is filled by the

ear patches, since these are often separated by a very

narrow median line only, which, as in Fig. 13, may con-

tinue posteriorly to separate the two neck patches

medially as well. In other cases (Figs. 1, 6) the failure of

the white nose stripe to extend farther posteriorly may

be due to the persistence of this patch.

The demarcation of the side from the rump patches is

indicated by the imperfect primary break across the

lower part of the back in Fig. 4, while in Fig. 5, a similar

primary break farther forward indicates the limits of the

shoulder and side patches. In each case the break is

incomplete transversely, with a narrow isthmus near the

median line. In dogs there is a marked tendency for the

ultimate centers of the side and rump patches to be close

to the median line, so that the corresponding patches of

opposite sides are confluent dorsally. This is especially

the case with the rump patches, with the result that it is

396 THE AMERICAN NATURALIST [Vou. XLVIII

very rare to see the two rump centers separated, but

instead, as in Figs. 10 and 14, they appear, when much

reduced, as a small median spot at the root of the

tail. That they were originally paired, there is no doubt,

as there is frequently (as in Fig. 9) a deep median

notch indicating the median primary break between the

centers, or (as in Figs. 11, 12) one of the lateral centers

drops out, leaving its fellow of the opposite side. The

` continued union of the side patches with the shoulder

patches is seen in Fig. 7, while in Fig. 9, though the union

is still present between these patches of the right side, on

the left side the shoulder patch has failed to develop, and

the side patch is so reduced that it does not meet its

fellow of the right. In Fig. 8 both shoulder patches are

present more or less bilaterally equal, and, as frequently,

are produced into narrow tongues on to the upper arm.

The two side patches in Fig. 8 are also reduced, so as to

be wholly separated from each other and from the neigh-

boring centers. They are further interesting in being

placed nearly median one behind the other instead of

nearly opposite. In Fig. 11, on the other hand, they are

far sundered, but this, in dogs, is a much less usual con-

dition. In Fig. 10 a single median dorsal patch repre-

sents the slightly developed side patches, but whether

this single patch corresponds to one or other of the two

centers, or whether the two are actually fused in the

dorsal line, I can not yet say.

The shoulder centers, when slightly reduced, are large

in dogs, and cover a considerable saddle-shaped area, as

indicated in Fig. 5, from near the center of the back for-

ward including the fore leg and part of the fore shoulder.

When further reduction takes place the pigment is drawn

away from the extremities and the saddle separates from

the neck patch (Figs. 2, 6) and then from the side patch

(Figs. 5, 9), and finally the shoulder patches separate

from each other (Fig. 8). One or other of the shoulder

patches may drop out entirely (Fig. 10) or be reduced to

a very small spot (Fig. 12) at what may be considered

No. 571] PATTERN DEVELOPMENT 397

the ultimate center of the pigment patch, near the upper

part of the body, near or just back of the shoulder. The

ear patches seem to be the last to disappear, and these,

too, may be variously reduced or only one may be present

(Fig. 15). The approximate outlines of the patches when

fully developed are indicated by dotted lines in Fig. 15,

in which 1 is the crown patch, 2 the ear patch, 3 the neck

patch, 4 the shoulder patch, 5 the side patch, and 6 the

rump patch.

In dogs, there is seldom seen any tendency for these

primary patches to divide. What has the appearance of

such a tendency is seen, for example, in the coach dog,

which is rather evenly flecked with rounded black spots,

with often in addition, black ears and more rarely reduced

rump patches. Fig. 9 shows such a dog in which both ear

patches, one shoulder, both side and both rump patches

are sharply indicated, though reduced. In addition there

are present on the white body areas between, many small

flecks of dark color, evenly distributed, which are clearly

not islands separated from the primary patches. Indeed

this spotting seems to constitute a wholly different cate-

gory of pigment formation, in addition to that of the

primary patches, which latter I have called ‘‘centripetal’’

pigmentation. As Professor Castle suggests to me, it is

probably homologous with the ‘‘English’’? marking or

spotted condition of domesticated rabbits, and possibly

the dappling of horses is a similar phenomenon. When

these spots and the primary color patches are of the same

hue, it is not possible to distinguish the two in visual

appearance, unless the latter are reduced areally, when,

as is sometimes the case in the coach dog, one or more of

the primary patches is seen with the spots, as it were,

proliferating from its edge. This second element no

doubt enters as a factor in the color pattern when the

small spots are of a different color from that of the

general body surface, as in case of the cheetah (Cyne-

lurus) or the leopard and jaguar.

I am inclined to think that the excessive breaking up of

398 THE AMERICAN NATURALIST [Vou. XLVIII

the primary patches, to be considered under the cow, is

not a wholly similar phenomenon.

Five diagrams illustrating the domestic cat are shown

in Figs. 16 to 20, and are interesting to contrast with

ik he I

17 26

Fics. 16-20. DIAGRAMS ILLUSTRATING PIGMENTATION IN THE DOMESTIC CAT.

those of the dog, also a carnivorous mammal. The

demarcation of the primary patches is usually less sharp

than in dogs, but is in general similar. The most.common

appearance is where the primary breaks occur in the

mid-line below, giving a white throat, chest or belly; or

the separation of the aural centers produces a white

streak on the nose or extends it up between the ears.

The ear patches in Figs. 17, 19, 20, show successive reduc-

tion, so that at first the hinder margin of the ears, as in

dogs, becomes white, then with further decrease in pig-

ment production, the inner bases only are colored. The

neck patch has its ultimate center farther back than in

dogs so that when much reduced, it is present as a pig-

mented spot at the very base of the neck or even at the

front of the shoulder (Figs. 16,17). In Fig. 16 the neck

patch of the right-hand side is only slightly reduced and

is in contact anteriorly with the ear patch, while poste-

No. 571] PATTERN DEVELOPMENT 399

riorly it does not meet the shoulder patch. The left-hand

neck patch, however, is quite separate from the neighbor-

ing patches and is reduced to.a small area at the junction

of the neck with the shoulder. It is absent in Fig. 17 from

the left side and is represented on the right side by a

similar small center, placed far back. In Fig. 20 the

neck patch or patches show a reduction to a single small

Square median patch at the base of the neck, but whether

this represents a median fusion of the two lateral centers,

or whether one only has persisted and has shifted to the

midline, I do not attempt to say, though the former

hypothesis seems on the whole more probable.

The shoulder patch in house cats is relatively small,

and, as indicated by the indentations in Figs. 17, 18, is of

the fore side of the upper arm, but the shoulder patch

when fully developed seems to cover the rest of the leg

and a small scapular area. It is shown much reduced in

Fig. 19, on the right-hand side, and is altogether wanting

in Fig 20. The conjoined shoulder and side patches in

Fig. 18 are shown reduced laterally, so as to form a

broad median stripe which I take to mean that the ulti-

mate centers are closely approximated dorsally. The

neck patch is wholly absent, but both ear patches are

present and joined medially. The sacral patches, as

commonly, seem fused or at least very close together.

There is a small break midway on the tail, which sepa-

rates off a pigmented tip, a phenomenon which I shall

refer to under ‘‘centrifugal pigmentation.’’ The side

patch is long comparatively, and extends forward to

cover the deficiencies of the shoulder patch, as in Fig. 17.

Here the left side patch has been reduced at its anterior

end, and its separateness from the patch of the right side

is indicated by the median indentations. It is often want-

ing in domesticated cats. o

The sacral patches, pigmenting the buttocks and tail,

seem to be fused or closely approximated at the root of

the tail, as in dogs. I have seen no instance of the crown

patch being shown in the cat, though such may occur.. |

400 THE AMERICAN NATURALIST [Vou. XLVILE

The approximate boundaries of the five bilateral patches

are indicated in Fig. 20 by dotted lines; 2 is the ear patch,

3 the neck patch, 4 the shoulder patch, 5 the side patch,

and 6 the rump patch.

Among domesticated rodents the pigment patches have

been studied in rats, house mice, and guinea-pigs. In

all, the same patches appear except that in rats and mice

the median crown patch appears to be lost, though in the

guinea-pig it is often present. Diagrams of parti-colored

mice are shown in Figs. 21-24, and sufficiently indicate

2a 23

Fies, 21-24. DIAGRAMS ILLUSTRATING PIGMENTATION IN DOMESTIC VARIETIES

OF THE HOUSE MOUSE,

the primary pigment areas. The white spot on the fore-

head of Fig. 21 indicates a primary break between the

two ear patches, and varies widely in different individ-

uals, from a few white hairs only to a large blaze The

inheritance of such a blaze has been studied by Little

(1914). The white mark at the base of the neck in Fig.

21 indicates the beginning of separation of the neck from

the shoulder patches and perhaps of the two neck patches

from each other, because of its longitudinal extension.

The white. band across the neck in Fig. 23, however,

indicates probably only the beginning of a separation of

the neck from the shoulder patches, which in Fig. 24 has

No. 571] PATTERN DEVELOPMENT 401

wholly sundered these two areas, so that a white-collared

mouse results. The condition shown in Fig. 22 is similar,

except that the separation has taken place on the right

side only, between the neck and the shoulder patches of

but one half of the body. A break between the two neck

patches of opposite sides is further indicated in this

figure by the deep median reentrant back of the ears.

In all four diagrams the areal restriction of the

shoulder patches is shown, but in varying degrees. In

Fig. 21, the pigment has not spread to the feet, leaving

these white, and so in the other figures, but to a greater

degree. A median linear break between the shoulders

indicates the restriction of the patches of opposite sides

at this point, which in Fig. 22 is more clearly perceptible.

The posterior limits of the shoulder patch are further

shown in this diagram, by the beginnings of a break

between the shoulder and the side patches. In Fig. 24

this break is no longer interrupted, but clearly separates

the two areas. Further, the side patch has dropped out

on the left. In Fig. 23 an imperfect separation of

patches on the posterior part of the body has taken place.

On the right-hand side the shoulder patch, which in mice

is of considerable extent, has broadly separated from the

side patch, while on the left-hand side a long transverse

break has taken place between the side and the rump

patches, with two island-like white spots between, the

anterior of which probably marks the transverse line of

stress between shoulder and side patches, the posterior

the median line of breaking between the two side patches.

A slight indentation in the pigmented area far back on

the right side of Fig. 22 points to the beginning of restric-

tion between side patch and rump patch. The separation

of these patches by a transverse mid-dorsal break is

shown in Fig. 21, and their complete separation on the

left side appears in Fig. 23 (the transverse white mark),

while in Fig. 24, owing to the failure of the left-hand

pleural patch to develop, the two rump patches, both par-

tially separate from each other, are wholly disconnected

402 THE AMERICAN NATURALIST [Vou. XLVIII

from the former except by a narrow isthmus on the right

side. The long tail is usually without pigment, or mainly

so Where areal restriction is present, and it is seldom that

pigment extends far on to the base of this member when

the restrictive tendency appears. In the domesticated

varieties of rats, the same patches may be distinguished.

There is, however, an interesting variety known as the

‘‘hooded” rat, in which the ear and neck patches appear

to be normal, but a narrow median dorsal area is pig-

mented for a varying length, sometimes quite to the root

of the tail. A separate factor seems here to be involved,

producing what may be called a ‘‘centrifugal’’ type of

pigmentation, which in many forms of mammals causes a

black spine stripe (Sorex wardi, Tupaia tana, certain

forms of Apodemus, Equus caballus), and others.

Among guinea-pigs the typical primary patches are

beautifully shown and may be seen in sundry figures

published in papers by Professor Castle on heredity in

this animal. The guinea-pig is one of the few mammals

yet known in which the median crown patch is visibly

present, a character which I take to be primitive.

In guinea-pigs the breaking up of the ticked color

pattern has progressed under long domestication to an

extraordinary degree, so that not only are black, tawny

or grizzled animals produced in various shades, but even

in the same individual, the different primary pigment

areas may be of different colors. This fact is of much

significance, for it indicates not only the mutual independ-

ence of the contiguous color areas, but further points to

the manner in which a variegated color pattern may have

been acquired. Among mammals the color pattern is in

general, not greatly developed in comparison with birds,

yet in many cases where some modification has taken

place, it is evident that this differentiation is confined to

the limits of one or two of the primary pigment patches.

Thus in the South American Tayra (T. barbara), the

head and neck are a grizzled gray, and the breaks occur-

ring in pied individuals show that the grizzled condition

No. 571] PATTERN DEVELOPMENT 403

FIGs, 25-32. DIAGRAMS ILLUSTRATING PIGMENTATION IN HORSES,

404 THE AMERICAN NATURALIST [Vou. XLVIII

is confined to the aural and nuchal patches only, for else-

where the animal is black. In this case, too, the black

condition is probably derived, for youngish animals are

uniformly grizzled, and sometimes, apparently, this is

the adult condition as well.

Among domesticated ungulates the same primary

patches are to be distinguished in cases where partial

albinism renders their bounds apparent, with the excep-

tion that in horses, cows and deer I have seen no clear

indication of the median crown patch which in mammals

is probably obsolescent.

In both horses and cows the patches show interesting

and peculiar modifications. A series of diagrams (Figs.

25 to 32) show these patches in ‘‘calico’’ horses, though

not so fully as could be wished. The first indications of

areal restriction of pigment in horses appear in the shape

of a white ‘‘star’’ or round spot in the center of the fore-

head. This is often accompanied by white at the base of

the hoofs, or sometimes the entire foot is white producing

the so-called ‘‘white stockings.’’ But there is no neces-

sary correlation between these white areas, such as

Brewer (1882) tried to show. The white on the forehead

may vary from a few white hairs to a broad blaze cover-

ing the entire front of the head between the eyes to the

muzzle. Sometimes the restriction of pigment is such as

to produce in addition to the white star on the forehead,

a white spot over each eye, and sometimes these three

spots are joined by a narrow unpigmented area. This

indicates that pigment production is weak at a spot

directly over the eye in comparison with neighboring

parts, and this no doubt accounts for the fact that

in black-and-tan or other dogs these are the pale spots

over the eyes where black pigment is not produced.

A white spot over the eye is also characteristic of many

rodents.

Next after the restriction of the ear patches and the

drawing away of pigment from the feet, the most common

No. 571] PATTERN DEVELOPMENT 405

white marking seems to be a primary break, as in Fig. 25,

from the shoulder back of the foreleg, which delimits the

posterior border of the shoulder patch. In the horse the

shoulder patch is large, and differs from that of any

mammal I have yet studied, in its great extent forward

along the dorsal side of the neck nearly to the head. In

Fig. 26 a small break at the back of the neck indicates the

beginning of separation between the ear and the neck

patches dorsally, and a long tongue of white running up-

ward from the forearm indicates the anterior limit of the

shoulder patch. This limit is marked still nearer the

dorsal line in Fig. 27 by a white spot on the side of the

` neck near its base. In Fig. 28 the shoulder patch has

entirely dropped out and the white space outlines very

nearly its extent. The ultimate center is perhaps shown

by the small shoulder spot in Fig. 31.

The area covered by the ear patches extends well on to

the upper part of the neck, and in Fig. 29 is shown at its

greatest spread, or, as in Fig. 28, cut off by anarrow white

collar from the neck patch. The neck patch is remarkable

from the fact that in its areal reduction it becomes re-

stricted first dorsally, and the ultimate center of each side

is nearly ventral on the throat, so that, as generally seen,

the two centers form a single median patch on the front

or ventral part of the throat. In Fig. 26 the neck patch is

seen to pigment the anterior side of the forearm and is

partly separated from the shoulder patch by a long

tongue of white. It seems to extend up diagonally to

reach the mid-line of the neck for a short distance only,

as indicated in Fig. 28, where its bounds are only slightly

contracted. In Fig. 29 it is so far lessened as to be absent

from the forearm, though still in contact at the throat

with the ear patch where, however, a deep indentation

locates the dividing line between the two patches. In

Fig. 31 a median ventral division of the conjoined neck

patches is seen indicated at the upper part of the area,

which in this case no longer reaches the ear patches.

Still further reduction of both ear patches and neck

406 THE AMERICAN NATURALIST [Vou. XLVIII

patches is seen in Fig. 32, but, as commonly, the neck

patches seem fused in the midventral line. This shifting

of the neck centers ventrally is a rather remarkable

phenomenon which may have some relation to the manner

in which the head is held erect. For this reason it might

be expected also in antelopes, and is perhaps evidenced

in such a species as the oryx, in which there is a black

median line on the throat as though strongest pigment

production centered there rather than on the gray sides

of the neck. The median reduction of the shoulder

patches in horses is sometimes indicated by a white mane.

The rump patches in the horse appear to be much as in

other mammals, restricted to the tail and posterior part’

of the buttocks and the entire foot. In Fig. 30 the patch

is shown at nearly its full development, except that it has

failed to extend to the entire hind foot. In Fig. 27 it has

drawn away still farther but remains in contact with the

side patch at one place. In Fig. 29 it is further restricted

to the tail and posterior border of the haunches, while in

Fig. 32 it covers only the root of the tail and that member.

The side patch is the largest of all and extends from

the shoulder to the fore part of the haunches and on to the

fore part of the hind leg nearly to the foot, as seen in Figs.

27 and 28, where it is still in contact with the rump patch,

or in Fig. 29 where it has become separated. In its fur-

ther reduction this patch may appear as a small spot

back of the ribs or, as often, a curious division takes

place, separating the patch into a dorsal area and a

lateral one. Occasionally this secondary break appears

in a horse which has most of its patches otherwise well

developed. In Fig. 31, the pigmented area of the tail,

buttocks and lumbar region consists of the conjoined rump

patch and a dorsal portion of the side patch, while the

ventral part of the side patch is present as the oval spot

at the groin. In Fig. 30 the latter spot only persists, but

in Fig. 32 the dorsal portion of the side patch alone is

present as a stripe along the entire back, except where it

breaks away posteriorly from the small rump patch.

No. 571] PATTERN DEVELOPMENT 407

This peculiarity of the side patch in horses is somewhat

paralleled in cows by a tendency to secondary breaking

up, though in a different way, as detailed below. It is

significant in this connection that in horses and donkeys

there is usually a black stripe along the spine from

shoulder to tail which may indicate that ‘‘centrifugal pig-

mentation’’ is also present (see beyond). The dotted

lines in Fig. 32 indicate the approximate boundaries of

the several primary patches. The crown patch seems to

be wanting in horses; 2 is the ear patch, 3 the neck patch,

4, 5 and 6 the shoulder, side, and rump patches,

respectively. : .

Of domestic ruminants I have studied the pigmentation

in the cow and show in Figs. 33 to 42 a few of the many

Fics. 33-36. DIAGRAMS ILLUSTRATING PIGMENTATION IN DOMESTIC Cows, SIDE

VIEW.

variations in partial pigmentation. These are all drawn

from photographs or from the animals themselves, and

are of cows in which, so far as I know, there has been no

attempt at breeding for pattern. Two types of spotting

may be distinguished in cows: first, that in which the pig-

mented areas are sharply outlined and solid or at least

408 THE AMERICAN NATURALIST [Vou. XLVIII

practically so; second, that in which there is a greater or

less tendency for the primary patches to be much broken |

up into small islands (as in Fig. 36) by secondary breaks,

though the main areas are still distinguishable. I take

this second or fragmental type to be a different phenom-

enon from the diffuse or dappled condition seen in the

coach dog or the dappled-gray horse.

In the cow, the ear patches as usual pigment each its

proper side of the head to a short distance behind the

Hics. 37-42. DIAGRAMS ILLUSTRATING PIGMENTATION IN Domestic Cows, AS

SEEN SPREAD OUT AND FROM ABOVE.

ears. The point of separation between ear patches and

neck patches is indicated by a small break back of the

skull in Fig. 38, while the posterior extent is shown by

the two ear patches in Fig. 42. These patches usually

No. 571] PATTERN DEVELOPMENT 409

draw apart first across the forehead making here a tri-

angular white mark, and on the muzzle, as in Fig. 34.

Further restriction broadens these white marks and joins

them by a narrow isthmus as in Fig. 35. In Fig. 40, the

two patches are still conjoined across the vertex, but are

much reduced, that of the right side more than that of the

left. In Fig. 42 they have failed to join medially, though

fairly well developed longitudinally. Still greater re-

duction, as in Fig. 37, confines them to the ears, the bases

of which appear to be the ultimate centers.

The neck patch in the cow is more extended posteriorly

than in the horse, and its center is strictly lateral rather

than nearly ventral. It is shown in Fig. 34 somewhat

contracted from the mid-line of the throat, but extends

squarely back against the foreshoulder at the base of the

neck, and is fused near its ventral corner with the

small shoulder patch, itself much reduced. As in

other mammals it appears to extend in its complete

development, to the front edge of the upper foreleg. The

animal in Fig. 41 shows a bilaterality in its pigmentation

that is rather unusual. What appear to be the reduced

neck patches are seen far back at the border of the fore-

shoulder. In Fig. 40 the left-hand neck patch has

dropped out, but that of the right side is still present,

though small, and in Fig. 42 it is reduced to a small spot

only.

The shoulder patch in cows is remarkably narrow,

and compressed between the neck patch and the body

patch, whence it extends as usual on to the foreleg. In

Fig. 33 a primary break back of the foreshoulder marks

the nearly vertical posterior outline of the shoulder

patch. In Fig. 34 the separation of this area from the

neck patch is all but complete and the patch itself some-

what reduced. Its narrow vertical outline is thus indi-

cated, as well as in Fig. 39, in which there is a narrow

tongue-like extension down on to the center of the foreleg.

In its further reduction it appears as a small center

at the base of the scapula, as in Fig. 35, or in Fig. 40, in

410 THE AMERICAN NATURALIST [Vou. XLVIII

which both shoulder patches are present, though small.

In Figs. 41 and 42 the shoulder areas are wanting. A

very common mark in cows is a white belt just back of

the foreleg. This is due to the development of a primary

break between shoulder patches and side patches, a con-

dition which is nearly realized in Figs. 33 and 38. It is

probable that this marking has been more or less fixed

through selection in breeding, and this has been the more

readily accomplished, since this break occurs in a place

which is one of the first in cows to cease pigment

production.

The side patch is large and covers the entire lateral

region of the body from the scapula to the hips, and on

to the front edge of the hind limb. When only slightly

reduced, it appears as a blanket-shaped area across the

back as in Fig. 38, where it has not wholly broken away

from the shoulder and rump patches, or as in Fig. 33,

where it has become nearly separated. In its further

reduction this dorsal blanket shows a peculiar manner of

breaking up into more or less transverse stripes directed

slightly backward. The beginnings of these secondary

breaks appear in Fig. 39 in which are seen on each side

posteriorly two deep indentations at the edge of the

patch, whose points if extended would meet the white

pigmentless islands already present within the patch.

In Fig. 34 a similar series of indentations points to the

trisection of the side patch which is realized in Fig. 35.

Here is a characteristic which if developed might even-

tually result in the actual production of white stripes on

the body, such as are found, for example, in certain ante-

lopes as the bongo and the kudu. The tendency of the

side patch to divide into three, as in these diagrams, is

rather marked in cows, and even with further reduction

the three centers persist fairly well. The first of these

secondary centers is just back of the shoulder patch, the

second about over the last ribs, and the third over the

lumbar region. In Fig. 40 the first two are present on the

left side, with a small spot between, which has become

No. 571] PATTERN DEVELOPMENT 41]

separated from one or the other of them, while the third

or lumbar spot has dropped out. On the right side, the

first and second divisions are still fused dorsally, but

the lumbar division is distinct. The same three divisions

are seen in Fig. 35, better developed, whereas in Fig. 42,

the two lumbars are present, one on each side, and con-

siderably in advance of them, what seem to be the rem-

nants of the first division of the side area, the left one of

which has further broken up.

The rump patches show no especial peculiarities, but

cover the posterior part of the buttocks and hind legs,

and the entire feet and tail. Though frequently the two

patches of opposite sides are conjoined medially, they are

often, under considerable reduction, well separated. The

beginning of such a separation appears in Fig. 38, where

there is a deep median tongue of white anteriorly, mark-

ing the line of union. In Fig. 41 the reduction has pro-

gressed still farther so that the two patches are quite sun-

dered medially and do not extend to the tail. In Fig. 40

the patch of the left side has become inactive, and that of

the right side is small.

A curious condition not infrequently seen is shown in

Fig. 37, in which all the patches are present, but those of

the right side are separated from those of the left by a

median dorsal white line, showing the distinct bilaterality

of these pigment areas. In the figure, the ear patches

are so restricted as not to reach the neck patches of

their respective sides, the shoulder patches do not extend

far on the forelegs, the side patches are reduced ven-

trally, and the rump patches, though in contact with the

side patches, do not pigment the tail or extremities of

the legs. A further reduction of pigment areas results in

Fig. 41, in which the paired centers of neck, side and

rump patches still appear.

The diffuse condition of pigmentation is illustrated in

Fig. 36, which is a photograph, inked in. The ear patch

is seen much reduced, but pigmenting the ear. The neck

patch is of most irregular shape, with several subsidiary

412 THE AMERICAN NATURALIST [Vou XLVIII

spots separated from its lower border. A clear line

separates the neck patch from the shoulder patch, which

is also of most irregular boundary. The side patch, at

its fore part, is broken into a series of small islands

which tend to arrange themselves in lines following the

direction of the ribs. The main part of the patch shows

a decided tendency to break into the usual three or per-

haps four portions. It is common for cows to have

patches with very irregular boundaries and tongues of

pigment, which may break off into isolated spots in a

most bewildering fashion, but even in such cases it is

possible to distinguish the main patches of which these

form part.

White patches occur in other domesticated ungulates

as the pig, the llama, the alpaca, the camel, the yak, the

reindeer, and the goat. In the water-buffalo, occasional

animals seen in Egypt show a beginning of pigment re-

duction through the presence of white in the forehead or

on the tail. I have had no opportunity to study the mark-

ings of these species.

(To be concluded)

. INTERNAL RELATIONS OF TERRESTRIAL

ASSOCIATIONS

ARTHUR G. VESTAL

UNIVERSITY OF COLORADO

CONTENTS

I. Introduction,

II. Internal activities of the association, as determined by the con-

stitution of the individual organism.

e organism.

- Ecolo

B. Farnar of the plant in relation to environment.

Constitution of the ppap: s relation to environment.

Internal annie of the

III, Relative spe of different segues with the association——domi-

nan

A. enon of dominance e animals.

iteria of dominance among animals.

C. Spe cialized and unspecialized animals.

IV. a ee in the association

Ai 4 in 5 ace.

ribution me.

Va Tatevdependencs of atrak trial plant and animal communities.

Aces rela spies of At ined “agen and animals.

eographie range: the pro

£ Distsbation apre the province: distribution of plants

and an ane communitie

B. Local relations Re plant and kaia assemblages (relations

n the esl

an Similarity of ecological type of plants and animals.

2. Relative hes smog ce of plant and animal a ssemblages.

3. Correspondence in distribution within the sane ciation.

4. Un porns ic of species composition of plant and animal

VE oe pate pres aor

VII. References

I. INTRODUCTION

THE material here presented is based on the writer’s

studies, during the past five years, of terrestrial associa-

tions of plants and animals, mainly in different parts of

the prairie region. The particular area chiefly used for

illustration in this paper is the sand prairie of the Illinois

River valley, plants and animals of which have been

studied by Hart and Gleason (1907) and by the writer

(1913b). A later study has been made of the vegetation

of inland sand areas of Illinois (Gleason, 1910) ; the Lake

Michigan beach area in northeastern Illinois has been

studied by Gates (1912) ; beach areas in Illinois and Indi-

ana by the writer (1914a). The chief representation of

413

414 THE AMERICAN NATURALIST [Vou. XLVI

the sand prairie is the bunch-grass association, well-

developed in parts of northwestern, central and north-

eastern Illinois, and in northwestern Indiana, in each

of which areas, as well as in the sandhills of Nebraska

and of eastern Colorado, the writer has studied. Discus-

sions of physical, vegetational and animal aspects of the

associations of the central Illinois sand prairie, together

with an annotated list of the animal species, with data on

food, habitat-relations, life-history, ete., are embodied in

the writer’s paper (1913b), to which constant reference

is made. Frequent citations to a more detailed study of

local distribution of grasshoppers, in a Michigan area

(Vestal, 1913a), and to the many associational studies

of Shelford, are to be found.

The data wal have accumulated relate nearly equally

to the botanical and zoological aspects of associational

study, but since the subject of plant ecology is at present

more advanced than that of animal ecology, it has been

possible to treat the vegetational side of the problem

very briefly, so that more of the discussion relates to

animals and animal assemblages.

The writings most frequently cited are indicated by

italic capitals, the full titles appearing in the list of spe-

cial references at the end of the paper.

The writer wishes to thank Dr. Charles C. Adams, Dr.

Max M. Ellis and Dr. H. A. Gleason for suggestions and

criticism.

II. INTERNAL ACTIVITIES OF THE ASSOCIATION, AS DETER-

INED BY THE CONSTITUTION OF THE

INDIVIDUAL ORGANISM

The internal activities of the association may be said

to be the sum-total of the activities of all the plants and

all the animals which make up the association. Such a

sum-total of activities may well be thought of as an intri-

cate and complicated mass of dependencies. It will

simplify the treatment of the entire system of relations

if the chief dependencies of the individual organism are

first discussed. A knowledge of the ecology of the asso-

No. 571] TERRESTRIAL ASSOCIATIONS 415

ciation is built up largely from a knowledge of the ecol-

ogy of all the organisms which compose it.

A. Ecouocican CONSTITUTION OF THE ORGANISM

The constitution of the organism is the sum-total of

those of its characters which enter into relation with

environment. These are commonly classified as structural

and physiological. For the purposes of this discussion

it would seem preferable to subdivide physiological char-

acters, restricting the term physiological to denote those

characters concerned with ordinary metabolic processes

of the organism, and excluding those having to do with

life-history and rates of reproduction (these may be dis-

tinguished as biographical and numerical') and also, when

dealing with animals, those related to behavior (psycho-

logical characters). The constitution of the organism in

relation to environment will be discussed in terms of

these classes of characters.

B. CONSTITUTION oF THE PLANT IN RELATION TO

ENVIRONMENT

-~ The environmental influences in the association are of

three kinds: (1) physical, (2) plant, (3) animal. Each

plant and each animal must obtain from each of these

three constituents of its environment certain necessaries;

it has certain structural and physiological characters

which enable it to obtain these necessaries, and to with-

stand adverse environmental influences.

The environmental relations of plants are very differ-

ent from those of animals. A tabular comparison of

these relations has been made by Shelford (4: 593). As

therein pointed out, structural characters are of greatest

importance in the adjustment of the plant to the environ-

ment, and plants in a given habitat are likely to have a

common structure or growth-form, indicating common or

ecologically equivalent physiological conditions within.

Different plants (and different animals), within a com-

1 Based partly on Forbes’, classification of adaptation to food require-

ments (1909: 292).

416 THE AMERICAN NATURALIST [Vou. XLVII

mon habitat, are similar in ecological constitution (eco-

logically equivalent) in so far as their presence is deter-

mined by the same environmental conditions. It should

be pointed out that there are local environmental differ-

ences within the area of the association which allow the

presence of differently constituted organisms, and that

the entire range of environmental conditions within the

habitat is usually much wider than that of the environ-

mental complex selected by a particular organism. The

environmental complex of the organism is not the same

as the sum-total of environmental conditions within the

association. Each organism differs in greater or less

degree from others in ecological constitution, and thus

selects a different environmental complex.

The physical factors of the environment are of great-

est importance in the life of the plant. Plants influence

one another directly to only a slight extent. There is

usually very little of the social relationship among eco-

logically similar plants which will compare with such

relationships as seen in animals. Competition among

plants is mainly a struggle to determine which plants

are to be most favored by physical conditions, and it is

probably most severe for the physical factor present in

minimal quantity. In desert associations plant competi-

tion is almost exclusively for water, and extensive root

systems are developed. In grassland it is very largely

for above-ground space; in forests it is principally for

light. The influence of the animal-environment is prob-

ably of greater importance than has commonly been

realized by plant ecologists; the study of economic ento-

mology and of the effects of grazing upon grasslands is

helping to bring about a realization of the importance of

animal influence upon plant life.

The structures of plants show frequent and great

modification in response to the physical conditions of the

environment. These modifications are most frequent and

important with respect to the factor present in minimal

quantity. Characters which may be associated with

No. 571] TERRESTRIAL ASSOCIATIONS 417

direct plant influence are infrequent. Certain plants

which become more abundant as a result of close grazing

are equipped with spines, or have acrid or pungent

juices; and many other characters may be correlated

with animal influence. The structural modifications are

most evident in the adjustment of the plant to external

conditions, though these are accompanied by physiolog-

ical pharaetens which are also in harmony with the en-

vironment.

C. CONSTITUTION oF THE ANIMAL IN RELATION TO

ENVIRONMENT

The animal, like the plant, selects an environmental

complex which is of three kinds: (1) physical, (2) plant,

(3) animal. Different animals show extreme variation

as to the degree in which the different parts of the en-

vironment are important to their existence. Endopara-

sites, for example, are most directly concerned with the

animal part of their environmental complex.

The existence of any animal is dependent upon a num-

ber of physical factors, all of which must be present in

proper degree or quantity. Minimal and maximal quan-

tities of any one of several factors mark the limits of

existence of any animal (A: 598—law of toleration of

physical factors). It is not necessary to consider these

factors in detail. The animal reacts to physical environ-

ment most evidently by its behavior: psychological char-

acters restrict activities more narrowly than do those of

other types. They are accompanied by structural and

physiological characters; hibernation, storage of food,

etc., are biographical characters correlated with seasonal

changes in physical environment. Animals which are

subjected to very severe physical conditions may produce

a larger number of offspring than those to which physical

conditions are favorable. This is an example of corre-

lation of a numerical character with the physical environ-

ment,

The plant environment reacts upon and modifies phys-

ical and animal environments, and has also direct influ.

418- THE AMERICAN NATURALIST. [Vou. XLVII

ence upon the animal. In addition to its effect in the con-

trol of temperature, light and other physical factors, the

vegetation ‘constitutes the basic food-supply for the ani-

mal community, and also provides shelter and materials

for abode (A: 601). Cases of direct association between

particular plants and particular animals are numerous,

but the majority of animals have no direct relation to

particular kinds of plants. Behavior characters are in

general of greater importance in the relation of the

animal to the plant environment, though such relations

are not confined to psychological characters.

There are two sets of relations between the animal and

its animal environment. These are: (1) social, and (2)

antagonistic. Social relations (inter-psychology and

inter-physiology of Shelford, A: 608, b) include those

between individuals of the same species, and between

animals of the same or similar mores? (ecologically equiv-

alent animals), in so far as these relations are not

antagonistic. Breeding and family relations are the

principal activities which come under this head. Be-

havior characters are of greatest importance, as compared

with structural and other characters. The antagonistic

relations constitute the intermores-psychology and phys-

iology of Shelford (A: 608, c). They are the antagonistic

relations between animals not ecologically equivalent, and

they are also antagonistic relations within a species and

between ecologically similar forms. These relations are

probably not greatly concerned with reproduction, but

center about the feeding activities of the animal. The

existence of the individual animal, in its relation to other

organisms, is dependent upon three conditions: (1) it

must obtain suitable and sufficient food, (2) it must be

free from destructive competition of animals of similar

requirements, (3) it must be able to escape or to with-

stand attacks of other animals (or, sometimes, of para-

sitic fungi or bacteria). The various characters of the

2 Mores (Latin for customs, habits) has been used by Shelford (1911a:

30) to supply the need for a term including all physiological and behavior

characters of the animal.

No. 571] TERRESTRIAL ASSOCIATIONS 419

animals are correlated with all three of these conditions.

The characters are both ‘‘adaptive’’ (fixed by heredity),

and regulatory (not fixed).

Following is a synopsis of correlations between the

various types of characters and the three conditions of

existence, in the relation of the animal to its antagonistic

animal environment.

(I) Characters Which Enable the Animal to Obtain Food

1. Structural Characters—Animals of selective food-

habits often have specialized structures, as in the case of

the long tongue of woodpeckers. Animals of non-selec-

tive food-habits have mouthparts that are not so highly

specialized; thus grasshoppers and cutworms have heavy

mandibles for cutting vegetation; tiger-beetles and

Chrysopa larve have sharp piercing mandibles. The

whole structure of the predaceous animal, its ‘‘action

system,’’ is sometimes suggestive of the manner of pur-

suit or holding of its prey.

` 2. Physiological Characters—The physiology of ani-

mals of different food-habits differs materially. Physio-

logical characters are not apparent, generally speaking,

and are secondary to psychological characters. The

range of food assimilable by the animal is usually much

wider than that selected by it, as is seen when animals of

selective habits take new kinds of food when the usual

food is exhausted, often thriving seemingly as well as

before.

3. Psychological Characters.—Selection of food is

determined chiefly by behavior characters of the animal.

These may be so widely variable that the animal will be

virtually omnivorous, as in the case of crickets, or so

narrowly restricted that it eats only a single species of

plant or animal, as the leaf-beetle Blepharida, a sand-

prairie insect eating leaves of the three-lobed sumac, and

the pentatomid bug, Perillus, which feeds on Blepharida

(cf. E: 49, 30). Selection is only one of the many psycho-

logical characters relating to food. The behavior cnar-

acters manifested in obtaining food are of great variety.

420 THE AMERICAN NATURALIST (Vor. XLVII

With these are accompanying structural and physiolog-

ical characters, which, however, play a subordinate part.

4. Biographical Characters—These may consist in

timing the life-history of the animal with that of the

food-species (plant or animal) in such a way that the

period of greatest activity of the former coincides with

the period of greatest growth or abundance of the latter.

This feature may be incidental to seasonal change of

physical environment. Whatever its cause, it is very

general in an established association, so general that it

is seldom recognized. It is of advantage to both animal

and food species.

5. Numerical Characters—The rate of reproduction

must be so adjusted to its food-supply (plant or animal)

‘that only the unessential surplus of this food shall be

appropriated, leaving the essential maximum product

undiminished’? (Forbes, 1909: 293). Species of re-

stricted food-habits must remain less numerous in indi-

viduals than general feeders, as the available food-supply |

is very much less.

(II) Characters Which Remove the Animal from the

ompetition of Other Forms

1. Structural Characters.—Structures which permit

animals to live in varied habitats, to take varied foods,

or to time their activities differently, remove each group

of animals from competition of all the others, resulting

in advantage to all. To that extent the fossorial forelegs

of the mole, the long proboscis of the butterfly, and modi-

fications of the eyes of nocturnal animals, are characters

which do away with competition. The structural char-

acters are, however, accompaniments of modifications of

behavior, and are secondary to the latter.

2. Physiological Characters.—Ability to digest food-

materials unavailable to other animals is an advantage-

ous physiological character. Thus the leaf-beetle Chry-

sochus auratus, which lives on doghane (Apocynum),

and the ‘‘skin-beetle’’ Tros, which eats animal tissues

in an advanced stage of decomposition, have few com-

No. 571] TERRESTRIAL ASSOCIATIONS 421

petitors for food. Physiological, as well as structural,

characters, are accompaniments to modifications of habit.

3. Psychological Characters—Apparent preference

for certain activities, certain habitats, or certain foods,

together with peculiar behavior complexes, seem to be of

greater importance in removing animals from competi-

tion than structural and physiological characters.

Highly regulatory habits permit certain animals to ad-

just themselves to changing conditions of competition.

4. Biographical Characters.—Professor Forbes (1909:

295-298) discusses the alternative timing of the active

period among close competitors for food. (It so happens

that the animals mentioned, having almost identical

habits, compete with each other in many ways, besides

with respect to food.) In the sand prairie it has been

found that different species of certain genera, having

otherwise the same habits, differ greatly in life-history.

Evidence of this biographical adjustment is more or less

complete for two species of Arphia (E: 21), two or three

species of Hippiscus (E : 21), two species of the milkweed

beetle, Tetraopes (E: 47), and three species of Procta-

canthus, robber-flies (E: 55). In these genera the term

of activity of one species is abruptly followed by that of

another, the successive periods usually covering most of

the summer season.

5. Numerical Characters—When a certain limited

food, place of abode, or other desideratum is used by two

or more kinds of animals at one time, a numerical adjust-

ment is likely to be found among these competing species.

The rate of multiplication of each species must be suffi-

cient to keep up its numbers, to allow it to hold place

with competing species. (Too high rates of multiplica-

tion, on the other hand, are disadvantageous because of

other influences.)

(III) Protective, Defensive and Concealing Characters

1. Structural Characters—Animals have various de-

fensive, protective and concealing structures. Stings,

beaks, mandibles, teeth, claws, hairs, spines, resemblance

422 THE AMERICAN NATURALIST [Vou. XLVIII

to surroundings in color or form—all are of advantage to

animals which possess them. Certain of the interstitial

or blowsand animals resemble in color the sand on which

they rest (Cicindela lepida, Stachyocnemis, Psinidia,

Spharagemon; cf. E).

2. Physiological Characters—Malodorous and ill-

tasting animals are to a considerable degree exempt from

attack. This is essentially a physiological modification,

though a structural basis in the form of glands may be

‘present. In the sand prairie Chrysopa (lace-winged fly),

a number of Hemiptera, ladybird beetles, soldier bugs

(Chauliognathus), blister-beetles (Epicauta), and others,

are ill-tasting (perhaps not to some animals). The

skunk’s lack of caution is well known.

3. Psychological Characters.—Self-preservation in

animals depends more upon their activities and behavior

than upon special structures. The ordinary methods of

resisting or evading attacks of enemies are generally

known and need not be discussed. Many specialized in-

stincts have arisen, such as feigning death, or dropping

to the ground when disturbed, as seen in many herbi-

colous beetles.

4. Biographical Characters. —It is to the advantage of

animal species preyed upon by others if their period of

greatest abundance is timed with the period of greatest

activity of the animals which feed upon them.

5. Numerical Characters.—Animals, as well as plants,

must produce a normal excess in numbers which will pro-

vide food for other animals and still leave a sufficient

number of individuals to continue the species.

It will be noted that the various kinds of characters

usually accompany.one another, all being parts of a

single modification. This modification may have rela-

tion to one or to several of the environmental influences

(physical, plant or animal) or to more than one kind of

antagonistic relation between the animal and others.

The modification is not necessarily advantageous to the

animal with respect to all or to any features of the

No. 571] TERRESTRIAL ASSOCIATIONS 423

environment, though a large number of characters do

result in advantage. Characters advantageous in one

relation may be disadvantageous or indifferent in an-

other relation. The origin of the characters is not at

present a subject which can be treated in a study of inter-

relations of organisms (cf. Shelford, 1912b: 342). Be-

havior characters appear to be of greatest importance to

the animal in determining its relations with other organ-

isms of the association, though usually these are accom-

panied by physiological or structural characters. The

animal is not adapted to a particular status in the asso-

ciation; its ecological constitution determines what place

it shall be able to find among the other animals of its sur-

roundings. The relations among the various animals,

when a state of equilibrium has been reached, are the

result of mutual accommodation | on the part of all the

animals involved.

D. [INTERNAL Activities oF THE ASSOCIATION

It has been indicated that the complex of activities

within the association is the synthesis of all the activities

of the individual organisms. Each plant and each ani-

mal is subjected to physical, plant and animal influences.

From the extreme complexity of the entire system of

relations within the association, it is hardly possible to

consider more than one or several of these at one time.*

It is possible, however, to see that each species finds a

status within the association, according to its particular

combination of internal and external relations. It con-

tinues in fairly constant numbers from year to year. A

change in these numbers, if at all great, may cause a dis-

turbance in the association, which is quickly regulated

by the activities of conflicting organisms (Forbes, 1880).

The entire association of plants and animals, by very

' 8 Very helpful diagrams are given by Shelford (C: 167, 168) which il-

lustrate the food relations of land (prairie) animals. There are also dia-

grams showing food relations of aquatic animals (C: 70, 71). Food rela-

tions of animals of plains and mountain streams are discussed by Ellis

(1914: 122-127; diagram on p. 125). References to studies dealing with

interrelations Sy organisms may be found in the recent handbook of Adams

(1913: 123 et seq.).

424 THE AMERICAN NATURALIST [Vou, XLVII

reason of the conflicting interests, the varying conditions

necessary for existence, and the varying methods of re-

sponse to these conditions, forms a self-contained and

self-regulating system of activities.

III. RELATIVE INFLUENCE OF DIFFERENT ORGANISMS WITHIN

ASSOCIA TION—DOMINANCE

The plant ecologist determines which plants in an asso-

ciation are of greatest importance (dominant) by ob-

serving which species tend to increase at the expense of

others, which are most abundant, most frequent, largest,

ete. Competition among plants in a grassland associa-

tion is mainly for space, and the dominant species are

usually determined with considerable accuracy after some

study. With the animals the consideration of dominance

involves greater complexity. The important relations

between conflicting animal species are those in which

they obtain food, are removed from competition, or

escape enemies. These relations are in each case most -

directly concerned with food. The plant-eaters of the

association thus form a dominant group within the asso-

ciation, since predaceous and parasitic animals, and

scavengers in large part, depend upon them for existence.

Individual species within the various food-groups, how-

ever, present such striking differences in importance,

that we can not speak of all plant-eaters as dominant

forms, or that all animals of other food-habits are un-

important. It is merely probable that the phytophagous

group will contain a larger proportion of dominant spe-

cies. This appears to be the condition in the bunch-grass

association.

A. Factors or DOMINANCE AMONG ANIMALS

The success of an animal species within an association

is due to the resultant effect of a large number of factors.

Among these may be mentioned number of individuals,

size, activity, voracity, concentration of food, rapidity of

growth, rapidity of reproduction, and wideness of dis-

tribution in space and in time. Dominance signifies more

No. 571] TERRESTRIAL ASSOCIATIONS 425

than mere ability of a species to thrive in its surround-

ings: the species of greatest influence are those on which

the greatest number of other animals depend; thus domi-

nant species are successful, but successful species are not

always dominant. Species which are relatively free from

competition or which have comparatively few enemies

may be successful, but are not dominant, and are usually

not numerous. Species which are successful and at the

same time extremely abundant, usually form the food of

a large number of other animals, as it appears to be the

rule that no considerable source of food within the asso-

ciation is left unused. Dominance in a species, then,

would seem to include the dependence of other animals

upon it, plus the ability to thrive in spite of the drain

upon its numbers.

B. Crrrerta or DOMINANCE AMONG ANIMALS

The factors mentioned as contributing to the success

of a species, and the numbers of animals dependent upon

the species, are all indications of the degree of its domi-

nance. It appears that another criterion is available,

which perhaps expresses the summation of many factors

which contribute toward dominance. This is the degree

of specialization exhibited by the species in its adjust-

ment to a particular place in the association. Dominant

animals appear to be those of moderately specialized

habits rather than those of highly specialized, or rela-

tively unspecialized, habits.

C. ĶPECIALIZED AND UNSPECIALIZED ANIMALS

Each species may be referred to a position in the scale

of specialization in habit. The degree of specialization

of the species is well seen in the food-habits, though all

the habits are to be considered. The most abundant food

in the sand prairie is plant material, bunch-grasses. The

majority of the plant-feeders are adapted to eat herbage

of nearly any kind: they are not restricted to particular

species or particular parts of plants. They are non-

selective feeders. Grasshoppers, cutworms and certain

~ 426 THE AMERICAN NATURALIST [Vou. XLVII

leaf-beetles are thus moderately specialized plant-eaters.

There are also non-selective predaceous animals, as tiger-

beetles and lycosid spiders, which eat any kind of small

animal. These are also moderately specialized. The

moderately specialized animals carry on the gross metab-

olism of the association; they constitute the dominant

group, and include the dominant species.

Selective feeders belong with the highly specialized

animals. In the bunch-grass association Languria bi-

color, an erotylid beetle, bores in the stems of the com-

posite Cacalia (Indian plantain), while Lygeus bicrucis

(hemipterous) feeds on the same plant; Perillus circwm-

cinctus eats Blepharida rhois. Others of the associa-

tion eat selectively. The majority of parasites are

greatly restricted in their selection of hosts. Such ani-

mals are particularly dependent upon special kinds of

food, which in many cases are not available to general

feeders. Highly specialized forms are thus enabled to

avail themselves of opportunities denied to animals of

generalized type; but while they avoid competition by

the adoption of special kinds of food, or by special habit

of some other kind, they lack the versatility of the less

specialized animals, being unable to adjust themselves

to changed conditions. They may, therefore, become

abundant at times; but as they depend wholly upon one

variable condition (perhaps the presence of a particular

plant species, which may be quite infrequent) they never

can become dominant species. Absolute numbers of the

insects which live upon Cacalia, for example, are insig-

nificant in comparison with such animals as the grass-

hoppers.

On the other hand, animals of relatively non-specialized

habits would also be ineffective in the association, for

whatever field of activity they were to enter, they usu-

ally would find already occupied by some animal better

constituted for that activity. Such non-specialized forms

would assume particular importance only when some

animal on which they might feed should become unusually

No. 571] TERRESTRIAL ASSOCIATIONS 427

abundant. Few animals are really non-specialized in

habits; many moderately specialized species, however,

may on occasion turn from their ordinary activities, per-

haps to appropriate a particularly abundant kind of food.

Many ants are thus habituated to certain ordinary kinds

of food, but are able to eat organic food of almost any

sort, and do vary their food with circumstance. When,

as frequently happens, some animal species becomes very

abundant,‘ the attacks of a great many species of flexible

habits becomes concentrated upon it, and the numbers of

the food-species are soon reduced to normal. Animals

with non-specialized habits, by taking whatever food is

easiest of access, act as regulators of disturbances within

the association. A clear exposition of the manner in

which species of generalized habits restore unbalanced

conditions to equilibrium is given in a paper by Forbes

(1883), in which the regulative action of birds upon

insect oscillations is discussed.

The animal’s status within the association is deter-

mined not only by its food-habits, but by the sum-total of

its physiological and behavior characters (its mores).

The degree of dominance is indicated not merely by the

degree of specialization of food-habits, but in all habits,

by the degree of flexibility of behavior. An extreme

specialization in nearly any behavior character, as habit

of abode in the pit-digging ant-lion larva, prevents the

species from becoming dominant. The degree of spe-

cialization of behavior is thus a convenient criterion of

the relative influence of animals in the association. The

dominant animals are moderately specialized, and carry

on the ordinary work of the association. The highly spe-

cialized animals make use of space otherwise unoccupied

and food material not demanded by other species. Cer-

tain of the first group, with habits more highly regulatory

than is usual, with perhaps some few unspecialized forms

4 With some animals sudden abundance is a matter of seasonal periodic-

ity, as in the case of May-flies (Hexagenia) along the Illinois River (E:17).

The adults on emerging become a sudden source of food for animals of ad-

joining terrestrial associations, as the bunch-grass.

428 THE AMERICAN NATURALIST (VoL. XLVIII

in addition, tend, by following the path of least resist-

ance, to act in opposition to forces tending to destroy the

biotic equilibrium.

IV. DISTRIBUTION WITHIN THE ASSOCIATION

The association may be subdivided into minor groups

of organisms, both in space and in time. Each group,

being thus removed from the immediate influence of the

others, is to some extent self-contained, having its own

environmental conditions, its own assemblage of organ-

isms, and its own system of interrelations.

A. DISTRIBUTION IN Space®

Different parts of the space occupied by an association

present different environmental conditions. In the ver-

tical distribution, four strata, the air (cf. E: 73), the

plant layer, the surface layer and the underground layer,

are usually present. In forest associations, the plant

layer is complex, plants of various heights giving rise to

minor strata (cf. A). In grassland associations the plant

layer is relatively uniform. Animals are most numerous,

during the feeding activity, in the plant layer. Others

find food at the surface or underground. Many of the

animals in the air or on the ground move about rapidly

from plant to plant. Predaceous animals (while active)

are frequently permanent members of air and ground

layers, depending for food upon the transient animals

and upon members of their own group. The ground

stratum is composed of the surface and subsurface layers

(E: 72), which are not, however, continuous horizontally,

but alternate to greater or less extent.

Local variability in horizontal distribution is due

partly to local discontinuity of the various strata. This

interruptedness is particularly conspicuous in open asso-

ciations, where the plants do not form a dense growth,

but are separated by open spaces. The subsurface area

is provided by cover of various kinds, which lies more or

less scattered about on the surface.

5 Cf. Shelford, A, B, 1912b, C; also D: 167; also p. — of this paper.

No. 571] TERRESTRIAL ASSOCIATIONS 429

The motility of the animal allows change in stratum,

and to some extent and in some animals, in habitat, with

change in activity. The food-stratum and the food-

habitat are apparently of greatest importance in the rela-

tion of the animal to other organisms.

B. DISTRIBUTION IN TIME

Physiological activities of the plants are subject to

diurnal variation, and are also greatly affected by varia-

tions in weather conditions. The greater part of the

animals of an association are active during the day.

Others are nocturnal. During the inactive period of the

day the animal rests in some more or less sheltered place,

perhaps in a burrow or nest. The inactive state is also

induced by unfavorable weather conditions.

Seasonal changes in the association are very great in

temperate climates, particularly in treeless regions,

where the winters are severe. Seasonal changes in

the vegetation are marked, certain groups of the

plants appearing in successive periods during a sum-

mer season, giving four or five successive aspects to the

plant cover. A corresponding seasonal distribution is

observed among the animals of the association (cf. D:

175).

Annual changes in the associations are indicated by

the very marked differences in the numbers of indi-

viduals, in certain species of plants and animals, in suc-

cessive years. This may be due (1) to fluctuation in the

numerical adjustment between different organisms, and

(2) to the effect of annually varying phenological condi-

tions upon the various organisms.

Oscillatory irregularities in the association take place

at indefinite intervals. The causes and nature of oscilla-

tions have been thoroughly treated in several of Forbes’s

writings (1880, 1883, 1887). -

V. INTERDEPENDENCE OF TERRESTRIAL PLANT AND ANIMAL

COMMUNITIES

The thesis of the following section is that, in terrestrial

climatic or extensive environments, the relations between

430 THE AMERICAN NATURALIST [Vou. XLVII

the assemblage of plants and the assemblage of animals

are intimate and regular of occurrence; so much so that

(1) the two are coextensive, (2) the two constitute to-

gether a community which may be called a biotic asso-

ciation, (3) neither plant nor animal assemblage usually

occurs independently of the other, (4) the geographic

distribution of many of the plant and animal species

which make up the assemblages are in general corre-

spondence, (5) the species composition of the association,

over its range, varies no more widely, relatively speak-

ing, than would an assemblage of plants alone. Perhaps

the single view-point of the botanist, on one hand, and

the zoologist, on the other, has tended to a neglect of the

dual character of the one problem. Probably most botan-

ists and zoologists agree that relations of animals and

plants within a habitat are most intimate, and there is

a tacit assumption that all the organisms in one place

constitute the true system of interrelations, but botanists

have spoken of plant communities, and zoologists of

animal communities. There are numerous disharmonies

and variations in agreement of plant and animal assem-

blages, but these must not be allowed to obscure general

facts of correspondence.

It is recognized that plants and animals of an area of

essentially homogeneous physical conditions are inter-

dependent, the animals as a group being wholly depend-

ent upon the plants for food, and many of the plants

being directly dependent upon animals, as in the matter

of pollination. All are directly or indirectly affected by

animals in some way. It is also recognized that the

plants are a good index to conditions for animal life, the

plant assemblage affecting animals locally in modifica-

tion of the physical environment, and more directly in

providing food, shelter, ete. (4: 601). It is further ac-

cepted that plants and animals respond to general en-

vironmental conditions in similar manner (Craig, 1908).

Thus considered, the character of the plant population of

an area is an index to general character, or ecological

No. 571] _ TERRESTRIAL ASSOCIATIONS 431

type, of the animal assemblage. These relations, how-

ever, are quite general, lacking detail. Detailed consid-

erations may be geographic, including geographic range

of species and of communities, and the distribution of

species and of individuals into communities; and they

may also be local, dealing with interrelations of plants

and animals within the area of the community.

A. QEOGRAPHIC RELATIONS OF TERRESTRIAL PLANTS

AND ÅNIMALS

1. Geographic Range: The Province.—If one were to

plot the geographic range of the plant species found to-

gether in a given climatic habitat, a general correspond-

ence in distribution would be made apparent, a large

number of the species ranging more or less continuously

over a common, rather definite area (cf. Transeau, 1905).

The similar ecological constitution of these plants and

their consequent selective distribution into similar envi-

ronmental complexes gives a uniformity to the vegeta-

tion over the geographic region in which these environ-

mental conditions are found, and the resulting vegeta-

tion unit is known as a vegetation province (Gleason,

1910: 42). The area of the province is generally uniform

in physical conditions. This uniformity is only relative, —

being subject to gradual geographic variation in climate,

perhaps giving rise to subregions in distant parts of the

province, and to abrupt local variations in soil, water-

content, exposure, etc., giving rise to local or edaphic

plant assemblages very different from those of the cli-

matic or geographic type. Thus the prairie province

occupies the winter-dry interior region of North America.

Environmental variations from east to west, climatic and

physiographic, divide the province into the three sub-

regions of Pound and Clements (1898). Certain plant

species range over one or all of these subregions, still

others establishing themselves over the whole area of the

province and also scatteringly eastward, in dry treeless

parts of the deciduous forest province, to the Atlantic

coast. These last are also typical prairie plants, though

432 THE AMERICAN NATURALIST [VoL. XLVIII

extra-limital in parts of an adjoining province locally

approximating the prairie environment.

The habitat-selection of different animal species re-

sults, in precisely the same manner, in similarity of geo-

graphic range among ecologically similar animals. These

correspondences of distribution point to the existence of

definite areas characterized by general similarity of the

animal assemblages. As the physical factors of the en-

vironment are the same ultimately for animals as for

plants, and as the vegetational environment for animals

has the same range as the physical environment, we

might expect animal communities to have the same geo-

graphic distribution as plant communities, and we might

expect the area of the plant province to be characterized

by distinctive kinds of animals as well as by distinctive

kinds of plants. The province is thus not simply a vege-

tation province, but a biotic province. This is not a

new notion. Ruthven (1908: 388-390) has stated a cur-

` rent viewpoint as follows:

Those who are acquainted with the literature of the field zoology of

North America are familiar with the fact that, since the time of the

Pacific Railroad surveys, naturalists have noted that there are in North

America well-defined biological regions. These have been pointed out

at various times by Allen, Cope, Merriam, and others, and the fauna of

each has been more or less investigated. . . . For example, we have forms

of birds, reptiles and mammals characteristic of the southeastern de-

ciduous forest region, and still others characteristic of the northeastern

coniferous forest region, ete.

Shelford (4: 604) bases his classification of animal

regions upon that of plant regions, as worked out by

Schimper (1903) and Transeau (1903, 1905).

How close the correspondence of distribution of par-

ticular animals with that of vegetation provinces may be,

is well shown in the case of North American rabbits

(Nelson, 1909). The distribution maps shown for certain

species and groups of these animals might almost serve

as maps of the provinces. Many other animals, verte-

brate and invertebrate, correspond in area with the plant

provinces. Among the insects listed by Hart (1907: 205)

No. 571] ‘TERRESTRIAL ASSOCIATIONS 433

as western species, those for which a number of locality

records are available are plainly to be assigned to the

prairie province, the range of most of them extending

west to the Rocky Mountains, north about as far as

Montana, east to Illinois or Indiana, and south to Texas.

Other animal species bear apparently no relation to

province boundaries. Such animals have been discussed

by Shelford (A: 606, footnote), who shows them to be of

three types: (1) Species of scattered but very wide

range, covering perhaps several plant realms (animals

of local associations of extreme habitats); (2) Species

occupying only a part of the plant realm in which they

belong (animals of such ecological constitution that their

range is restricted by some conditions unfavorable in

certain parts of the province); (3) Species occupying

intermediate ground between two realms—these are few

(Ruthven). These exceptional species are found also in

plants, so that local associations are occupied by both

plants and animals of the secattered-but-wide type of

range, while certain subregions, as the Great Plains area

of the prairie province, contain associations with both

plant and animal species restricted to these less extensive

areas.

Associations of two adjoining provinces may inter-

grade, if ecologically similar, or may alternate if dis-

similar. Similar associations of two provinces may con-

tain the same or closely related species, as with certain

grasshoppers which range in both northeastern and west-

ern coniferous provinces (D: 173). But these same asso-

ciations contain also plant species in common, so that

irregularities of range are no greater in animals than in

plants.

2. Distribution Within the Province: Distribution of

Plants and Animals into Communities.—It is seen that

plant and animal species may correspond closely in geo-

graphic range. There may be also more local corre-

spondence in distribution. The plant community has

been found by the writer to be the convenient index of the

434 ` THE AMERICAN NATURALIST [Vow. XLVII

area of the habitat for animals. It has been observed,

in an area in Michigan, that grasshopper species corre-

spond closely in local distribution with plant communities

(D). There is evidence that local distribution of ani-

mals is seldom promiscuous as a result of motility (D:

159). It appears also that the local variability of envi-

ronmental conditions within the area of the climatic plant

community is sufficiently great, usually, to supply all

necessary conditions for a large number of animals, so

that the limits of the plant community need not be

passed, ordinarily.

The animal community of the area may be thus, in large measure, self-

contained, and coextensive with the plant community (D : 161).

One of the problems of plant ecology has been the

differentiation of plant communities or associations.

Mere comparison of lists of species is not sufficient; rela-

tive abundance of various species must be considered as

well. Animal assemblages in contiguous areas must be

separated in the same way. Given two adjoining habitats

differing in plant population, it has been found that, in

addition to differences of animal species, there are also

differences of relative abundance in those animal apoa

common to the two areas (D: 154, 167).

The local area of a plant community is determined by

(1) local distribution of the physical environmental com-

plex, and (2) influence (competition, etc.) of adjoining

plant communities. Local area of the animal community

depends upon (1) local distribution of physical environ-

ment, and (2) local distribution of vegetational environ-

ment, the latter being uniform over the area of the plant

community. Contiguous areas differing in physical and

vegetational conditions will be expected to differ also in

animal population, in a degree comparable to that of the

differences in environmental conditions.

Physical habitats, and plant communities, sometimes

alternate, sometimes intergrade; it is not unreasonable to

expect accompanying alternation or intergradation of

6 Differences in species, both plant and animal, are accompanied by dif-

ferences in ecological constitution.

No. 571] TERRESTRIAL ASSOCIATIONS 435

animal populations. Certain of the animal assemblages

of sand habitats, as studied in central Illinois, intergrade;

others, as oak forest and bunch-grass, differ radically.

The above considerations, if correct, appear to signify

that, in ordinary climatic development of plant and ani-

mal life in temperate land environments, the area of the

animal assemblage is that of the plant assemblage, both

resting basically upon the physical environment. The

plant and animal assemblages are therefore coextensive

parts of a biotic association, composed of both plants

and animals, and this association as a whole constitutes

the real terrestrial community of living organisms.

B. Locat RELATIONS or PLANT AND ANIMAL ASSEMBLAGES

(RELATIONS WITHIN THE ÅSSOCIATION)

The more intimate relations between plants and ani-

mals are seen in the detailed study of a single associa-

tion. The bunch-grass association of sand prairie is

selected for illustration (E: 68).

1. Similarity of Ecological Type of Plants and Ani-

mals —Shelford has shown (A: 593-594) that animals

and plants may evince ecological similarity by similar

response to the same general environmental conditions,

behavior responses in animals’ corresponding to struc-

tural responses in plants,’ so that mores of the animal

may be in accord with growth-form in the plant. Shel-

ford states (B: 87) that ‘‘plants and animal communities

are in full agreement when the growth-form of each

stratum of the plant-community is correlated with the

conditions selected by the animals of that stratum.’’

In the bunch-grass there is general agreement, ac-

cording to this criterion. The herbaceous stratum is oc-

cupied mainly by tuft and mat plants—bunch-grasses,

cactus and a few half-shrubs. Associated with the tuft or

mat growth-form is the sedentary mores of the plant-

inhabiting animals (leaf-beetles, stem-borers, ambush-

bugs, ete.). A considerable proportion of ground surface

7 Or motile organisms, ef. C: 305.

8 Or sessile organisms.

436 THE AMERICAN NATURALIST [Vou. XLVIII

is bare sand; in the interspaces between the dominant

plants are slender annuals (interstitial plants), and here

are also found animals of the roving mores of the ground

stratum (interstitial animals). Many of these are swift-

running and predaceous (six-lined lizard, tiger ae

lycosid spiders).

Correspondence in ecological type of plants and ani-

mals in the bunch-grass is not complete in several re-

spects. Shelford mentions types of disagreement (B:

88; C: 306-308), and there is a further important kind of

disharmony, in mixed associations, due to presence of

diverse types of plants and animals (D: 163). Mixed

associations are quite frequent in forest border regions,

and in the transition area between two provinces. The

plant and animal assemblages of a given habitat, partic-

ularly if climatic and extensive, are usually in general

ecological agreement, and the exceptions are likely to be

infrequent or temporary (Shelford, B: 88).

2. Relative Dependence of Plant and Animal Assem-

blages——There is evidence that the agreement of plant

and animal assemblages of terrestrial associations is

often a matter of accommodation on the part of the

animal assemblage. In the early stages of development

of vegetation, local physical conditions dominate; in

later stages the vegetation assumes the type determined

by climatic conditions, and exerts nearly complete con-

trol over local physical factors. In established associa-

tions, therefore, the locally dominating environmental

feature is the vegetation. Shelford states that in the

several associations of a successional series, the domi-

nating animal mores are correlated with the dominating

conditions (B: 94) and that, as the forest increases in

density, the animals make use of the vegetation in in-

creasing degree, particularly for breeding-places, and as

places of abode (B: 90). Many grasshoppers of open

grassland depend upon a particular kind of soil for egg-

laying, while those of closed forest lay eggs in fallen

logs—a condition of the plant environment (D: 163).

No. 571] TERRESTRIAL ASSOCIATIONS 437

The sand-prairie vegetation is in an intermediate stage,

certain animals depending chiefly on the presence of

loose bare sand, others on the bunch-grass vegetation.

With development of bunch-grass into closed grassland,

the interstitial animals are eliminated. The animals of

established associations, while in accord with climatic

physical conditions, are perhaps more intimately affected

by vegetation conditions. Since established associations

are very much more extensive than primitive associa-

tions, the importance of vegetation as a dominating part

of the environment for animals becomes apparent, and

we may conclude that the character of the plant assem-

blage determines, to a large extent, the ecological type

of the animal assemblage.

3. Correspondence in Distribution within the Associa-

tion.—The uniformity of physical and vegetational con-

ditions is only relative. There are spots in the bunch-

grass association in which local invasion of blue grass

has occurred, darkening and binding the soil. In such

partly humified situations, small colonies of the corn-

field ant, not occurring elsewhere in the bunch-grass (E:

57), have been found. There are also areas some few

feet in diameter in which the bunches of grass are few,

small and seattered. In these relatively bare patches the

abundance of interstitial animals is greatly increased.

More direct relations are seen in the case of animals

associated with particular species of plants. Within the

association, any animal species, like any plant species,

may be distributed generally throughout the area, or it

may be restricted to a part of the area characterized by

a slight environmental difference, or it may occur in

seattered parts of the association, characterized by

seattered local differences (D: 168). There is evidence

that, in so far as the vegetational environment is con-

cerned, distribution of animals within the association is

usually a direct function of similar distribution of plants.

4. Uniformity of Species Composition of Plant and

Animal Assemblages.—It has been seen that plant assem-

438 THE AMERICAN NATURALIST [Vou, XLVII

blages of definite ecological type, as regards growth-

form, etc., are regularly accompanied by animal assem-

blages of similar ecological type, as regards mores. In-

terest attaches also to the problem whether associated

plant and animal assemblages show definite species

relations. —

One familiar with a certain association, who visits a

representation of that same growth in a different part

of the same climatic region, will be struck with the fact

that a large proportion of both plant and animal species

is well known, while a certain proportion, perhaps con-

siderably smaller, is new to him. The writer has been

impressed with the similarity of the plant and animal

populations of the sandhills of central Nebraska and of

eastern Colorado, to those of the sand prairie of central

and western Illinois, despite the fact that certain species

are not common to the two areas. Tiger-beetles, blow-

snake, grasshoppers, box-turtle, six lined lizard, western

meadow-lark, white-footed mouse, among the animals;

prickly-pear, lead-plant, bunch-grasses, sand-bur, sand

evening primrose, among the plants; are represented in

the two areas either by the same or by closely related

varieties and species. There are no yuccas or sand-sages

in the Illinois sand prairie, no lizard Holbrookia nor

lubber-grasshopper Brachystola; and there are certain

eastern species not found in the western sandhills. But

on the whole the species (particularly the important

species) common to the two areas are more numerous.

This is the more remarkable in view of the fact that dis-

tribution of sand prairie is discontinuous, the largest,

nearly uninterrupted gap being several hundred miles in

extent. Many of the animals, as well as plant species, of

dry mixed prairie-grass in loamy soil, are the same along

the mountain-front in Colorado (Vestal, 1914b) as in

north-central Illinois. The likenesses become much more

impressive as distance is decreased.

Absolute identity of species composition, where large

numbers of species are involved, is an ideal condition,

No. 571] TERRESTRIAL ASSOCIATIONS 439

never actually attained. No one can say just what pro-

portion of species-in-common is necessary for two

growths to be’said to represent the same association. In

addition to likenesses and differences of environment, of

aspect, and of history, which must be weighed, the differ-

ent plant and animal species vary so much in importance

in the association, in physiological variation and in range

of environmental tolerance, that associations can hardly

be separated or placed together on a statistical basis. A

comparison of species is fair if the following kinds of

plants and animals are left out of consideration; (1)

those of limited range within the climatic region or prov-

ince, including species belonging more properly to other

provinces; (2) those of very indefinite habitat-relations,

which are found in nearly any kind of habitat; (3) those

of special restricted habitats, which may be scattered

about in many kinds of associations, as moist dead wood,

in which particular fungi, beetles, perhaps snails, myrio-

pods and pill-bugs, are usually found; or as excrement of

grazing animals, in which certain molds, certain dipter-

ous and scarabeid larve, etc., regularly occur, irrespect-

ive of surrounding conditions; (4) invaders from near-

by associations; (5) ruderal and introduced species; and

possibly one or two other groups. The second and third

groups may be called the irregular element; the fourth

and fifth may be known as the derived element. While

these groups make a formidable list, their representa-

tives constitute usually a very small proportion of the

organisms of the association. The other organisms, and

some of these, follow habitat-differences, as represented

in different associations, very closely.

Since hardly any two species are identical in habitat-

relations, geographic and even local variation must be

looked for, but since many species resemble each other

more or less closely in general ecological relations, there

come to be recognized certain ecological groups of spe-

cies, each characterized by a general type of growth-form

in plants, or by a general kind of mores in animals, and

440 THE AMERICAN NATURALIST [Vou. XLVIII

these groups may be considered to be small or large,

according as we emphasize minor differences or general

likenesses.

Now within any limited region (let us postulate first

an area removed from the influence of an adjoining prov-

ince) there are only a limited number of ecological

groups, of growth-forms of plants, and of mores among

animals, each group represented by a limited collection

of species. Each habitat within this restricted area will

be characterized by definite physical conditions, and with

these will be correlated certain growth-forms of plants

and certain mores of animals, each represented by as

many of the species as can migrate into and survive

within the area, as determined first by capabilities of

migration and by habitat-selection, and second by inter-

relation of species and of individuals. It follows that

physical complexes which are alike will become populated

with similar complexes of ecological groups, represented

by similar collections of plant and animal species, and

that unlike physical areas will be occupied by different

combinations of ecological groups, and will be composed

of different species. Two areas within this region which

have similar physical conditions and similar plant

growths will be expected to have a large number of ani-

mal species in common, although direct relations between

species of animals and species of plants obtain only

rarely (between comparatively few associated plant-and-

animal pairs). It is to be noted that species composition

of the animal assemblage varies proportionately no more

widely than does that of the plant assemblage.

No terrestrial continental region is sufficiently isolated

to be free from influence of surrounding areas, and since

the influences are different from different directions, and

since there is continual change of physical conditions, and

of range and abundance of plant and animal species,

there must be more or less local and geographic varia-

tion of species composition within similar but separated

habitats. Geographic variation is wider with distance.

No. 571] TERRESTRIAL ASSOCIATIONS 441

because the geographic and physiographic complexes

vary geographically, as well as the entire collection of

plant and animal species which may invade the habitat.

Within the area of the climatic province, however, or at

least within the area of a subregion of the province,

climatic, physiographic and biotic complexes are likely

to be relatively constant, that is, likenesses of two areas

are likely to be greater and more striking than differ-

ences. Within the province or subregion, therefore, it

is to be expected that species composition of association

of closely similar habitats will be relatively constant.

Particular plant and animal assemblages will be found

together, both associated with a particular habitat. Field

observation bears out these expectations.

Conditions within the transition zone between two

climatie regions or provinces are much more complex

than in an area in the middle of a sub-region or province;

climatic and physiographic conditions vary to wider

extremes and are less stable; the total number of species

near enough at hand to invade a given habitat is much

greater. Mixed associations, often transitional as re-

gards physical conditions, are composed of representa-

tives of both provinces. Animals of a particular associa-

tion of one province, may be found with plants of a

similar or equivalent association of the other province.

When three geographic elements are represented, as at

the southern end of Lake Michigan (cf. C, and Vestal,

1914a), the complication of conditions is extreme. Even

here, on the dry sand of old lake beaches, fairly typical

representations of sand prairie can be seen; and though

such habitats are shared with deciduous forest associa-

tions, and with associations of the northeastern coniferous

forest province, and with mixed associations, the bunch-

grass growth can still be recognized in dry shifting sterile

sand, with bunch grass plant species, and bunch-grass

animal species. The tendency towards uniformity of

association of plant and animal assemblages is even here

to be made out.

442 THE AMERICAN NATURALIST [Vou. XLVIII

If the foregoing considerations relating to relations

between plant and animal communities are correct, the

thesis mentioned at the beginning of part V would seem

to be justified, though the evidence is far from complete.

Plant and animal assemblages are mutually interdepend-

ent; the plant assemblage dominates in established

associations. Plant and animal assemblages correspond

in geographic distribution, in distribution into commu-

nities, and in more detailed distribution within the

habitat. They are made up of ecologically similar groups

correlated with the same physical conditions or with

‘each other. Though there are few direct relations be-

tween particular species of plants and animals, it so

happens that within any restricted region, particular

collections of animal species come into regular associa-

tion with particular collections of plant species, the spe-

cies composition within the habitat exhibiting a greater

or less degree of uniformity, except for minor irregular

and derived elements. The more restricted, or uniform

in biological conditions, this region is, the greater the

uniformity of the collection of species. Climatic and ex-

tensive associations, and established associations, show a

greater degree of uniformity than local or primitive

associations,

VI. SUMMARY AND CONCLUSIONS

The discussion is based principally upon the writer’s

study of prairie associations, the bunch-grass associa-

tion of sand prairie in Illinois being chiefly used for illus-

tration. Internal activities of the association are a com-

plex of activities of all the organisms. Environmental

influences are of three classes, physical, plant and animal.

The characters of plants and animals are interpreted in

their relation to these influences. Characters of plants

may be classed as structural, physiological, biographical

and numerical. Animals have, in addition, behavior or

psychological characters. These groups of characters

are intimately related, one to another. The relations of

the animal to its animal-environment are of two kinds,

No. 571] TERRESTRIAL ASSOCIATIONS 443

social and antagonistic, the latter relations being with

food-species, competitors and enemies. Correlations of

the various kinds of characters with relations involving

food, competition and enemies, are given. According to

its ecological constitution, each organism finds a status

in the association, the whole being a self-contained and

self-regulating system of activities.

Dependencies within the association are concerned

mainly with sources and interchange of material and

energy. Dominant plants (the most influential species)

are those most intimately correlated with physical en-

vironment, as indicated by aggressiveness, abundance,

frequence, size, ete. Dominant animals are most numer-

ous among phytophagous forms. Dominance in an ani-

mal species includes dependence of other animals upon it

(for food) plus the ability to thrive in spite of the drain

upon its numbers. The degree of specialization of be-

havior is a convenient index of the relative influence of

animals in the association. The dominant animals are

moderately specialized, and carry on the ordinary work

of the association. The highly specialized animals make

use of space otherwise unoccupied, and food material not

available to other species, or not taken by other forms.

Least highly specialized animals act as a check upon

undue departure from biotic equilibrium.

The association may be divided into minor groups of

organisms, both in space and in time. Space-division is

vertical, resulting in strata, and horizontal, resulting in

sub-habitats of greater or less magnitude. The strata

and sub-habitats present a larger or smaller degree of

discontinuity and of internal variability. Time-distribu-

tion is diurnal, seasonal and annual. There are also

time-variations produced by variability of weather condi-

tions and by oscillatory disturbances.

The relations between plant and animal assemblages

have long been known, in a general way, to be intimate.

Plants and animals agree in similar response to common

environmental influence, and in types of geographic dis-

444 THE AMERICAN NATURALIST [Vou. XLVIII

tribution. Upon investigation, it begins to appear that

plant and animal assemblages are coextensive parts of a

biotic association, composed of both plants and animals,

and this association as a whole constitutes the real ter-

restrial community of living organisms. Plant and ani-

mal assemblages are mutually interdependent; the plant

assemblage dominates in established associations. Plant

and animal assemblages correspond in geographic dis-

tribution, in distribution into communities, and in more

detailed distribution within the habitat. They are made

up of ecologically similar groups correlated with the

same physical conditions or with each other. Though

there are few direct relations between particular species

of plants and animals, it so happens that within any

restricted region, particular collections of animal species

come into regular association with particular collections

of plant species, the species composition within the habi-

tat exhibiting a greater or less degree of uniformity,

except for minor irrégular or derived elements. The

more restricted in area, or uniform in biological condi-

tions, this region is, the greater uniformity of the collec-

tion of species. Climatic and extensive associations show

a higher degree of uniformity than local or primitive

associations.

VII. REFERENCES

Special References

(A) mugen Vv. 0%

11b. ae ae Animal Geography. Jour. Morya: 22: 551-

(B)

1912a, Ecological Succession. IV. Vegetation and the Control

of Land Animal Communities. Biol. Bull., 23: 59-99

(C)

1913. Animal Communities in Temperate America. Geogr. Soc.

of Chicago, Bull. No. 5, p. 362. Chicago.

(D) Vestal,. A. G.

13a. REF Distribution of Grasshoppers in Relation to Plant

ssociations. Biol. Bull., 25: 141-180.

CE a nas

1913b. An Associational Study of Illinois Sand Prairie. Bull.

Ill. State Lab. Nat. Hist., 10: 1-96.

Other Articles Cited

Adams, C.

1913. Guide to the Study of Animal Ecology, p. 183. New York.

No. 571] TERRESTRIAL ASSOCIATIONS 445

Craig, W.

1908. eee Dakota Life: Plant, Animal, and Human. Bull. Am.

gr. Koc., 40: 321-332, 401-415.

Ellis, M. M.

1914. Fishes of Colorado. Univ. Colo. Studies, 11: 1-136.

Forbes, S. A.

1880. On Some aas of Organisms. Bull. IU. State Lab. Nat.

Hist., Vol. 1, No. 3, pp. 3-17

1883. The Benais Action of Birds upon Insect waa ce Bull.

Ill. State Lab. Nat. His., Vol. 1, No. 6, pp. 3

1887. The Lake as a a Reprint from Bull Sei. Assoc. of

Peoria, Il., pp. 1-15.

1909. The General bien Ecology of the Indian Corn Plant.

Am. Nat., 43: 286-301.

Gates, F. C.

1912. The Vegetation of the Beach Area in Northeastern Illinois and

Southeastern Wisconsin. Bull. IU. State Lab. Nat. Hist., 9:

55-372

Gleason, H. A.

1910. The Vegetation of the Inland Sand PEMP of Illinois. Bull.

Ill. State Lab. Nat. Hist., 9: 23-1

pees C. A., and Gleason, H. A.

190 ‘On the cones of the Sand tomas of Illinois. Bull. Ill. State

Lab. Nat. Hist., 7: 137-2

Nelson, E. W.

1909. The Rabbits of North pei Bur. Biol. Surv., U. S. Dept.

Agr., N. Am. ea No. 29, pp. 314.

pe R., and Clements

1898. The Varant ait of the Prairie Province. Bot. Gaz., 25:

381-394.

Ruthven, A. G.

1908. The Faunal Affinities of yf ae Region of Central North

merica. AM. NAT., 42: 388-393.

Schimper, A. F.

1903. Plant pen upon a Physiological Basis.

Shelford, V. E.

1911a. Ecological Succession. I. Stream Fishes and the Method of

hysiographie repens eee Biol. Bull., 21: 9-35.

1912b. Ecological Succession. V. Aspects of Physiological Classifi-

eation. Biol. Bull., 23: 331-370.

Oxford.

Vestal, A. G.

1914a. The Status of Prairie oe in the pap Beach Areas

of Lake Michigan. r. Ecology. (In press.)

1914. Prairie Vegetation of a a front Area in | Colorado. Bot.

Gaz. (In press

Transeau, E. N.. i

1903. On the Geographie Distribution and Ecological Relation of Bog

Plant Societies. Bot. Gaz., 36: 40 seq.

1905. Forest Centers of Eastern heiatins: Am. Nart., 39: 875-889.

SHORTER ARTICLES AND DISCUSSION

ANOTHER HYPOTHESIS TO ACCOUNT FOR DR.

SWINGLE’S EXPERIMENTS WITH CITRUS

THE results of the cross-breeding experiments with forms of

Citrus by Walter Swingle have given rise to quite a number of

different hypotheses, to account for the facts observed.

The facts are simply these. All the different.forms of Citrus

used in the experiments, Citrus trifoliata, the lemon, orange and

other citrous fruits have, so far, proved to reproduce their own

type through seed.

Nevertheless, the plants raised from one single cross are ex-

ceedingly different among themselves. And yet, all these new

forms, for so far as tested, have proved truly to reproduce their

own kind only, if sown.

The theories offered to account for PEN facts are rather com-

plex. So far, we have not seen the simple hypothesis which we

want to add to the others.

The fact, that the F, from almost every cross between types of

Citrus is multiform, can only be accounted for on the assumption,

that the parent plants are impure (heterozygous) for a number

of genes. The difficult question is this: how can a tree, impure

for a number of genes, produce seed which always only repro-

duces the type? We know, that if a plant reproduces itself

by an asexual method, all its daughter plants are pure for

those genes in respect to which it was pure, impure for those

genes for which it was impure. Is it possible that in these

trees the seeds normally produced are not derived from a union

between two normal gametes? In Citrus, with its adventitious

embryos, this is very well possible. If the forms of Citrus used

by Dr. Swingle are self-sterile, the seeds normally produced by

these trees, are not produced by the union of two gametes, but

as buds, asexually.

This hypothesis, that the Citrus used are self-sterile, and that

the seeds normally produced, are produced asexually, fully ac-

counts for all the facts. All the daughter plants from un-

crossed seeds are genotypically identical with the mother plant,

as in all clones. On pollenization by another tree, normal se

are produced, each the result of the union of two real gametes.

446 ` :

No.d571] SHORTER ARTICLES AND DISCUSSION 447

These seeds contain different combinations of the genes, for

which the parent plants are impure, as normally. The F, gener-

ation for this reason becomes as diverse as such generations

always are, if the parents are impure for numerous genes.

But these daughter plants, although impure for a number of

genes, can, because of their self-sterility, in their turn only pro-

duce seed asexually and therefore their offspring will be like

themselves,

It should not be difficult to test our hypothesis. It seems

easier to find out, whether the seeds produced without crossing in

Citrus contain the embryo formed by fertilization of the em-

bryo sac, or embryos formed adventitiously by the adjacent tis-

sue, than to test any of the other theories, which assume a pecul-

iar behavior of the chromosomes.

Our hypothesis, that a variable F,, of only true-breeding

plants (from the union of two true-breeding forms), results

from habitual self-sterility and asexual production of seed, with

real fertilization in the case of a cross taking place, not only

accounts for the facts found by Swingle, but also for those found

by Rosen with Erophila verna. These facts were somewhat dif-

ferent. The F, plants were all identical, and somewhat inter-

mediate. They gave rise to a variable F, generation of which

all the plants bred true to their type. These facts can be ex-

plained on the assumption, that Erophila verna is self-sterile,

and that, in the absence of cross-fertilization, unfertilized egg-

cells develop parthenogenetically. Such F, plants, which are

impure for a number of genes, should therefore produce as many

different kinds of F, plants, as there are female gametes pro-

duced, and in the same proportions. In the case of such a plant

being impure for two genes, we should expect it to produce

plants of the four different types, not in the usual proportion of

9:3:3:1, but in equal proportions, 1:1:1:1. The F, plants

from such seed could only be pure for all the genes present.

It would be possible in Erophila verna to find out whether F,

plants, impure for two genes, produced daughter plants of the

four kinds, AB, Ab, aB, and ab, in the proportion of 9:3:3:1,

or in proportion 1:1:1:1, and thus to test our hypothesis.

To find out, whether it is possible, that a plant, impure for a _

number of genes, produces a variable F, generation of only

completely homozygous plants, we have begun a series of experi-

ments with squashes. Some hybrid plants have not produced a

448 THE AMERICAN NATURALIST [Vou. XLVIIT

single fruit from carefully sealed female buds, others have given

plenty of empty fruit, but some hybrids have produced several

fruits, full of viable seed. If this seed is formed by the par-

thenogenetic development of unfertilized normal egg-cells, as

we have reason to believe, we expect to raise a variable F, gene-

ration of exclusively homozygous plants. If these seeds have

developed by apogamy, or any other asexual process, we expect

to obtain a second generation consisting exclusively of plants

like the original hybrids. Thus we will have a non-cytological

test to decide between apogamy and true parthenogenesis.

A. C. HAGEDOORN,

A. L. HAGEDOORN

Bussum, HOLLAND,

March 18, 1914

VOL. XLVIII, NO. 572 K AUGUST, 1914

THE

AMERICAN

NATURALISE

A MONTHLY JOURNAL

Devoted to the Advancement of the Biological Sciences with

Special Reference to the Factors of Evolution

CONTENTS

Page

I. Multiple Allelomorphs in Mice. Professor T. H. MORGAN ~ ~ -~ - 449

II. Thirteen Years of Wheat Selection. T. B. HUTCHESON - - = 469

OI. Pattern Development in Mammals and eae GLOVER M. ALLEN - - 467

IV. The Meadow Jumping Mouse. Dr. H. L. BABCOCK = s - 485

V. Shorter Articles and Discussion: nak on ee Dr. RAYMOND

PEARL. Parallel Mutations in (nothera biennis L. Dr. J. 8tomps, Dr

BRADLEY M. Davis. The Theoretical ne as seers Multiple Allelo-

nates nd Close —— TRETE. T. H , Professor W. E.

— 491

VI. Pipers and p ee Monetti YMOND p Paani. i New Mode of

Segregation in Gregory’s Tetraploid ai HERMANN J. MULLER ~ 505

THE SCIENCE PRESS

LANCASTER, PA. GARRISON, N. F.

NEW YORE: SUB-STATION 84

The American Naturalist

intended for inte and books, etc., intended for review should be

MSS

sent to _ ae of THEA

ICAN NATURALIST, Garrison- -on-Hudson, New York.

rticles containing summaries of research work bear ring on the

problems of organic evolution are especialiy welcome, and will be given preference

in sng

canoes reprints of contributions are supplied -to authors free of charge.

Further 1 reprints will be supplied a

cos

criptions and sivertiesments should be sent to the Larong

ar. Foreign postage

sclisvigiion price is four dollars a

Canadian postage twenty-five cents additional.

The

s fifty cents and

The eho for single copies is

forty cents. The advertising rates are Four Dollars for a p

THE SCIENCE PRESS

NEW YORK: Sub-Station 84

Entered as second-class matter, April 2, 1908, at the Post Office at Lancaster, Pa., under the Act of

ongress of March 3, 1879.

Lancaster, Pa.

Garrison, N. Y.

F E

ARCTIC, ICELAND and GREENLAND

BIRDS’ SKINS,

Well cage Low Prices

articulars of

_ DINESEN. Bird Collecto

Husavik, ‘North Iceland, Via Leidle, England

WANTED TO PURCHASE

a set of BIRDS OF AMERICA by J. J. Audubon,

7 or 8 volumes, please report, stating cash price, stat-

ing condition, binding and dates of volumes.

F. C. HARRIS,

Box 2244 Boston, Massachusetts

For Sale Entire

An important collection of Indian Birds’

Eggs, containing approximately 3,300

specimens of about 620 species, and in-

cluding among other varieties the Green-

ish Willow Warbler (Acanthopneuste

viridanus), Red-browed Finch (Calla-

eanthis burtoni), Tibetan Sand-Grouse

(Syrrhaptes tibetanus), Spotted Sand-

Grouse (Pteroclurus senegallus), Ibis-

bill (ibidorhynchus struthersi), White-

bellied Minivet (Pericrocotus erythro-

pygius), Bearded Vulture (Gypaétus

barbatus), numerous Cuckoos with their

hosts, ete., ete.

Particulars and full list may be had

rom

W. F. H. ROSENBERG

57 Haverstock Hill

London, N. W.

Marine Biological Laboratory

Woods Hole, Mass.

INVESTIGATION

Entire Year

ue for research in Z

for rie ip a > table is

laboratory on

Comps Pet ures are offered in ao

INSTRUCTION

July 1 to August 11,

Beoloey a of grey Phy wad and Bog

Vi stati tion. Each o

the full time of te “student E Fee

$50.00. A lectu uree io

rg wig ne “Aspects of of mi

UPPLY Animals

DEPARTMENT

and ts, pre

living, and embryonic ol

Preserved material of all typet ip

Open the Entire Year animals and of Algae, Fungi, ed for

wo Mosses furnish aah

classwork, or for the, muse,

Living material furnished in serasi

ordered. ice lists no torial

logical and Bı cal pT

sent on application. State i all

is desired. For prios 1 obe iiai,

information regarding

GEO. M. GRAY, Curator, Woods Hole, Mass

The annual announcement will be sent on application to

| The Director, Marine Biological Laboratory, Woods

| Mass-

THE

AMERICAN NATURALIST

VoL. XLVIII August, 1914 No. 572

MULTIPLE ALLELOMORPHS IN MICE

PROFESSOR T. H. MORGAN

COLUMBIA UNIVERSITY

Some breeding experiments with mice that I have been

carrying on during the last two years have shown that

yellow, gray gray-belly, gray white-belly and black are

allelomorphs. To this series a fifth allelomorph may pos-

- Sibly be added which for the present may be called new

gray. This quadruple (or quintuple) system of allelo-

-morphs fulfils the conditions of a multiple series in that

only two of the allelomorphs can exist at the same time

in any individual. In other words, a mouse may be pure

for any of these genes (except for yellow, in which the

pure form is not viable), or a mouse may be heterozygous

in any two of the genes, but never in more than two.

The evidence that establishes this series of allelomorphs

may be briefly stated as follows:

In 1911, I pointed out that if yellow mice (producing

-yellow and chocolates) are bred to agoutis (grays), and

their yellow offspring mated, they should produce not

only yellow and agoutis (as they did) but some choco-

lates (or blacks) also; but no chocolates appeared. I

stated that the results obtained were explicable if yellow

and agouti are allelomorphs.!

1 The discussion in the same paper of the presence of chocolate yellow and

black bars in the ticked hair in relation to the occurrence of chocolate, yel- .

low and black color in domesticated races may only confuse the ontogenetic

production of characters with the gametie inheritance of factors. The

449

450 THE AMERICAN NATURALIST [Vou. XLVIII

Sturtevant (1912) showed that the results are also

consistent with the hypothesis that there is close or com-

plete linkage (genetic coupling) between yellow and

agouti. In principle this is the same as saying that when

yellow and agouti enter from different sides (mother and

father) they separate in gametogenesis, or in other words

they ‘‘repel’’ each other and behave, as I said, like

allelomorphs.

The numerical results would be the same whether

yellow and agouti are treated as though completely

linked or whether they are treated as allelomorphic.

What I had vaguely seen in my 1911 paper was clearly

explained in the following year by Sturtevant’s treat-

ment of the same data, to which he added that of Little

and Miss Durham.

Sturtevant showed, from an analysis of Miss Dur-.

ham’s results, in which she used ordinary gray (gray

‘‘oray-belly’’) mice, that her results are consistent with

the hypothesis of absolute linkage, or, on my interpre-

tation, with the hypothesis of allelomorphism. Sturte-

vant’s conclusions were promptly contradicted by C. C.

Little on the evidence furnished by some of his earlier

experiments, in which he obtained yellow, grays and

black (or chocolates) in offspring from yellow to black

(or chocolates). Such a result would be inconsistent

with Sturtevant’s hypothesis. Little also appealed to

certain experiments of Miss Durham, in which, he stated,

results like his own are given. Since Little has been

unable to get again his former results, but has obtained

evidence in favor of Sturtevant’s view, and since it is

clear that he misunderstood Miss Durham’s evidence,

his contradiction ceases to have any weight.

factorial hypothesis relates to those differentials that serve to separate

different types in inheritance and is not concerned with the problem as to

how those differentials produce their effects. Breeding experiments show

that gray differs from black by one differential, from yellow by another, and

from cinnamon by a third. So far as Mendelian segregation of these dif-

ferential genes is concerned it is es no consequence that the gray hair is

made up of a black, a yellow, and a chocolate band.

No. 572] MULTIPLE ALLELOMORPHS 451

After the publication of my own and of Sturtevant’s

paper I set to work to obtain crucial evidence in favor

of, or opposed to, the view that yellow and gray are

allelomorphic. Little, also, it appears, has carried out

some new experiments which he has recently published,

with the results just stated. My own data have been

_ ready for some time, but I have withheld them in order to

get a sufficient body of evidence to make the case con-

vincing, especially in the light of the possibility that the

crossing over might occur in one sex and not in the other.

For, if no crossing over occurred in the male, there

might be crossing over in the other sex, which would not

reveal itself unless the experiments were deliberately

planned so that both sexes are tested. This consideration

seems to have been overlooked by Little, for he has

omitted in his confirmatory paper to give the sexes of the

animals used. Without a knowledge of this relation even

his confirmation fails to confirm (as he supposes) the

view that he formerly combated.

Since Miss Durham worked with common gray and I

with gray white-belly, and both are ‘‘repelled’’ by yellow,

i. e., both are allelomorphs of yellow, it follows that these

two grays are also allelomorphic to each other.

The evidence that black belongs to the same series of

allelomorphs is obtained in the following way: If a given

yellow is mated to black, and yellow and gray offspring

are obtained, and if then the yellow offspring are mated

to black again and now give yellow and black only, the

proof is furnished; for in the first mating yellow and

agouti have repelled each other, and the yellow-bearing

gametes have united with the black gametes of the other

sex to give the yellow offspring. The second mating

shows that black is now repelled in turn by yellow and is

therefore allelomorphic.

This may be illustrated in the following way: Let

BY — yellow, b= black and B = gray. These three

factors may be treated as allelomorphs, then:

452 . THE AMERICAN NATURALIST — [Vou. XLVIII

Yellow B'B by | black bb.

Gametes of P, yellow BY-B,

Gametes of P, black b—b.

F BYb = yellow.

1 7

Gametes of F, yellow B-b.

-~ Gametes 5 _— black b-b

Yb = yellow.

Fy bb = black,

But if yellow and black and gray are not allelomorphic

the same matings should ie the following results:

Y’ = yellow. y’=not yellow = "i gray’? (not black).

Yellow Y’y 'BB or. Sng PAS

Gametes of P, yellow Y’B-B.

Gametes of P, pure black y'b-y’ b.

F Y’B y’b = yellow.

2 yB yb = gray.

Gametes of F, yellow Y’b—Y’B-y’b-y’B.

Gametes of pure black y’b-y’b.

yB y’b=gra

On the second assumption yellow, gray and black

should appear in the back cross. The former and not the

latter view is therefore consistent with the actual results.

THE SYMBOLS EMPLOYED

It is, of course, a matter of secondary importance what

system of symbols is followed. The requirements are

simplicity, consistency and suggestiveness, but one can

not always arrange to have all three at the same time.

The simplest scheme, for a system of allelomorphs like

these, would be to have some common letter to indicate

their relation and an exponent to suggest the different

characters for which each stands. If we take the symbol

b (black) for the common letter, and use capitals for

dominance, the allelomorphs will be:

b- = bink.

a pny y.

BW = gray white belly.

BY = yellow

i one preferred to take Y Kaw as the common letter

the series would be y’, y”, y'", Y’; or, if one preferred

No. 572] MULTIPLE ALLELOMORPHS ` 453

to take G (gray), as the common letter, the series would

be g’, g", G, g”. On the whole the first series seems to

me somewhat preferable.

The factor for cinnamon is entirely independent in

heredity of the preceding series of allelomorphs. This

factor may be represented by ci and its normal allelo-

morph by Ci. The formula for the wild gray would then

be Ci Ci, and that for cinnamon would be ci ci. Black

would be bb, and the double recessive cinnamon black

(or chocolate) would be bb ci ci. Chocolate is one of the

commonest types of domesticated mice and since I have

used it very extensively in my matings, its relation to the

other types may be further stated. It is known that if

chocolate is bred to wild gray, and if the gray offspring

that are obtained are then inbred, they give, in F,, the

following classes: 9 wild gray, 3 cinnamon, 3 black, 1

chocolate.

It is clear that chocolate is the double recessive type.

Of the two genes, that differentiate chocolate from wild

gray, chocolate has one in common with cinnamon and

the other with black. In other words, chocolate is cin-

namon black, and technically should receive this name.

THE EXPERIMENTAL EVIDENCE

Is There a Separate Factor for White-belly?

The first series of experiments was made in order to

determine whether the peculiarity of white-belly, shown

by the wild race of white-bellied grays, is due to a factor

that may be separated from the gray white-bellied mice,

or whether it is completely linked to gray (or allelo-

morphic to it). As wild gray house mice offer some

drawbacks in breeding work, I used cinnamon blacks

(chocolates). Gray white-bellied mice were bred to

2It is not possible to make a system of allelomorphs (in which the

‘feompounds’’ are serially epistate to each other) consistent entirely with

the system of nomenclature that I have suggested for the usual cases in

which mutant allelomorphs are contrasted with the normal allelomorphs of

the wild (or supposed original) type.

454 THE AMERICAN NATURALIST [Vou. XLVIII

chocolates.2 The gray white-bellied offspring were se-

lected and these were bred again to chocolate. The cross,

in regard to sex, was made both ways. If there is an

independent factor for white-belly that can separate from

the factor for gray gray-belly, then some gray gray-

bellied mice should appear. None were obtained, as the

following table shows. We must conclude either that

there is one factor that gives the gray white-bellied coat,

or else that the postulated factor for white-belly is so

closely linked to the gray factor that it has not sepa-

rated once in 100 times. Therefore unless such a sepa-

ration occurs it is simpler to assume one factor for gray

white-belly that is allelomorphic to black and to gray

gray-belly, ete.

TABLE I

Gray or Cinna- | | t

mon White- Black Chocolate White

Mating belly :

EE E be | 9

Gwb S by Chao.... 7 21 2 9 5 11 3 1

Ch 92 by Gwb d'.... 2 14 3 10 4 10

Totas. o eaS 9 35 5 19 9 21 3 1

Taking both crosses together, there are 44 grays to 54

blacks and chocolates, which approximate at least to

expectation. To these numbers I may add the follow-

ing data taken from similar experiments made for other

purposes in which one parent was, as before, gray white-

belly.

Gray-white Belly. Black or Chocolate.

m - 25 20 20

Presumably, therefore, the results may be treated as

though a single gene for gray white-belly exists. It will

be observed that the experiment has been made in two

ways, for at the time I was aware of the possibility that

crossing over, if it occured, might be limited to one sex.

-3At the time when the experiment was made all the gray white-bellied

mice were heterozygous for black and for agouti (including some with the

factor for cinnamon).

No. 572] MULTIPLE ALLELOMORPHS 455

We are justified, therefore, in treating gray white-belly

as an allelomorph of gray gray-belly, the former domi-

nating. If crossing over should occur, it might perhaps

only be realized in the gray or cinnamon mice, since it is

possible that the ticked condition of the hair (that is,

common to gray and to cinnamon) is necessary to realize

this condition. The expected crossover that would be

observed would be gray gray-belly mice. The contrary

class would then be black or chocolate mice ‘that carry

the factor for white-belly that might or might not show

the influence of the supposedly separable factor.

My white-bellied stock of mice had been killed after

my earlier results had been published, but Mr. B. B.

Horton had kept some of my original stock alive, and

from him I obtained a few of these mice in 1912 to carry

on the above experiments.

An extraordinary sex ratio appears in the next to the

last table, where there were 26 males to 76 females, ap-

proximately 1:3. The mice were entered when about

three weeks old. The sex was noted, but no special atten-

tion given to the determination. There is some chance

of mistaking the sex of young mice, but one familiar

with these animals can determine with certainty the

sex at three weeks if sufficient care is taken. I have no

reason to suppose that I made such errors which would

have to be frequent to give these results. If taken, then,

at their face value, the data seem to show that there is a

sex-linked ‘lethal gene present here. It is not linked to

any of the factors involved, and this is not expected,

since neither black nor agouti is sex-linked. If further

work confirms this conclusion (and I hold it as a provi-

sional conclusion until it can be further studied) we have

here the first evidence of a sex-linked gene in mice. A

sex-linked lethal should give a sex ratio of 14:29.

THE ALLELOMORPHISM oF YELLOW, Gray AND BLACK

The allelomorphism or ‘‘repulsion’’ of yellow and

agouti (gray) may be tested in various ways. One of the

456 THE AMERICAN NATURALIST [Vou. XLVIII

simplest tests is the following: Yellows were bred to

chocolates. The combination gave yellow and agouti off-

spring, when certain yellows are used, and yellow and

chocolate offspring when other yellows are used. Mixed

litters of yellow, agouti and chocolate do not appear.

Now when yellow and agouti appear in a given litter (as

above) the yellow parent must have carried agouti. If

her yellow gene ‘‘repels’’ the agouti gene, then none of

the yellow daughters should contain agouti genes, con-

sequently if such yellows are next bred to chocolate the

offspring should be only yellow and chocolate (or black)

and never yellow and agouti. This, in fact, is what my

experiments have shown. In the two following tables the

results of crossing yellows by chocolates are given by

litters. The yellows that were used at first were for the

most part heterozygous for gray white-belly, hence in

the earlier litters yellows and grays were generally ob-

tained. The yellow offspring of these earlier litters were

for the most part used in the later experiments, hence

the later litters are made up of yellows and chocolates.

The records (not given here) showed in every case that

yellow mice from litters of yellow and gray gave, when

bred to chocolate, only yellows and chocolates.

TABLE II

YELLOW ¢ BY CHOCOLATE 9?

LITTERS

| BE Re

iin eee sazos assiza zanz

ray.. | 5422 Pa: seas ailas oa

Chocolate....... ae: alt. lia 2848. 33}. 28/2: 3/428

yen ACEP EER BELICEPELEEPEEECEELEEL LEBEL

TABLE III

YELLOW 9 BY CHOCOLATE ¢

LITTERS

Yellow....4/1/4/4|/3|4|4|2|4|1/4/8/8|1/2/2/s/3/2\/2{1/1 34

Gray..... Pia ue st eve deer of co a, 1 Mt meee Me tio ee pee ee oe eon

Chocolate eae Fla] 2/2}../1],.}2/3}4/2/38]}1

Black.. | fefee | | : .|2

White. PEETER RT E i

4 Probably two litters combined,

No. 572] MULTIPLE ALLELOMORPHS 457

TABLE IV

SUMMARY OF LITTERS

Yellow and Gray Yellow and Chocolate

T Gray Yel. Choc.

101 78 70 67

The experiment is not demonstrative, however, unless

both the yellow daughters and sons are bred to chocolate,

for it might be that yellow and agouti are linked and

crossing over might occur in one sex and not in the other

sex. For instance, if we start again with yellow by choc-

olate, then if their yellow offspring contain agouti linked

to yellow that does not cross over in one sex, let us say in

the males, it follows that a yellow male bred to chocolate

would give only yellows and chocolates, for the agouti

gene would go with the yellow. Therefore, both sexes

must be tested. This essential element in the proof has

been overlooked by Little, for he fails to state whether his

test experiments were made with both sexes. In my

main experiments I have used yellow sons only, and the

tables are based on those data, but in a few cases I have

mated the yellow daughters (whose brothers were agouti)

also to chocolate and have found that these females give

only yellows and chocolates, which shows for both sexes

that no crossing over of yellow and agouti occurs.

A specific case will illustrate this point. A yellow

male was bred to a chocolate female and gave 5 yellow

and 7 gray offspring in two litters. One of the yellow

daughters was bred to chocolate and in four litters pro-

duced 11 yellows and 9 chocolates. A yellow grand-

daughter gave 9 yellows, 7 chocolates and 4 whites.

A yellow female bred to chocolate gave 8 yellows and

16 chocolates, but as I have no record of the preceding

generation, I can not be sure that this result is compar-

able to the last. It shows at least that a yellow female

gave only two kinds of offspring.

A ‘‘New Gray’’ Factor

A word may be added about the ‘‘new gray.’’ In the

original stock obtained from Mr. Horton there was a

458 THE AMERICAN NATURALIST [Vou. XLVIII

gray female with a not-pure-white belly. She was not

used in the main lines of the experiments described above.

But she was kept in stock and bred with chocolates.

About a year later I noticed in the offspring of a pair of

cinnamon white-bellied mice a few mice that looked like

chocolates, but which showed, on closer inspection, dis-

tinctly ticked hair. One of these new grays bred to

black (heterozygous) gave some chocolates, blacks, new

grays, and one very dark, almost black, mouse with

ticked hair. The female was bred next time to a house

mouse (gray gray-belly) and produced all -gray gray-

bellied offspring that had a dark coat, but not nearly so

dark as that present when the new gray is heterozygous

for black. Until further tests have been made it can not

be stated whether or not the new factor belongs to the

yellow-black system of quadruple allelomorphs.

5 The resemblance of this mouse to the rabbit ‘‘agouti-black’’ homozygous

for black is very striking (Punnett, Jour. of Genetics, II, 1912).

THIRTEEN YEARS OF WHEAT SELECTION

T. B. HUTCHESON

ASSOCIATE AGRONOMIST, UNIVERSITY OF MINNESOTA

INTRODUCTION

Ty 1901 the Minnesota Agricultural Experiment Sta-

tion planted a number of varieties of wheat from the

polonicum, spelta, turgidum, durum and vulgare types in

foundation beds in order to have specimens of these differ-

ent types always on hand for class work, hybridiza-

tion or demonstration purposes. Six of these varieties—

hedgrow (turgidum), Russian (vulgare), common speltz

(spelta), kamouka (durum), and Polish (1) and Polish -

(2) (varieties of polonicum)—have been grown continu-

ously since that time and an effort has been made to

improve them by selection. The method followed was

that introduced at this station by Professor W. M. Hays

and called the ‘‘centgener’’ method.

The centgener method consists, briefly, in starting with

individual plants, planting one hundred selected kernels

from each plant at equal depths and at equal distances

apart in separate plots. A plot of one hundred plants is

called a centgener. Careful notes are taken on the plants

in each centgener and at harvest time five or more of the

highest yielding plants are selected from which the seeds

for planting the next year are taken. From these five

best plants from five to ten of the best heads are selected

and thrashed together. One hundred of the largest and

plumpest kernels are then selected out of the seed ob-

tained by thrashing these selected heads, and these are

planted in the centgener test the next year. This work is

continued from year to year, each season the hundred

best kernels from the five or more best plants being

planted in succeeding centgeners. —

459

460 THE AMERICAN NATURALIST [Vou. XLVIII

In 1908 an experiment was planned with the object of

developing a strain of wheat which would have a minimum

amount of culm exposed between the base of the spike and

the upper leaf sheath, or in other words, to produce a

short-necked variety of wheat. The ultimate purpose of

reducing the neck lengths was to reduce the area of the

stem exposed to the black stem rust. Since this rust

ordinarily does little damage to that portion of the culm

enclosed in the leaf sheath, it was thought that a short-

necked wheat would be more likely to escape serious

damage from stem rust than a long-necked kind. For

this work individual plants were selected which had short

necks and the seed from these were planted in separate

centgeners. Each year at harvest time ten or more plants

which appeared to the observer to have the shortest necks

were selected from each centgener and measurements of

their neck lengths were made and recorded. One hundred

kernels were saved from these shortest necked plants

each season for subsequent centgeners, thus making a

continuous selection for short neck lengths.

The data derived from the above experiments seems to

throw some light upon the much-discussed question as to

whether or not selection within a pure line can increase

yield or change type enough to make it a desirable prac-

tise from the practical breeder’s standpoint. In both of

the experiments, we have the requirements for a pure

line satisfied. Wheat is a normally self-fertilized plant.

Each centgener was started from a single head in 1901

and these heads have bred true to type ever since.

The long period of years over which this experiment

has extended makes the data particularly valuable. One

of the adverse criticisms to most pure line work is that

it has not extended over a long enough period of time.

Thirteen years are about as long as any practical breeder

would be apt to keep up selection on one pure line and

covers the longest period of continuous selection for a

self-fertilized plant yet reported.

Another criticism to pure line investigations is that in

No. 572] THIRTEEN YEARS OF WHEAT SELECTION 461

many cases it has not appeared certain that the material

studied was a pure line. Since the plants have bred true

to type throughout the whole period of study, it is obvi-

ous that this criticism will not hold for the data herein

presented.

The work has been conducted at this station under the

direction of Professor W. M. Hays from 1901 to 1905,

under Professor E. C. Parker 1905 to 1908, under Pro-

fessor Andrew Boss from 1908 to 1911 and under Pro-

-fessor C. P. Bull 1911 to 1913.

m3 N

= 2. [ N

gS IEA |

z AN I

\ "A SE EE /

2 Atti l

a F , eae: |

a p LIL X

m | A

Z VY

Puate I. Average yield per plant for all varieties. X-X, fitted straight

line.

SELECTION To [INCREASE YIELD

The varieties studied, the average annual yield of each

variety and the average yield per plant for the six vari-

eties under test are shown in Table I. In the years 1903

and 1904 weather conditions were unfavorable, making

it impracticable to obtain correct average yields per

plant, so data for these years were omitted. However,

selections of the best plants were made in these two

seasons as in the others and the best seed from them

were kept for planting, so the continuous selection for

increased yield was uninterrupted.

462 THE AMERICAN NATURALIST (Vou. XLVII

TABLE I

SHOWING YIELD PER PLANT— YEARS 1901-1913

Yield per Plant in Grs.

Name of Variety % j i

1901 | 1902 | 1905 | 1906 | 1907 | 1908 | 1909 | 1910 | 1911 i 1912 | 1913

| | | |

FACUITOW . choos oc eee 3.10} 2.80 3.69 2.48) 1.27) 3.75 2.49) 2.55] 2.02) 99 3.67

Riman fos Sea re es 1.00} 1.70} 3.57 1.96) 1. 71) 2.74| 2.71) 2.17 1.95} 1.37} 2.70

BOC. 5 os cece eo ee 2.40} 1.80) 3.99 2.99, 1.38 3.38) 2.40, 2.86, 2.01} 2.14) 2.59

SSHBIOUME ioc ccs a ae ee 1.50} 2.50} 1.99; 2.69) 1. 39 3.31) 2.19} 2.48) 1.67 1.35| 2.16

Po (Be ica. ei a4 ye Bee -80] 1.30) 2.52, 2.04) 1.03) 1.48) 1.91) 1.70) 1. 56, 1.12] 1.74

POMS CA) Dea NEINA, 1.10} .95| 2.83) 1.97 1.26 . $; 61 Lal}. 78 51} 1.33

i |

PA T A ET EEA 1.65| 1.84| 3.10 2.35 1.34 2.93 2.22 2.18) 1.83 ioe 2.36

SELECTION TO Increase HEIGHT

The average height of the plants for each year of the

test is shown in Table II. Though no attempt was made

to select for increased height, since a number of workers

have shown that height in the small grains is distinctly

correlated with yield, it is natural to suppose that the

selected plants were among the tallest as well as being the

highest yielders of each year’s crop. When this experi-

ment was begun, it was not known that height and yield

TABLE II

SHOWING AVERAGE HEIGHT PER PLANT—1901-1913

Height in Inches

Name of Variety T

1901 | 1902 | 1905 | 1906 | 1908 1909 | 1910 | 1911 | 1912 | 1913

Hedgrdee 2... 961-41 | 42 Ea | 46 l 40 | BB t Al 86 1. 86

Roslan... n... 34 | 37 O | 38 44°) 41 | 38 | 36 | 35° | 33

Soltero: S438 Oe | AF 4 at l 80) 40 80: 1 3S

Kamouka........ 36 | 34 | 34 | 38 | 40 | 40 | 32 | 38 | 36 | 33

Polish (1)....:...| 40 | 38 | 41 | 38 | 42 |.42 | 33-| 39 | 88./| 83

Polak (0). 6.4 cc 98480 He SP BT ois ss 35 | 84: | 3$ | ar r 3

Average......... 35 | 36 | 38 | 39 | 43 | 40 | 34 | 39 | 36 | 33

were correlated, so the figures on height were kept merely

as a matter of general interest and with no idea that they

would have bearing on the problem. Among those who

later found height correlated with yield are Deneumostier

(710),1 Love (711),2 Myers (712),? Leighty (’12)* and

1 Deneumostier, C., ‘‘ Correlations in Wheat,’’ Ann. Gembloux, 20, No. 5,

1910.

No. 572] THIRTEEN YEARS OF WHEAT SELECTION 463

SELECTION TO DECREASE Necx-Lenctus

The result of the selection for short neck-lengths is

shown in Table III. This is a clear illustration of how

misleading short-term experiments may be. Had the

experiment been discontinued at the end of the third year,

the figures would have indicated that it was possible to

modify this character very rapidly by selection. How-

ever, in the following two years the neck-lengths seemed

to revert to the mean of the pure lines, and the last year

they were actually longer than when the experiment was

started. The reduction in the first three years was prob-

ably due to growing conditions.

TABLE III

SHOWING RESULT OF SELECTION FOR SHORT NECKS

Average Neck L ngth in Curve

| 1909 1910 1911 1912 | 1918

Dei ed Beever uh Fok 1.86 24 7.34 | 9.54

ite Me ee C 6i 1.12 .79 Gis | 218

bie Ce a | 58 1.65 ‘56 753 |

Sete Di | B22 2.08 59 10.47 | 13.82

Discussion

From the data presented in these tables, it is evident

that there has been no permanent gain for these thirteen

years of selection either in yield per plant, height of

plant, or shortening of neck-lengths. The expected sea-

sonal variations occur. A comparison of the yield of

Haynes Blue Stem, which is grown extensively in Minne-

sota, and was continued in the variety test without any

attempt at selection throughout the whole period, with

Hutcheson (713).5

2 Love, H. H., ‘‘A Study of the Large and Small Grain Question,’’ An.

Rep. Am. Br. tele: 7: 109-118, 1911

3 Myers, C. H., ‘‘ Variation, Oaren noa and Inheritance of Characters of

Wheat and Fea’ Cornell University Thesis, 1912.

4 Hutcheson, T. B., ‘‘Correlated Characters in Avena sativa, with Special

Reference to Size of Bd Planted,’’ Cornell University Thesis, 1913.

5 Leighty, C. E., ‘‘Studies in Variation and Correlation of Oats, Avena

sativa,’’ Cornell University Thesis, 1912.

464 THE AMERICAN NATURALIST [Vou. XLVII

the average yield of the selected varieties, is shown in

Table IV. The average yield in bushels of the Haynes

Blue Stem is also platted in comparison with the average

yield of the selected varieties in Plate II. In 1912 a

severe hail storm injured the variety plats so much that

2 AN N

z ENSEN z

= LiNE EAEN /

<2. MANIN ee EN / .

a EF, NA | } N pi

aha ery CNT FA

ARE Y7 Kit z

= 4% Meaty VY

> N

| —— ewe mes

Prate II. Comparing seasonal fluctuations in selected lines with unse-

lected Blue Stem. Solid line, yield per plant in grams for selected lines;

dashed line, yield in bushels per acre for Blue Stem

it was thought best not to include the yield of the Haynes

Blue Stem for that year. This gives an incorrect appear-

ance to the curve, as it was extended just as if this year

was present and midway between 1911 and 1913 in yield.

It will be noticed from Table IV and Plate II that the

TABLE IV

COMPARING SEASONAL FLUCTUATIONS IN SELECTED LINES WITH UNSELECTED

LUE STEM

1901 | 1902 | 1905 | 1906 | 1907 | 1908 | 1909 | 1910 | 1911 | 1912 | 1913

Yield in grs. per

plant for selected

HEM. cele: 1.65] 1.84) 3.10} 2.35; 1.34] 2.93) 2.22) 2.18) 1.83) 1.24 | 2.36

Yield g ong per

acre blue

Mem cc eS. 22.9 |23.9 |30.4 |24.00'21.00|26.00/26.6 |24.6 24.2 |..... 23.2

No.572] THIRTEEN YEARS OF WHEAT SELECTION 465

fluctuations from year to year agree very closely. These

data indicate that increased yield is due to favorable

environmental factors and not to improvement by

selection.

A comparison of the yield of each variety for the first

five years of the test with that of the last five years is

shown in Table V. The data in this table show that

there is no significant difference in yield for these two

periods. In Russian and Polish (1) there is a slight

increase in favor of the latter period, but in the other

four varieties there is just as much decrease for this

period. However, there is not enough difference in any

case to indicate either permanent improvement or de-

crease in yield. As far as these varieties are concerned,

it seems that selection has brought about no permanent

improvement.

TABLE V

COMPARING THE YIELD OF THE Jima YEAR ees WITH THAT OF THE LAST

YEAR PERI

ist 5-year Period Last 5-year Period

Name of Variety

Height Yield Height Yield

Hedgrow 41.6 2.67 38.4 2.34

38.0 1.99 35.4 2.18

Speltee es 40.0 2.51 39.2 2.40

BAMOURA. o is Vi. s 36.4 2.01 35.8 1.97

Posh soys eae: 39.8 1.54 37.4 1.61

Eola 19) i obras Sue be 33.4 1.62 33.4 131

t AA oa et 38.2 2.06 36.5 1.97

A curve of the yields of the six varieties under con-

sideration for the thirteen years of the test was plotted

and a straight line was fitted to it, by the method of the

least squares, to indicate the trend of the yield. This

curve is shown in Plate I. There is a slight downward

tendency in this straight line, but it is not enough to indi-

cate a tendency toward decrease in yield. The line fitted

to the curve of height (Plate III) also shows a slight

tendency downward.

The data herein cited are not sufficient for definite con-

466 THE AMERICAN NATURALIST [Vou XLVIII

clusions. However, the indications are that from a prac-

tical breeder’s standpoint permanent improvement in

pure lines in small grains, if possible, is certainly not

rapid or apt to be very marked. Thirteen years of selec-

w4 Nf W

z L \ I

a7 + Am

= ; ~

3 Vy l

pom -_ = —

a Ba Ba = =

Puate III. Average height of all varieties. X-X, fitted straight line.

tion covers considerable time and expense, and, as far as

can be seen from the varieties reported in this paper, it

has resulted in no permanent improvement. This would

suggest that some other line of improvement must be

sought. It is probable that much more rapid progress

could be made by segregating pure lines from mixed

populations and combining the desirable characters of

these lines by hybridization.

PATTERN DEVELOPMENT IN MAMMALS

AND BIRDS

GLOVER M. ALLEN

Boston SOCIETY or NATURAL HISTORY

PARTIAL ALBINISM IN Witp MAMMALS

Partially albinistie individuals of species that normally

are wholly pigmented, occur frequently in a wild state,

and almost any large series of a given species may con-

tain a few. I have examined many such, in which it was

perfectly evident that the white mark was due to areal

restriction of some one or more of the primary pigment

areas just as described in the various domestic species.

It is apparent that the white markings in both are quite

comparable, but in species under domestication no agency

seems present whereby such pied individuals are elimi-

nated, whereas in a wild state the sudden acquisition of a

large amount of white in an individual would not only

render him too different from his fellows, but might put

him at a disadvantage because of a conspicuousness to

which as a species he had not yet become accustomed.

There are many other species in which, as we now see

them, white markings form a permanent and normal part

of the pattern. Among those in which these white mark-

ings are few or simple, it is often evident that they are

merely primary breaks between the pigment patches that

have become more or less fixed by long periods of selec-

tion, whether natural, sexual or otherwise. As I shall

endeavor to show, there are species in which a beginning

has already been made towards the development of a

pied pattern, though it has not yet become well fixed.

Still other species show a more complicated white and

pigmented pattern, the white portions of which can not

readily be derived from primary breaks alone. Such I

take to be highly developed patterns and make no attempt

467

468 THE AMERICAN NATURALIST [Vou. XLVIII

to analyze them here. Examples of this type are seen in

the zebra, the spotted skunks (Spilogale), the striped

weasel (Ictonyx). Probably more than one factor is

responsible for some of the combinations of stripes and

spots seen, for example, in certain spermophiles (Citellus

13-lineatus), but I shall not now attempt a discussion of

these.

One of the most frequent manifestations of pigment

reduction in mammals is the presence of a white spot in

the normally pigmented forehead. This is due primarily

to the reduction of the ear patches, which fail to meet at

their median edges. Perhaps, too, the apparent loss of

the crown patch in some mammals still further tends to

lessen the amount of pigment production at this point.

Rabbits and hares very often have more or less white in

the forehead, but none of the species has developed this

sufficiently to make it a permanent mark. Moseley in his

. ‘“‘ Naturalist on the Challenger,” speaks of a ‘‘black

variety” of wild rabbit—doubtless introduced—‘‘with a

white spot on the forehead’’ as occasionally found on

Teneriffe, Canary Islands, but this mark is common,

and I have seen it in such widely sundered species as the

eastern varying hare of New Hampshire and the black-

necked hare native to Java. A specimen of Leisler’s bat

(Nyctalus leisleri) in the Museum of Comparative Zool-

ogy has a white spot in the middle of the forehead and

another on the mid-ventral line of the abdomen—the first

a primary break between the ear centers, the second

probably a ventral primary break between those of the

sides. Among the Insectivora, the West Indian Solen-

odon paradoxus has a white patch at the nape of the

neck which has become a permanent part of its pattern.

It is clearly the enlargement of a primary break sepa-

rating the ear patches and neck patches on the median

dorsal line. It is a fact of much interest that in a con-

= siderable series of this species in the collection of the

Museum of Comparative Zoology hardly two have it

developed alike, but it varies from a few white hairs to

No. 572] PATTERN DEVELOPMENT 469

a large patch 15 X 10 mm. wide. Evidently it has not

yet become precisely defined in its limits, though now a

permanent mark of the species.

White marks in the forehead are common among the

species of the Mustelide or weasel family. A narrow

white median line is present in the Javan mydaus and in

the skunks (Mephitis) as part of the permanent pattern.

In the badger (Taxidea) a white line is not only pres-

ent on the forehead, but it is often extended medially so

as to separate the pigment patches of both sides of the

body. In the New York weasel (Mustela noveboracensts)

of the eastern United States a few white hairs are often

present on the forehead, and other instances could be

multiplied. Among monkeys, a white spot on the nose is

present in some species of Lasiopyga, and in an allied

genus Rhinostigma, it is elongated vertically to form a

white streak.

A yet more illuminating case is that of the Muskeget

Beach mouse (Microtus breweri) a derivative of the

common brown meadow mouse of the New England

mainland. On this island of white sand off the Massa-

chusetts coast, a pale variety has developed which is very

distinct from that of the neighboring shores. Not only

is it a paler race, but albinism also has begun to appear,

so that occasional individuals have a white fleck between °

the ears, showing the drawing apart of the ear patches.

Of a series of 62 specimens in the collections of the

Museum of Comparative Zoology and the Boston Society

of Natural History, no less than 13 had such white flecks,

and one had in addition a white spot just in advance of

the shoulders, marking the line of separation between

neck and shoulder patches. In our studies on the hered-

ity of coat colors in mice, Professor Castle and I dis-

covered (Allen, 1904; see also Little, 1914) that the pied

condition is recessive in the Mendelian sense towards

the self colored, so that partial albinos bred to wholly

pigmented mice produce in the second generation, if

interbred, 25. per cent. of spotted young. The figures

470 THE AMERICAN NATURALIST [Vou. XLVIII

given above (13 in 62) are near this in case of the Muske-

get mouse, but the matings are of course more promiscu-

ous. The case is interesting in connection with the

studies of Ramaley (1912) and Pearl (1914), tending to

show that in a mixed population the recessives may in-

crease so as to exceed the dominants. Although the

spotted mice do not, in case of this species, exceed the

unspotted individuals, they nevertheless are of far more

frequent occurrence than they are in the mainland repre-

sentatives of the species. This accords with the fact that

island-living mammals are very commonly albinistic, and

the cause is doubtless that the population is much more

inbred, so that the recessive condition of partial albinism

is more likely to be propagated than if successive genera-

tions have a wider range over which to spread. It seems

probable that heredity will tend to increase the propor-

tion of spotted mice of Muskeget, and that if this condi-

tion is disadvantageous, a large part of the spotted indi-

viduals will be killed off, yet in the course of time they

may become adjusted to this condition and will survive

in increasing proportion till the white mark becomes

characteristic of all the animals. Cory (1912) records

the capture of seven muskrats at Hayfield, Iowa, all of

which were uniformly marked, having a white ring around

the neck and the entire underparts, feet, and end of tail

white. I can think of three causes influencing the status

of such white markings. These markings may be in-

herited in a purely automatic way as unit characters;

but if thus inherited they may be (1) increased through

selection, natural or sexual; or (2) eliminated by the same

agent; or (3) they may be, at first, of’no influence at all

in the economy of the animal and persist or not, accord-

ing as they are heritable.

I have mentioned that island mammals tend to be more

albinistic than their mainland representatives. Other

cases may be mentioned, as the common squirrel (Sciurus

vulgaris leucurus) of Great Britain, which differs nota-

bly from that of the continent in having frequently a

No. 572] PATTERN DEVELOPMENT 471

white or whitish tip to the tail, often for one half its

length. A similar white tip is occasionally seen in our red

` squirrel (S. hudsonius) as an albinistic mark, and is due,

of course, to the terminal restriction of the rump patches.

The deer of Whitby Island, Puget Sound, are said to be

much marked with white, and sundry marsupials of

Papua as well as the monotreme Zaglossus are subject to

white markings. In the cuscus (Pseudochirus) the pig-

ment is sometimes restricted to small patches and round

spots scattered on the back, those in the region of the

shoulder of a different color from those of the side and

rump patches. Another instance is that of the white-

footed mouse of Monomoy Island, Massachusetts, the

mid-ventral parts of which are pure white to the roots of

the hairs, an albinistic condition to be clearly distin-

guished from that in which the belly appears white, but

only because of the white tips to the hairs whose bases

are dark-pigmented.

The restriction of the rump patches so as to produce

a white tail-tip is common among mammals. It is found

in occasional specimens of many species as the shrew mole

(Blarina), Brewer’s mole (Parascalops), the meadow

jumping mouse (Zapus), the white-footed mouse (Pero-

myscus), and squirrels (Sciurus). In some it has be-

come developed as a permanent and characteristic mark,

as in the woodland jumping mouse (Napeozapus), the

red fox (Vulpes), such genera as Hydromys, Tylomys,

the Virginia opossum (Didelphys virginiana), the tree

kangaroos (Dendrolagus). In many others a pure white

belly is developed through ventral restriction of the

shoulder and side patches.

Among ungulates the break between the ear patches

has been developed to form a broad white blaze from

forehead to nose in case of the blesbok (Damaliscus

albifrons) of South Africa and in related species in

East Africa. The chevron-mark on the forehead of cer-

tain antelopes is possibly a specialized development of

the same thing.

472 THE AMERICAN NATURALIST [Vou XLVIII

White buttock patches are present in several unrelated

ungulates—as the pronghorn (Antilocapra), the wapiti

(Cervus canadensis), and the Rocky Mountain sheep:

(Ovis canadensis). Probably these are the result of

restriction or total inactivity of the pigment patches

covering the rump.

Fic. 42a. DIAGRAM SHOWING THE PIGMENTED PATCHES OF A PARTIALLY ALBINO

Among the deer family white is generally confined to

the under surfaces and the primary white breaks have not

been developed to form patterns. Albinistic deer are

fairly common, however, and in Fig. 42a I have made a

tracing from a photograph showing the side of a par-

tially albino doe in which areal restriction of pigment has

taken place in such wise that the primary patches are

all indicated, and separated from those of the opposite

half of the body by a median dorsal white line. The ear

and the neck patches are joined, but a few small islands

of pigment are left here and there, much as in cows.

In the young of many deer and in the adult of such

species as the axis deer, a spotted pattern is developed.

No. 572] PATTERN DEVELOPMENT 473

There is an obvious tendency for the spots to become

arranged in longitudinal rows, and intermediate stages

may be found in which they coalesce to form broken lines,

There is little doubt that the complete white stripes

occurring in part of this pattern were formed originally

through the coalescence of rows of white spots. In the

tapir a somewhat similar spotted pattern is found in

the young, while the adult Malayan tapir has lost the

shoulder and side patches, producing thus a white-bodied

animal, pigmented to the back of the foreleg and on the

buttocks and hind legs. Among the ground squirrels

(Citellus) a beautiful series can be picked out showing

the transition from a uniform grizzled mixture of ticked

hairs to indistinct spotting, then rows of white spots, and

finally broken and complete longitudinal stripes. The

production of these stripes I believe to be due, not to the

development of breaks between the primary pigment

patches, but to the action of a factor which is the negative

of the so-called ‘‘ English’? marking in rabbits, so that

instead of the development of scattered small pigments

spots there are formed, instead, spots without pigment.

That it is possible to evolve a striped pattern from spots

through selection, I have no doubt, and indeed, it is gen-

erally believed. On the other hand, it is quite possible

that the converse may happen, and spots result through

the breaking up of stripes. According to the experiments

of Professor Castle and Dr. MacCurdy, however, it seems

to be a difficult matter to fix a given marking by rigid

selection, yet it must be admitted that a few years’ work

even of careful breeding is nothing in comparison with

the age-long selection that may have been at work on the

species: That it is a difficult matter to produce a given

pattern is further evidenced by the fact that in many

species in which white markings regularly occur as part

of the pattern, these are subject to great individual

variation in their extent, showing that sie are even yet

not wholly definite.

It was formerly urged against a dpetvine

474 THE AMERICAN NATURALIST [Vou. XLVIII

that we do not now see its processes in action, that species

are stable and subject to very little variation. This view,

however, was found to rest on faulty observation, for,

though some species are fairly stable, others are very

plastic and exhibit before our eyes various steps in

development. So in case of the development of a partic-

ular pied pattern, it is possible to see in certain species

the actual course of its formation. Among mammals,

the Mustelide or weasel family show several instances in

is A Oe an 7 4$

Fics. 43-48. DIAGRAMS SHOWING RESTRICTION OF PIGMENTATION ON THE VEN-

TRAL SURFACE OF MINKS (Mustela Aani,

point. The common mink (Mustela vison) of north-

eastern North America is now in process of developing a

pure white under side, such as is present in the New York

weasel (M. noveboracensis) or the smaller Bonaparte’s

weasel (M. cicognani). The diagrams shown in Figs.

43-48 are from the fine series of mink in the collection of

the Museum of Comparative Zoology and depicit the

No. 572] PATTERN DEVELOPMENT 475

under side of the specimens. In the large coastal race of

mink found from southern Maine to the Carolinas (M. v.

lutreocephalus), the entire pelage is usually brown, ex-

cept for the chin which is white. Occasional white marks

are present in some specimens along the mid-ventral

line of the throat and chest, and between the hind legs.

In the smaller typical M. vison of northern New England

northward the white marking is apt to be more extensive,

and in no two individuals exactly alike. The diagrams

show the ventral markings of a few specimens from New

England and Nova Scotia. In Fig. 43 the amount of

white is very small. The chin spot, which represents the

beginning of a break between the two ear patches at their

antero-ventral extremity, is always present and has be-

come now a fixed mark of the species, though variable in

extent. A slight break in the center of the chest shows

where the two shoulder patches have failed to meet, and

a white spot at the anal region indicates a like restriction

of the rump patches. Similar spots appear mid-ventrally

in Fig. 44, with the addition of a few white hairs, medially

at the upper throat, where the ear and neck patches join,

and a few more on the lower throat at the line of union

of the neck patches of opposite sides. In Figs. 45 and 46

no break is present on the abdomen, but in the former

figure, a large transverse break has appeared on the

upper throat where the ear patches fail to unite with the

neck patches and with each other, and a median line runs

forward to join the white of the chin, showing the greater

restriction of the ear patches ventrally. An imperfect

separation of these patches along the center of the throat

has taken place in Fig. 47, and a more considerable break

occurs in the same place in Fig. 46. In the Pacific Coast

mink (Mustela vison energumenos) a well-developed

white patch on the chest is rather characteristic, some-

what larger than in Fig. 45. This is due to the ventral

restriction of the shoulder patches which fail to meet

below. In Fig. 46 this white area is seen with a tongue

extending upon the center of the lower throat, and on to

476 THE AMERICAN NATURALIST [Vou. XLVIII

one fore leg, as well as in the mid line of the thorax, mark-

ing nearly the anteroposterior limits of the shoulder

patch. The neck patches are not separated in this figure

but have become so in Fig. 48, so that a continuous line

of white runs from chin to chest. In Fig. 47 the shoulder

and the side patches have both failed to join ventrally,

and thus a broad white line is formed down the center of

the: belly from the conjoined neck patches to the rump

patches. If all these breaks were to be present in a single

animal, there would be a narrowed white area along the

entire ventral side of the body from chin to anus, extend-

ing on to the lower side of the fore legs. Practically

this condition exists in another species of the same

genus, Streator’s weasel (Mustela streatori) of the

Pacific Coast, in which the throat, chest and belly are

white but the width and boundaries of the white area are

very variable in different individuals. It is therefore in

a stage beyond that which the minks have reached, yet it

has not attained the stage in which the white area is of

definite and rather constant bounds, as in certain other

weasels, for example Mustela noveboracensis, in which

the white, of the belly extends nearly or quite to the

lateral border of the body, but in different individuals

varies slightly, and M. cicognanii, in which the white area

of the belly constantly extends to the lateral boundary of

the venter from throat to anus. This is the condition

toward which the mink is tending.

Another interesting case in which a pattern mark ap-

pears to be evolving through the fixation of a primary

break between pigment patches is that of the so-called

tayra of South America (Tayra barbara) a large Muste-

lid. The Central American race (biologie) of this animal

is wholly black, but the typical subspecies of Brazil and

northern South America is subject to a varying amount

of reduction in pigmentation. Curiously, this takes place

at the posterior end of the neck patches or at the anterior

part of the shoulder patches. Three of five specimens in

the Museum of Comparative Zoology are marked in this

No. 572] PATTERN DEVELOPMENT 477

way. All have a triangular patch of white at the base of

the throat ventrally, as a break between neck and shoulder

patches and a partial separation of the neck patches

from each other. Each has a dorsal mark of white; in the

first a narrow linear break between the shoulders; in the

second a broader transverse mark, and in the third a

square patch of white occupying nearly the width of body

between the shoulders to the base of the neck. The white

throat marking increases in extent from first to third,

just as does the dorsal marking. Probably in time this

white mark, now of irregular size and appearance indi-

vidually, will become a permanent part of the pattern. In

this animal the entire head and neck are a grizzled gray

as far back as the posterior limit of the neck patches, and

the rest of the body is black. This, then, shows that the

pigment patches of head and neck are differentiated in

color as well, from the patches of the rest of the body.

The occurrence of white markings in the back is relatively

uncommon in mammals, though white on the under sur-

faces is common, and, as shown by Mr. Abbott H. Thayer,

may be of real service to the animal as a factor in con-

cealment.

In the development of white pattern-marks, the evi-

dence seems to show that these come in at first as small

and fluctuating spots, which may be of little effect in the

economy of the animal. Their further development might

lead to the extinction of the species if they render it too

conspicuous to enemies, unless the species at the same

time makes use of them or accommodates itself to their re-

vealing effect. Often, no doubt, they may not be a source

of danger at all. A case in point may be that of Sciurus

finlaysoni, a Malayan squirrel, most of the individuals of

which are largely marked with white, and of which speci-

mens may be found side by side, varying from an almost

entirely pigmented condition to one of completely white

coat and black eyes. White squirrels are occasional in

other species, as albinos, but these rarely survive more

than a generation in the cases I have known, whereas

478 THE AMERICAN NATURALIST [Vou. XLVIII

Finlayson’s squirrel seems to have accustomed itself by

gradual stages to the white condition, so that it is prob-

ably not at a great disadvantage by reason of its

whiteness.

Piement Patcues In Bms

In birds the same primary pigment patches seem to be

present as in mammals, and they are homologous in the

two groups. In defining the extent of the pigment patches,

however, allowance must be made for the fact that the long

feathers may cover a part of the body remote from their

origin. The distribution of the feathers or the pterylosis

of the species in hand must also be r bered. In

order to arrive at the true interpretation of the patches,

it is necessary to consider the pigment as projected back

from the vanes of the feathers to the part of the body

at their bases. By so doing, it becomes evident that a

feather variegated with pigmented and unpigmented (or

white) areas indicates none the less that the feather

arises from a place of pigment formation. It is only a

wholly white feather or patch of feathers that can be

considered albinistic in the sense here intended. The

factor determining the intermittent formation of pig-

ment in the individual feather is probably a wholly

different one from that determining the presence or ab-

sence of pigment formation at certain places on the

body, though not necessarily different except in its inter-

mittent action.

In the domestic pigeon of our streets and buildings,

we have a species that in its wild state is normally fully

pigmented except for a white rump patch. Under semi-

domestication it has developed partial albinism to a large

degree, so that it is possible to obtain a complete series

representing on the one extreme a totally pigmented

bird without a trace even of the white rump patch, and

on the other extreme a bird of pure white plumage. A

few of the intermediate stages in areal reduction of pig-

mentation are shown in Figs. 49 to 53, selected from

birds raised for the market and, so far as known, not

No. 572] PATTERN DEVELOPMENT 479

bred for pattern. The first steps in reduction are shown

in Fig. 49. Here there is seen first a crescentic band of

white feathers passing from eye to eye around the occi-

put. This is a primary break marking off the crown

patch posteriorly. This patch in birds, in contrast to its

development in mammals, is the main patch of the head,

4g so oF

Fics. 49-53. DIAGRAMS SHOWING PIGMENTATION IN THE DOMESTIC PIGEON.

covering the area from the base of the bill to the eyes

and occiput. In Fig. 50 its posterior limit is similarly

defined by a primary break separating it from the neck

patches, and although it does not extend forward quite

to the eye in this specimen, it shows a beginning of sepa-

ration from the more lateral ear patches by virtue of the

indentations on each side posteriorly. In Fig. 52, the

crown patch is shown slightly reduced in extent and

wholly separate from the ear patches, which have become

inactive altogether. In Fig. 53 it has dropped out with

the latter. It is evident then that by greater or lesser

480. THE AMERICAN NATURALIST [Vou. XLVIII

reduction of the crown patch alone it is possible to pro-

duce a pigeon with a mere white spot at the back of the

head, one with a white stripe from the base of the beak

through the eyes to the back of the head (or some part

of such a stripe) to a pigeon in which by the total reduc-

tion of the patch, the entire top of the head is white.

Such specimens can be found in most any miscellaneous

flock. There is a tendency often for the patch to be irreg-

ularly broken, sometimes divided almost into two parts,

a result of the pterylosis to some extent.

The ear patches in pigeons, and probably in all birds,

are rather insignificant, and the smallest of all the pri-

mary pigment areas. “They include the feathers from

the posterior angle of the lower mandible to the angle of

mouth and thence back, including the ear coverts. I do

not feel sure that the patches of opposite sides may not

join on the chin, but the present evidence tends to show

that the chin is pigmented by a forward extension of the

neck, patch, which, under reduction, often leaves a small

island of pigment between the mandibular rami. In Fig.

50 the neck patches are seen to have broken away ante-

riorly from the crown and ear patches and the separa-

tion of the latter from the crown is indicated by deep

reentrants along the line of the separation. In Fig. 51 a

remnant of the ear patch of the left side alone remains

in dorsal view, consisting of a small tuft of pigmented

feathers at the fore end of the aural area and a single

pigmented feather just behind it. In this specimen there

are a few pigmented feathers on the chin as well, which I

take to be an isolated bit of the neck patches.

The neck patches are bilateral in origin, and pigment

the entire throat and neck back to a point corresponding

to the base of the neck vertebre. They meet the crown

patch and separate the ear patches at the occiput. In the

domestic pigeon the neck patches correspond very closely

to the area of differentiated feathers that give the metal-

lic reflections. In the reduction of this area it is common

for the anterior part of the throat to be white, and then a

No. 572] PATTERN DEVELOPMENT 481

break occurs between the neck patches and those of the

head as in Fig. 50. Posteriorly the neck patch under

reduction may become separated by a white ring at the

base of the neck, from the shoulder patches as in Fig. 50.

The ultimate centers of these patches seem to be in the

pigeon well back on the base of the neck. These are

shown, of small extent, in Fig. 51, as two small areas of

pigmented feathers, one on each side of the base of the

neck. In Fig. 53, further reduction has taken place, so

that the patch of the left side only remains as a small

center. In Fig. 52 there is a large median dorsal patch,

which, as in mammals, may represent the two centers of

opposite sides which even under much reduction have not

in this individual become divided medially.

A very common manifestation of pigment reduction in

pigeons is to have the primaries or some of them white,

as in Figs. 49 or 50., This indicates a failure of pigment

to develop at the extremities of the shoulder patches,

just as in mammals white forefeet mark a slight reduc-

tion of the same areas. It is a fact of much interest that

in the guinea fowl (Numida), which has been under

domestication but a short time comparatively, a distinct

breed has arisen in which this same reduction of pigment

is present, resulting in a speckled bird with pure white

primaries and often a pure white area on the breast. In

the pigeon, further reduction cuts off a narrow ring of

pigment encircling the breast, or, it may be, broken in the

mid-ventral line. This ring represents the reduced

shoulder patches, and is to be seen in many wild species

as a permanent part of the pattern. The white collar at

the base of the neck in Fig. 50 marks the separation

between the neck and the shoulder patches at the ante-

rior border of the latter. In other specimens the patches

are separated medially by a white area down the back.

The ultimate centers of these patches seem to be near

the elbow or on the upper arm at the base of the tertiaries,

as seen in Figs. 52 and 53.

482 THE AMERICAN NATURALIST [Vou. XLVIII

The side patches are rather small and seem to center,

as in Fig. 51, near the groin on either side. They pig-

ment the belly back of the breast area included by the

shoulder patches, and extend on to the hind legs as well.

In a specimen before me, the shoulder patches pigment

the bases of the wings and the entire breast correspond-

ing roughly to the length of the sternum, and tend to be

separated by encroaching white feathers midventrally.

The side patches are much more reduced, and are con-

fined to a small area at the top of each thigh. The re-

mainder of the patches has become inactive, so that a

completely white belly and back result. A very common

occurrence is the white rump patch due to the restriction

of the side patches, so that a break occurs between them

and the tail patches. The rump patches in birds are

situated far back, as in mammals, and pigment the tail

coverts and the rectrices as in Figs. 49-51. The bilater-

ality of the two patches is often indicated in pigeons by

the occurrence of a few pure white rectrices in the center

of the tail. Other birds show pure white feathers at

either side of the tail, with a tendency to bilateral sym-

metry, a most important fact, since it indicates restriction

at the outer extremes of these centers. In the restriction

of pigment formation, the rectrices are the first to be-

come white, as one would expect, since they are situated

at the extremity of the body and farthest from the center

of the patch. In Fig. 52 these centers are seen to be at

the base of the tail above, and include the upper tail

coverts. They are still joined medially, but that of the

left side is more extensive than the patch on the right side.

The approximate boundaries of the several pigment

patches are indicated in Fig. 53 by dotted lines; 1 is the

crown patch, 2 the ear patch, 3 the neck patch, 4 the

shoulder patch, 5 the side, and 6 the rump patch, as they

appear in a dorsal view. Ventrally the neck patch runs

forward to the symphysis of the mandibles.

In a flock of domesticated mallard ducks which I

studied, the same patches were found indicated, and

No. 572] ‘PATTERN DEVELOPMENT 483

some of the details of these are shown in Figs. 54-56.

In the male wild mallard there is no white in the pattern

of the head and neck except a white ring at the base of

the neck. In one of the domesticated breed, shown in

Fig. 54, the crown patch was very beautifully marked

off, as in the pigeon (Fig. 49), by a white band from eye

to eye passing about the occiput. This duck was further

interesting in showing the median division of the two

neck patches, as a narrow white line running down the

Jsf 5e Fé

Fies, 54-57. DIAGRAMS SHOWING PIGMENTATION IN DOMESTICATED MALLARD

DUCKS AND IN THE (WILD) LABRADOR DUCK (57).

back of the neck medially, from the occipital stripe. An-

other duck shown in Fig. 55 had lost the neck patches

entirely, but showed the same occipital stripe bounding

the crown patch posteriorly, and the ear patches dorsally.

The ear patches still adjoin the crown patch anteriorly.

In Fig. 56 is represented another of these ducks in which

both ear patches are distinct and separate on either side

of the head. The crown patch appears as two narrow

lines of pigmented feathers which are not quite in con-

tact posteriorly. I have not obtained a satisfactory

explanation for the apparent tendency of this patch to

484 THE AMERICAN NATURALIST [Vou. XLVIII

divide medially. Probably for some reason the forma-

tion of the pigment is more intense at the sides of the

crown than in the center where the nerve and blood

supply is less. In the pterylosis of this area the develop-

ment of feathers is seen to be greater at the sides also.

The neck patches and the side patches are absent en-

tirely, but the shoulder patches are both present, in Fig.

56, that of the right side covering the scapulars and

middle of the upper back, that of the left side including

a few only of the scapulars.

The tail patches are both present, and separate from

each other, as shown by the median white rectrices.

In this same flock of mallards was a female which had

a white ring at the base of*the neck in the same situation

as the white ring which in the male is a part of the per-

manent pattern. It was not quite complete dorsally,

however, in this female, and was somewhat broader than

regularly in the male. Nevertheless, it is apparent that

this white collar in the male is merely a primary break

between neck and shoulder patches that has become

developed as a part of the normal pattern.

- Stone (1912, p. 318) in his paper on the phylogenetic

value of color characters in birds, hints at the existence

of these patches. He says, in part:

In matters of pattern there seems to be a deeper problem involved,

i. e., the determination of the cause governing the appearance of a dif-

ferently colored patch on corresponding parts of the plumage of birds

belonging to wholly different groups . . . or the presence of a mystacial

stripe, a superciliary stripe, a light rump patch. .. . In fact if a bird

exhibits a bright or contrasting patch of color, it is, in the vast major-

ity of cases, found on one of several definite portions of the plumage,

as the crown, the throat, the bend of the wing, the rump, ete.

These contrasting areas are due to the development of

one or more of the primary patches, or of breaks between

them, or again paler areas, as at the bend of the wing or

on the rump, indicate often a lessening of pigment inten-

sity at a distance from the respective primary centers.

(To be concluded.)

NOTES ON THE MEADOW JUMPING MOUSE

(ZAPUS HUDSONIUS) ESPECIALLY RE-

GARDING HIBERNATION

H. L. BABCOCK, M.D.

DEDHAM, Mass.

THE jumping mouse is the only one of the wild mice of

this region (Massachusetts) which exhibits the habit of

regular hibernation. Regarding this habit there are a

number of references in the literature on the subject.

Barton! was one of the first to refer to the fact that this

mouse became dormant in winter. He says, in describing

the actions of one he had in captivity:

On or about the 22d of November it passed into the torpid state. It

is curious to observe that at the time it became torpid the weather was

unusually mild for the season of the year, and moreover the animal was

kept in a warm room, in which there was a large fire the greater part

of the day and night . It was frequently most active while the

weather was extremely ld in December.

This was in Philadelphia, Pa.

: Audubon and Bachman? regret that they live in a region

where the species does not exist and can not speak from

personal observation on the subject.

Godman,: Thompson‘ and Kennicott® speak of its habit

of hrboriation:

Tenney® gives an account of a specimen of this species

taken alive on January 18, 1872, near Vincennes, Ind. It

was dormant, coiled up tightly, ‘‘the nose being placed

upon the belly, and the long tail coiled around the ball-like

1‘*Some Account of an American Species of Dipus or Jerboa,’’ by Ben-

jamin Smith Barton, M.D., Translations of the Am. Philosophical Society,

Vol. IV, No. XII, 1799.

2 Viviparous Quadrupeds of No. Anais: ’ Vol. II, 1851, p. 255.

8 Godman, ‘‘Am, Nat. Hist.,’’ Vol. I, 1

4 Rev. Zadoe Thompson, Ni at. and Civil Hk of Vermont,’’ 1842.

5 Kennicott, Patent Office Report for 1857.

6 Tenney, ‘‘ Hibernation of the Jumping Mouse,’’ AM. NATURALIST, June,

1872, Vol. VI, No. 6, pp. 330-332.

485

486 THE AMERICAN NATURALIST [Vou. XLVIII

form which the animal had assumed.’’ It was taken from

a nest about two feet below the surface, made of bits of

grass. The mouse showed no signs of life at first, but on

being held in his hand, soon became feebly active, and on

being placed in a warm room, came out of its dormant con-

dition entirely. It again became dormant that night, but

was aroused twice again by the application of heat, within

the next few weeks, in spite of very cold weather.

Merriam’ tells of taking an active male at Easthamp-

ton, Mass., on February 11, 1872, and states that during

the mild winter of 1881-82, in Lewis County, Northern

New York, he saw jumping mice active several times.

Seton® speaks of finding a Zapus Hudsonius on Sep-

tember 27, 1888, at Carberry, Manitoba, in a nest of leaves

under the roots of a stump, nearly torpid. He says:

In the country near Carberry, I never saw it active after September

Stone and Cram? believe that this mouse passes six

months or more of every year hibernating underground.

They speak of seeing a family of them turned up by a

plough in May and exhibiting not the slightest symptom

of life, on being handled or breathed upon.

Burroughs!’ tells of a female jumping mouse in cap-

tivity that began hibernating early in November and con-

tinued until May, with several intervals of activity, espe-

cially after warm weather came on.

Preble! says:

Hibernation varies with the locality, but usually begins about the time

of the first heavy frosts and lasts until Spring. The fall pelage is

usually assumed and the animals become exceedingly fat before entering

winter quarters. Although they often lay up stores of food in nests or

urrows during summer, it is not known that they use this food during

winter. The animals are generally found singly (sometimes in pairs)

in nests at a depth varying from a few inches to two or three feet below

the surface. Hibernation sometimes takes place above ground.

TC. H. Merrian, M.D., ‘‘Mammals of the Adirondack ai 7? 1884.

8 E. T. Seton, “Life histories of Northern Animals,’’ Vol. I.

9 Stone and Cram, ‘‘ American Animals,’’ p. 103- 104.

10 John Burroughs, ‘‘ Squirrels and Other Fur Bearers,’’ pp. 121-12

11 E, A. Preble, ‘‘ Revision of the Pa Mice of the Genus ue

U. S. Dept. Agr. N. A. Fauna Series, No. 15, 1899.

No.572] NOTES ON MEADOW JUMPING MOUSE 487

On June 25, 1912, a female Zapus Hudsonius was taken

alive, by the writer, on the edge of a small pond in eastern

Massachusetts. It was placed in a small wire cage, and

after a few frenzied efforts to escape, became quite tame.

On July 5 it gave birth to five young, blind and hairless;

but when the family was transferred to a larger cage, the

mouse deserted the young and they soon died. One dis-

appeared mysteriously, and may have been eaten by the

mother. The young measured at birth: total length 33

mm.; tail 9 mm.; hind foot 4 mm.

Throughout the summer the mouse ate chiefly rolled

oats and shredded wheat, and was also very fond of straw-

berries and blueberries. It refused most of the common

fruits and vegetables.

It was almost wholly nocturnal in its activity, sithough

when disturbed during the day it would immediately begin

to eat and remain active for half an hour or more.

Toward the latter part of the summer, it seemed to grow

quite fat. Rhoads!” says in this connection:

When going into winter quarters they are exceedingly fat, as I can

testify from experience in removing this tenacious yellow blanket from

the skins of them. This fat is their fuel. By spring it is nearly gone.

During the latter part of August there were several

very cool nights (49° F. minimum) and on the night of

August 28 it did not come out. This fact was apparent

from the clean drinking dish, which was placed in such a

position that the mouse could not approach without scat-

tering saw-dust in it. The absence was repeated on Au-

gust 30, and September 1. Throughout September its

actions were irregular. Every night until the 21st, with

the exception of the 12th and 17th, it was active, but on

the 22d disappeared for four nights. It was then active

for two more nights (26 and 27) and following that,

inactive for six (September 28 to October 3). From Oc-

tober 4 to 28 it was out every night, although not as vigor-

ous as formerly, neither did it eat as much. When ap-

proached it seemed to pay no heed, as if in a sort of

stupor.

12 S. N. Rhoads, ‘The Mammals of Pennsylvania and New Jersey.’’

488 THE AMERICAN NATURALIST [Vou XLVIII

OrriciA. OBSERVATIONS

From US. Wearner Bureau, Gos row Sranion.

CHART vlad MINIMUM. NIGHTLY

Aveust SEPTEM

Aaa 12 1431F 1516 IT IF 19-10 H 222324 3S 3 2) 38 3430 Hf 23¢F6 759, Utd 7 70.2 y 729 293P 72347

70"

éo’

ts?

vo"

30%

There was no evidence of any attempt at storing away a

supply of food, although there was ample opportunity.

This habit of storing food is mentioned by Hornaday**

who says:

In the autumn it stores in the ground quantities of food for winter

use, but despite this fact, under certain conditions, it becomes so Bas

oughly dormant in winter that it seems to be quite acinar

According to Seton,® `

It is quite ready to respond at any time to any spell of unusually fine,

unseasonable weather, even in the depths of winter, and it is probably

for these arousing times, as much as for the spring time famine, that it

lays up its abundant stores of food.

Preble"! also mentions this habit, but Shufeldt'* denies

it. He says, in speaking of the deer mouse (Peromyscus

Leucopus) :

Is it to meet the requirements of his condition that this mouse lays up

a goodly stock of food during the autumn? one the Zapus does

not do.

Following the period of activity through October, the

mouse was inactive on the four nights of October 28, 29,

30 and 31, and reappeared for the last time on the night

of November 1, after which it retired for the winter. The

cage was placed by an open window of an empty box stall

in a stable where the temperature was practically that of

out doors. The mouse built its nest in the side of a large

sod placed in one corner of the cage.

In spite of a very mild winter, the lowest official tem-

13 Hornaday, ‘‘The American Natural History.’’

14R. W. Shufeldt, M.D., ‘‘Chapters on the Natural History of the United

States.’’

No. 572] NOTES ON MEADOW JUMPING MOUSE 489

TEMPERATURES, AVGUST — NOVEMGER, 1912,

OCTOBER ; NOVEMBER

STII 1M RISHSKIZI GG Wr PALAGIAN 23 ESE JE FION IMIS 17M IPOH > ia

AWN BY

f H.L. BABCOCK:

perature for this section being only 3° F. (February 10,

1913,) the mouse did not survive the cold weather, and

was found dead, when the cage was opened on June 17,

1913.

The nest was found to be located in the extreme end of

the sod, only 14 inches from the top and about 1 inch from

the edge. It was roughly oval in shape, being hollowed

out of the loam and lined with a few blades of grass. It

_ measured roughly 13 inches by 14 inches and was just

large enough to contain the mouse when curled up into a

ball. The opening was on the side. Death was probably

caused from exposure to continued cold owing to the un-

protected location of the nest.

The poor judgment shown in not building the nest

securely in the middle of the large sod, and other similar

instances of poor management, have led the writer to

believe that the intelligence (if that term may be used) of

the Zapus Hudsonius is of comparatively low grade, much

lower, for instance, than that of the deer mouse (Peromys-

cus Leucopus).

The accompanying chart, which is a record of minimal

nightly temperatures, according to the official observa-

tions of the U. S. Weather Bureau for this section (Bos-

ton), shows the activity of the mouse in relation to the

temperature during August, September, October and No-

vember, and brings out some rather interesting facts. For

example, on October 15, 16 and 17, with the minimum

nightly temperature 42°, 36° and 42°, respectively, the

mouse was active, while on September 22, 23, 24 and 25,

490 THE AMERICAN NATURALIST [Vou. XLVIII

with the minimum nightly temperature of 48°, 49°, 56°

and 53° F., respectively, a much warmer series of nights,

it remained inactive, And again, after November 1, dur-

ing a warm spell in which the minimum nightly tempera-

ture for November 7 was 64° F., the mouse did not appear.

A study of this chart suggests the question as to how

much the temperature has to do with this habit of hiber-

nation.

It is a somewhat general belief that temperature regu-

lates the degree of torpidity.

Barton! maintains that

the torpid state of animals is altogether an accidental circumstance and

by no means constitutes a specific character. The same species becomes

torpid in one country and not in another. Nay, different individuals

of the same species become torpid or continue awake in the same neigh-

borhood or even on the same farm.

Seton® believes that

while torpor is more or less controlled by temperature, the habit of tor-

pidity, like the changing pelage of the white-hare, is so deeply ingrained

constitutionally that there is a strong tendency to torpify at a given

time without regard to the original cause.

It is evident from this chart that torpidity develops

gradually, at first for only one night at a time. Whether

this process is explained by a cerebral anemia, a slow

toxemia of the brain centers or some other of the theories

regarding sleep, it seems to require about two months in

which to become sufficiently developed to control com-

pletely voluntary body functions. During that interval

the animal occasionally awakens, probably from hunger

and habit as much as from any effect in change of tem-

perature.

After torpidity is thoroughly established, changes of

temperature may be important external factors, as has

been demonstrated on numerous occasions in producing a

temporary activity by the application of heat. It is safe

to say, however, that the temperatire is not the only ele-

ment which influences the length of the period of hiberna-

tion.

SHORTER ARTICLES AND DISCUSSION

STUDIES ON INBREEDING—IV

~ ON A GENERAL FORMULA FOR THE CONSTITUTION OF THE NTH

GENERATION OF A MENDELIAN POPULATION IN WHICH

ALL MATINGS ARE OF BROTHER X SISTER ?*

_ I. In a former paper in this series’ the constitution of a Men-

delian population in which all mating was of the brother X

sister type was worked out empirically. The results there pre-

sented may be put in the form of a general formula, by means

of which the constitution of any generation may be written down

from a knowledge of the preceding generation; that is from a

knowledge of the n— 1th generation the nth generation may be

at once written down.

II. This general formula may be developed as follows. A

single character pair will be considered, A denoting the dominant

character and a the recessive. Equal fertility for all matings

is assumed, the number of individuals per family being taken as

2s, of which s are males and s are females. One family will then

make s matings and produce s families in the next generation.

Each mating is, by hypothesis, of a brother with his sister.

Starting as before with a pair from a population in which all

individuals are of constitution Aa there will be in the next

generation one family of the AA + Aa+aA- aa type. In all

succeeding generations there will be six types of families, viz.:

(1) AA families.

(2) AA + Aa families.

(3) Aa families.

(4) Aa + 2Aa-+ aa families.

(5) Aa\+ aa families

(6) aa families.

1 Papers from the ee Laboratory of the Maine Agricultural Ex-

periment Station, No.

It seems dieiis ny pe as a general series of ‘í Studies on In-

ts o hi

sis of the Problem of Inbreeding,’’ AMER. Nat., Vol. XLVII, pp. 577-

615, 1913. II. ‘‘Tables for Calculating Coefficients of» Inbreeding,’’ Ann.

Rept. Me. Agr. Expt. Sta. for 1913, pp. 191-202. III. ‘‘On the Results

of Inbreeding a Mendelian Population: A Correction and oa of

Previous Conclusions, ”? AMER. Nart., Vol. XLVIII, pp. 57-62,

2 AMER. Nat., Vol. XVLIII, pp. 57-62, 1914.

491

492 THE AMERICAN NATURALIST [Vou. XLVIII

The proportionate number of each of these types of families |

will change in successive generations according to the following

system.

Let on denote the number of AA families in the n— ith

generation, and

Pn. denote the number of AA-+ Aa families in the

n— 1th generation, and :

Qn, denote the number of Aa families in the n— ith

generation, and

Yn, denote the number of AA and 2Aa and aa families

in the n— 1th generation, and

Un_, denote the number of Aa- aa families, and

Un_, denote the number of aa families.

It will be possible to write down u and v in any case without

calculation because of the symmetrical relations of a Mendelian

population, since always under normal conditions such as are

assumed in the general treatment, we have

Un. = Pn-1

Un- TED On-1;

Un = Pn,

Un = On

It is necessary, therefore, to consider only the coefficients for

the first four types of family. In the nth generation the consti-

tution of the population in respect of families (not individuals)

will be as follows:

Families in nth generation

` =S (0n + 1/4pn +1/16r,_,) AA families

+ s(1/2pn_,+ 1/4rn) AA + Aa families

+ 8(1/8rn_,) Aa families

a 8(1/2pn_. + Qn-1 + 1/41.) AA

+ 2Aa-+ aa families

+ s(un)Aa-+t aa families

j+ s(n) aa families.

Or, taking coefficients alone we have

= Ona +1/4pn_,+1/16rn_,,

Pn=1/2pn_, + 1/4tn_1,

Gn aoe 1/ Brn,

n= 1/2pn- + Ina +1/ ea

Un = 1/2tty:, + 1/40n4 =

Un = Vna + 1/4tn + Wik. A

(i)

No.572] SHORTER ARTICLES AND DISCUSSION 493

III. Let us see how this formula works out in a concrete case.

Assume the same conditions of fertility as in the former paper,

that is, put 2s = 32, or s=16. Start with a single AA + 2Aa

+ aa family.

Then

Cig, SD,

Pn-1 rs

Qn-1 = 0,

fas

Then in the next generation we shall have

16{0 + 1/4(0) + 1/16(1)}=1AA family

+ 16{1/2(0) +1/4(1)}—4(AA+ Aa) families

+ 16{1/8(1) } =2Aa families

+ 16{1/2(0) +0+1/4(1)}= 4(AA +2Aa-+aa) families

+ 4(Aa-+ aa) families

+ laa family.

This is the fact.

In the next generation we shall have

16{1 + 1 +1/16(4)} 364A families

+ 16{1/2(4) + 1/4(4) }— 48 (AA + Aa) families

+ 16{1/8 (4) } — 84a families

+ 16{1/2(4) + 2 + 1/4(4) }=—= 80 (4A + 24a + aa) families

+ 48(Aa + aa) families

- +36 (aa) families. —

This is the fact.

In the next generation we shall have

16{36 + 1/4(48) + 1/16 (80) } —16 X 53 — 8484A families

+ 16{1/2(48) + 1/4(80) } 16 X 44=704(AA + Aa)

familie

+ 16{1/8(80) } —160Aa families

+ 16{1/2(48) + 8 + 1/4(80) } 16 X 52=—832(AA + 24a

+ aa) families

+ 704(Aa-+ aa) families

+ 848aa families,

Succeeding oe follow the same law and need not be

worked out in

IV. So far ae dietaniGn has confined itself to families, as this

must be the basic unit in the theory of any form of inbreeding.

Turning to individuals we have the following simple relations to

pass to individuals.

In the nth generation the number of

494 THE AMERICAN NATURALIST — [Vou. XLVIII

AA (or aa) individuals = 2s (0n) H- s(pn) + 1/28 (ra),

Aa (or aA) individuals = 2s (qn) + s(1/2pn) +1/2s(1n).

The first of the above expressions multiplied by 2 gives the

total homozygotes, and the second multiplied by 2 gives the total

heterozygotes.

RAYMOND PEARL

PARALLEL MUTATIONS IN ŒNOTHERA BIENNIS L.

In the summer of 1912 I cultivated pure strains of O. biennis

L. and of the O. biennis cruciata de Vr. of our Dutch dunes, as

well as of their hybrids, made with the purpose of studying the

behavior of the cruciata-character in crosses. In one of these

cultures I unexpectedly obtained two mutants, which because of

their similarity to corresponding variants derived from O.

Lamarckiana have been called O. biennis nanella and O. biennis

semi-gigas. The first mutant, O. biennis nanella, occurred in the

second generation of the cross O. biennis X O. biennis cruciata

and. differed from O. biennis in all those points which separate

O. Lamarckiana nanella from O. Lamarckiana. The other vari-

ant, O. biennis semi-gigas, appeared in the second generation of

the reciprocal cross, O. biennis cruciata X O. biennis, suggesting

immediately by its much more vigorous habit and especially by

the larger size of its buds and flowers the differences between O.

Lamarckiana and O. gigas. A count of its diploid number of

chromosomes proved it to deserve the name semi-gigas, 21

chromosomes being shown by nuclear plate-stages in the meris-

tematic tissue of young buds. From these facts, showing that

O. biennis is in a mutating condition, I drew the conclusion that

the phenomenon of mutation in the genus @nothera is older than

the species O. Lamarckiana—O. biennis generally being consid-

ered to be an older species than O. Lamarckiana—and further,

that the mutations in this group can not be the result of hybridi-

zation, as was assumed by some authors at that time—nobody

doubting of the purity and constancy of O. biennis. As a mat-

ter of fact, both of my mutants have been derived from crosses

between O. biennis and O. biennis cruciata. But I laid special

emphasis on the fact that O. biennis and O. biennis cruciata have

exactly the same germinal constitution except for the factors that

determine the shape of the petals, O. biennis cruciata being prob-

No.572] SHORTER ARTICLES AND DISCUSSION 495

ably a mutant from O. biennis itself. Therefore, hybrids between

these two forms can be looked upon as pure O. biennis except for

floral characters.

With this conception Bradley Moore Davis does not agree.’

He thinks that the O. biennis and O. biennis cruciata of our dunes

are not so closely related types, that a cross between them can

be treated ‘‘as though it were the combination of forms within

the same species which have similar germinal constitutions.’’ He

says:

It should be made clear that the form “ O. biennis cruciata” is recog-

nized in the more recent taxonomic treatments as a true species sharply

distinguished from types of biennis by its floral characters. ... O

cruciata is found wild in certain regions of New ean and New

York and is consequently a native American specie . Whatever

may have been the origin of O. cruciata or its Haat HisHionehip to

O. biennis, a cross between these types must certainly be regarded as a

cross between two very distinct evolutionary lines and its product a

hybrid in which marked modifications of germinal constitution are to

be expected.

From Davis’s point of view I ‘‘really made a cross between

two rather closely related species’’ and obtained in the second

generation ‘‘two marked variants due to some germinal modifica-

tions as the result of the cross.’’ In so far as my observations bear

upon the problem of mutation Davis’s interpretation is exactly

the reverse of mine. To him they further illustrate the same

phenomenon which he is obtaining through his ‘‘hybrids of

biennis and grandiflora, namely, that behavior by which these

hybrids in the F, generation throw off variants that in taxonomic

practise would be considered new species readily Gieree piano

from the parents of the cross and from the F, hybrid.’’

It will be shown in the following lines that the objections made

by Davis are not sufficiently justified. My argument consists of

two points,

In the first place, Davis is mistaken as to the nature of the O.

biennis cruciata de Vr. of our dunes. This strain is in reality

quite another type than the different forms of the American 0.

cruciata Nutt., called by some authors O. biennis cruciata. With

this species it has in common only the character of the narrow

1 Bradley Moore Davis, ‘‘ Mutation in @nothera biennis L.?’? THE AMERI-

CAN Naturauist, Vol. XLVII, 1913, pp. 116-121; ‘‘Genetical Studies on

Enothera,’? IV, THE AMERICAN Naturauist, Vol. XLVII, 1913, pp.

546-571.

496 THE AMERICAN NATURALIST [Vou. XLVIII

petals, all other features of the stem, foliage, flowerspikes and

fruits being exactly those of the Dutch O. biennis L. It must

certainly be looked upon as a mutation from the O. biennis L. of

our sand dunes. Until now it has only been found a couple of

times in single individuals in the midst of the ordinary O. biennis,

the first time in 1900 by Dr. Ernst de Vries in the dunes in the

neighborhood of Santpoort, Holland, in one individual—and

from this one specimen all the subsequent generations of O. bien-

ms cruciata in the cultures grown by de Vries and by myself have

been derived. Besides this, our O. biennis and O. biennis cruciata

are so similar to one another except for floral structure that

plants of both types can not be separated before the flowers open.

Therefore we have the right to assume that the crossing of these

two forms is concerned alone with the floral characters and that

with respect to all other characters parents as well as hybrids are

mere biennis. Therefore the two variants which arose in my cul-

tures from crosses between O. biennis and O. biennis cruciata

obviously prove the faculty of mutation in O. biennis.

In the second place I have found now that it is not necessary

to cross O. biennis with O. biennis cruciata in order to obtain the

above named mutants, as Davis seems to believe. Already in his

new book Professor de Vries figures a dwarf derived from O.

biennis cruciata grown in pure line. Shortly afterwards I myself

obtained six mutants from the O. biennis of our sand dunes grown

also in pure line. A few details about these cultures of last year

may be given here. In all they counted 920 individuals, 430 of

which belonged to the third and 490 to the fourth generation of

a pure line, the point of departure for which had been one in-

dividual brought into the experimental garden in the rosette

stage from the dunes near Wyk aan Zee in the beginning of 1905

and self-fertilized in the same year. The six mutants which ap-

peared in these pure cultures of O. biennis were the following.

First a dwarf, then a biennis semi-gigas having 21 chromosomes

and finally four individuals of the O. biennis sulfurea, a pale-

flowered form of O. biennis, which had been found already several

times in our dunes in the midst of the ordinary biennis, but was

not with certainty known to be a mutant from the latter form

until now. The two first named mutants and one sulfurea ap-

peared in the third generation of our pure line, the nanella and

the semi-gigas coming from the same mother. The three remain-

ing sulfwrea-individuals appeared in the fourth generation, all

No. 572] SHORTER ARTICLES AND DISCUSSION 497

descending from the same motherplant. Of these mutants the

nanella and semi-gigas are especially valuable because similar

forms have been produced by O. Lamarckiana. It will be seen that

the biennis-dwarfs seem to be somewhat rarer than the dwarfs of

Lamarckiana. Whilst for the latter the mutation coefficient is about

1 per cent. our O. biennis nanella appeared as the only dwarf among

920 individuals. The above cited O. biennis cruciata nanella

was the only dwarfed individual in a culture of 500. And the

dwarf which I got in 1911 was the only one among about 600

plants. In this connection I wish to recall the conclusion

reached by de Vries that in O. Lamarckiana the pangen for tall

stature must be assumed to be present in the labile condition on

both sides, in O. biennis, however, only in the male sexual type,

whilst in the female sexual type active alta-pangens have to be

supposed. The way from biennis to biennis nanella might there-

fore possibly be somewhat longer than the one from Lamarckiana

to Lamarckiana nanella. The biennis semi-gigas which appeared

in the last summer corresponded in all points exactly with the

mutant of 1911. Moreover a count of the chromosomes, as shown

‘by nuclear plate-stages in the meristematic tissue of young buds,

determined them to be 21 in number. Even as the specimen of

this type, that appeared in 1911, and as the semi-gigas mutants

produced by O. Lamarckiana, the plant of last year proved to be

almost absolutely sterile.

In his second above-mentioned paper Davis says about i 0.

biennis of our dunes: ‘‘No species of @nothera is perhaps so free

from suspicion as to its gametic purity. If Stomps can obtain

mutations from tested material of the Dutch biennis grown in

pure lines he will have the basis of a strong argument,

Fortunately the experiment asked for by Davis, has been had

in the same year as his criticism. The Dutch biennis L., culti-

vated in pure line, has produced a dwarf, a semi-gigas acd some

sulfurea-individuals, proving its mutability beyond all doubt. I

therefore trust that the conclusions arrived at in my first paper,

concerning this mutability and its consequences, may now be ac-

cepted as thoroughly valid.

THEO. J. Stomps

AMSTERDAM, HOLLAND

498 THE AMERICAN NATURALIST (Vou. XLVIII

IN a recent review’ of Stomps’s studies on @nothera biennis

L.? from the sand dunes of Holland I protested against his desig-

nating as mutants a nanella type and a semi-gigas type which

were obtained in the second generation of crosses between Eno-

thera biennis Linneus and its variety O. biennis cruciata de

Vries. The criticism was presented on the general ground that

however close the possible relationships between the two parent

forms, they nevertheless constituted lines so far apart as to

render unsafe a conclusion that marked variants obtained from

their crossing are mutants in the sense of de Vries and Stomps.

Such variants, it seemed to me, might have been the result of

_ hybridism between two lines sufficiently divergent to upset the

similarity of germinal constitution shown in their vegetative

morphology, for the species biennis and its variety cruciata are

said to differ only in their flower structure.

In that review I incorrectly associated O. biennis cruciata de

Vries with O. cruciata Nutt., an American species entirely dis-

tinct from the variety cruciata of de Vries, which has been found

only once (in the year 1900) on the sand dunes of Holland

among plants of O. biennis. I greatly regret my confusion of

these two types, since I was led in my criticism to regard Stomps’s

crosses between biennis and biennis cruciata as though they were

crosses between two distinct although possibly closely related

species. In this I was clearly mistaken, since all of the evidence

short of experimental proof, which Stomps may yet obtain, indi-

cates that biennis cruciata de Vries is a variety of biennis L. and

arose as a mutation on the sand dunes of Holland. The crosses

of Stomps are, therefore, to be regarded as between a species and

its mutant variety. I trust that the mutationists will accept this

acknowledgment of an error.

There is, I believe, a body of naturalists for whom the value

of evidence for mutation rests fundamentally upon the unques-

tioned purity of the parent stock, and to them any cross, no matter

how close, is open to criticism. Stomps has justified his first con-

clusions by obtaining in later studies the same mutants biennis

nanella and biennis semi-gigas from lines of the pure species O.

biennis Linneus. Had he waited for these later results before

1 Davis, B. M., ‘‘Mutations in @nothera biennis L.?’’ AMERICAN NAT-

URALIST, Vol. XLVII, p. 116, 1913.

2 Stomps, T. J., ‘‘ Mutation bei @nothera biennis L.,’’ Biol. Centralb., Vol.

XXXII, p. 521, 1912.

No. 572] SHORTER ARTICLES AND DISCUSSION 499

publishing on the first there could have been no objections to his

main contention that O. biennis from the sand dunes of Holland

is capable of giving rise to true mutants.

Stomps is continuing his studies on this same Dutch biennis

with the view of determining its possible powers of mutation, and

it is a pleasure to review his second paper? which presents some

extremely interesting data, a paper in which no important criti-

cism can be based on the source and character of the material em-

ployed. No wild species of evening primrose has been so long

under experimental and field observation or is better known to

the workers with cnotheras than this plant. The species has

proved uniform in culture to a remarkable degree and it would

be difficult to find a type of @nothera so free from suspicion of

gametic purity. The species appears to have been in Holland

since pre-Linnwan days and is therefore very old. As material

for experimental studies on mutation the Dutch biennis seems to

the writer the best of all the œnotheras so far brought into the

experimental garden.

The starting point of Stomps’s cultures of @nothera biennis

was a plant transplanted from the sand dunes in 1905. From

seed of this plant, self-pollinated, a second generation was grown

in 1910, three selfed plants of which gave the seed for a third

generation of 430 individuals, and a fourth generation of 490

plants was grown from two selfed plants of the third generation.

Thus in all 930 individuals were observed in the third and fourth

generations from the plant that gave rise to these pure lines.

It is true that these lines have not been under selection for many

generations, but, considering the stability of the species and its

habit of close pollination, it is very improbable that the source of

the cultures should have been a plant not representative of the

type. Furthermore, Stomps presumably will continue indefi-

nitely the lines now established and thus determine through

later generations whether their mutating habits remain constant.

Among the 430 plants of the third generation there appeared

1 biennis nanella, 1 biennis semi-gigas and 1 individual of biennis

sulfurea; the first two came from the same mother plant. Among

the 490 plants of the fourth generation appeared 3 individuals of

biennis sulfurea, all from the same selfed mother. The variety

sulfurea differs from the species biennis in having flowers of a

3 Stomps, T. J., ‘‘Parallele Mutationen bei Gnothera biennis L.,’’ Ber.

deut. bot. Gesell., Vol. XXXII, p. 179, 1914.

500 THE AMERICAN NATURALIST [Vou. XLVIII

lighter yellow, and is reported by de Vries to be not uncommon

in the wild state mixed with the species proper. Sulfurea has

been held systematically to be a variety of biennis but this is the

first time that it has appeared in the experimental garden as a

derivative of that species. Thus out of a total of 920 plants

there were 4 individuals of the color variety sulfwrea, 1 nanella

and 1 semi-gigas, in all 6 mutants, a showing that may well

gratify Stomps.

The mutant biennis nanella differed from typical biennis in

much the same way that Lamarckiana nanella differs from

Lamarckiana and like the latter dwarf showed evidence of a

bacterial infection. Certain selfed flowers set no seed because

of diseased stigmas. Other flowers pollinated from pure biennis

set good fruit. The ratio of the appearance of biennis nanella

is much lower than the mutation coefficient of one per cent. which

de Vries has reported for Lamarckiana nanella. It should also

be remembered that de Vries* obtained a cruciata nanella in a

culture of 500 plants from O. biennis cruciata.

“The mutant biennis semi-gigas in comparison with typical

biennis showed a stronger habit, broader leaves, thicker buds,

larger flowers, supernumerary stigma lobes, and the presence

of 4-cornered pollen grains. Counts of the chromosomes in

meristematic tissue determined the number to be 21. Therefore

in this plant, as in the biennis semi-gigas obtained by Stomps

from the cross cruciata X biennis, there is clear cytological evi-

dence that one of the gametes which formed the zygote contained

14 chromosomes, i. e., double the number characteristic of the

gametes of Gnothera. This is another case of triploid mutants

in @nothera to be added to the list of Stomps and Miss Lutz.

The plant was self sterile, but set fruit when pollinated by

biennis, although the yield of seed was very poor.

Stomps is justified in calling attention to the agreement of his

second biennis nanella with the plant derived in 1911 from the

cross biennis X cruciata, and of the agreement of his second

biennis semt-gigas with the plant from the cross cruciata X

biennis. It is to be hoped that he will next obtain the cruciata

variety as a direct mutant from the Dutch biennis and thus

establish its relationship and origin beyond all possible doubt.

_ Stomps has before him the opportunity of making through the

study of Gnothera biennis very important contributions to our

4 See ‘‘Gruppenweise Artbildung,’’ p. 299 and Fig. 108.

No. 572] SHORTER ARTICLES AND DISCUSSION 501

knowledge of the frequency of mutations and their importance

in organic evolution. That retrogressive mutations take place is

not likely to be seriously doubted by any one who has followed

the experimental work of recent years both botanieal and

zoological. The loss of characters through germinal modification,

even in what seem to be ‘‘pure lines,’’ appears to be not

uncommon.

Most of all is desired information on the possibilities, fre-

queney and character of progressive mutations. Can the muta-

tion theory satisfactorily explain progressive advances in organic

evolution or must amphimixis chiefly carry that responsibility ?

-~ Mutants of the tetraploid gigas-like type would appear to be pro-

gressive, and we can see the reason in their doubled chromosome

count which gives larger nuclei, larger cells and modified tissues.

Gigas-like forms are however very rare and in O. Lamarckiana

gigas the fertility is relatively low. More common are the trip-

loid semi-gigas forms, but these seem to be sterile or almost

sterile when selfed, and the work of Geerts indicates that the

triploid number in @nothera returns to the normal through

the elimination of supernumerary chromosomes. Very inter-

esting is the recent paper of Gates and Thomas*® which offers

evidence that lata-like characters are associated with the pres-

ence of a single additional chromosome.

And what of the series of forms which differ from the

@nothera parent types with as yet no evidence of peculiarities in

their chromosome count, brevistylis, levifolia, rubrinervis, obo-

vata, scintillans, ete. Will forms similar to these and perhaps

others in addition be represented in a series of derivatives from

@nothera biennis? The mutants biennis nanella and biennis

sulfurea belong to this group and have already been obtained

by Stomps. One may almost envy him his opportunity for an

intensive study of this species.

BRADLEY Moore Davis

UNIVERSITY OF PENNSYLVANIA,

June, 1914

5 Gates, R. R., and Thomas, N., ‘‘A Cytological Study of Gnothera mut.

lata and @. sinks semilata in Relation to Mutation,’’ Quart. Jour. Mic. Soi.,

Vol. LIX, p.°523, 1914.

502 THE AMERICAN NATURALIST (Vor. XLVII

THE THEORETICAL DISTINCTION BETWEEN

MULTIPLE ALLELOMORPHS AND

CLOSE LINKAGE

PROFESSOR CASTLE’S difficulty in understanding the distinction

made by Mr. Dextert is owing to his unfamiliarity at first hand

with the phenomenon of linkage. The distinction between allelo-

morphs and close linkage has already been given several times

elsewhere and need not be repeated; but if Professor Castle has

failed to note it, or to see its significance, it is probable that

others may have done the same. I may be pardoned, therefore,

for attempting once more to show why, for clear thinking, it is

important to keep in mind the difference between allelomorphs

and close linkage. Furthermore, since we have here one of the

newest developments of Mendelism, it seems to me that it may be

worth while not to let Professor Castle’s criticism pass un-

challenged.

Dexter pointed out that the mode of treatment that Nabours

followed in the analysis of his results is the procedure of multiple

allelomorphism, although Nabours does not seem entirely con-

versant with the fact, but treats the results as though they were

regular phenomena. In one case, however, Nabours got an un-

conformable individual. Dexter points out that if this case is

not due to non-disjunction (a known process that will cover

such cases) it shows that here at least the factors involved are

not allelomorphs, but must be treated as though closely linked.

How could the matter be put more directly? I confess I am

at somewhat of a loss to discover why Professor Castle is con-

fused. Perhaps it is the subsequent development of Dexter’s

explanation that has troubled him. Let us again try to make the

distinction clear. `

If the factors B and E are not allelomorphic to each other then

each must have another allelomorph. This is nothing but pure

Mendelism, which no one will, I suppose, dispute. It is entirely

irrelevant whether we use small letters or none at all (as Castle

prefers) for the allelomorphs. If they are a part of the Men-

delian machinery, who cares very much what we call them?

If then we have here two pairs of allelomorphs, crossing over

may take place, as it does in other cases where two pairs of linked

1 THE AMERICAN NATURALIST, June, 1914, p. 383.

No.572] SHORTER ARTICLES AND DISCUSSION 503

genes are involved.? This is all there is to the matter. We need

not dwell, therefore, at length on Professor Castle’s statement

that here is another case of an erroneous conclusion reached in

consequence of using small letters for ‘‘ absent’’ characters,

except to remark that Dexter did not use small letters for absent

characters, and that the erroneous conclusion has been drawn by

Professor Castle himself.

T. H. MORGAN

COLUMBIA UNIVERSITY

PRoFEssorR MorGAN has called my attention to the fact that in

criticizing a single point in Mr. Dexter’s review I have given the

impression, to some at least, that I regarded Dexter’s views as

erroneous. Such was not my intention, and I wish to correct the

impression, if I may. I do not for a moment question the reality

of ‘‘unit-character’’ inheritance’ or indorse the idea of ‘‘the or-

ganism as a whole’’ as the only inheritance unit. I agree here en-

tirely with the view which I understand Dexter to hold. If Na-

bours has encountered nothing but simple allelomorphs among his

grasshoppers (which I neither assert nor deny), this by no means

proves that only simple allelomorphs exist even among said grass-

hoppers. An organism which seems to have only one variable

““gene’’? may nevertheless possess any number of other genes

which are not varying so far as we can discover, and in which con-

sequently all zygotes are homozygous and all gametes similar to

each other.

It is only in Dexter’s discussion of the significance of the ex-

ceptional ‘‘B E T” individual that I should dissent from any part

of his excellent review. Nabours’s explanation of this case, ac-

cording to Dexter, is essentially that of ‘‘non-disjunction,’’ in-

stead of which Dexter himself offers the explanation of ‘‘link-

age,’ and proposes a repetition of the experiment to decide

between them. Now I do not question for a moment the genuine-

ness of either ‘‘non-disjunction’’ or ‘‘linkage,’’ as they occur for

example in Drosophila. Through the kindness of Professor Mor-

gan I have been able to demonstrate both these phenomena re-

peatedly to classes in geneties in the course of their laboratory

work upon Drosophila. The point which I wished to make in com-

2 Crossing over would not take place if the factors in question were allelo-

morphic. If the ease is one of non-disjunction the subsequent generation

would also give a different kind of result from that of linkage. (See

Bridges, Jour. Exp. Zool., 1913.)

504 THE AMERICAN NATURALIST [Vou. XLVIII

menting on Dexter’s review (and this is the only point in which `

I dissent from his opinions) is that the repetition of the experi-

ment, provided it had the outcome suggested by Dexter, would

leave us as much in the dark as we were before concerning the

correct interpretation of the result. Very likely, however, addi-

tional facts might be observed which would give some clue, so that

I quite agree with Dexter’s suggestion that the case should receive

further study. But I can not see that at present linkage has

more in its favor as an interpretation than non-disjunction.

The ‘‘demonstration’’ which Mr. Dexter gave of his argument

by introducing duplicate ‘‘symbols’’ instead of the single set used

by Nabours, seemed to me quite superfluous and possibly to have

been a real stumbling block in the logical process. This is why I

raised the question as to the significance of the small letters. The

terminology is that of the ‘‘presence-absence’’ hypothesis, as

commonly understood; but Professor Morgan assures me that

such is not the significance which Dexter attaches to the symbols

used. It seems to me therefore that the significance attached to

the symbols is vital to the argument in the ‘‘demonstration.’

I quite agree with Professor Morgan, however, that symbols

are a matter of small consequence. Suppose we omit the ‘‘dem-

onstration’’ by means of symbols altogether. Should we then

have any reason to favor linkage as an interpretation rather than

non-disjunction? I can not see that we should have. It seems to

me quite possible that neither explanation will prové adequate.

When albino mammals are crossed with colored ones, piebalds

sometimes are obtained in later generations. So far as we know,

these result neither from ‘‘non-disjunction’’ nor from ‘‘cross-

overs.’’ Perhaps the B E I individual also is a ofaa quid.

W. E. CASTLE

NOTES AND LITERATURE

BIOMETRICS

An ImporTANT CONTRIBUTION TO STATISTICAL THEORY

ONE of Pearson’s most valuable contributions to statistical

theory is his test for goodness of fit.t It enables one, with the

aid of Elderton’s? tables, easily to determine the probability that

a given system of observed frequencies does or does not differ

significantly from a series of theoretical frequencies supposed to

graduate the observations. The significance of this criterion in

Mendelian work has recently been pointed out by Harris.*

Hitherto this criterion has found an important limitation in

the fact that, as originally developed by Pearson, it was appli-

eable only to frequency systems. It could be used to test good-

‘ness of fit only where the observations were counts of the number

of times particular classes of events occurred. But, of course,

frequency systems comprise only one kind of observational data

to which one has occasion to fit curves. Much more often there

is need for a criterion of goodness of fit where the observations

are of the nature of true ordinates, rather than frequencies.

Such cases include all data of the sort where a mean y is deter-

-mined for each x, as in a growth curve; or in the regression

observed in a correlation table, where for each successive value

of one of the variables the mean value of the correlated variable

is caleulated. There has been no method of testing the good-

ness of fit for such curves. From a visual inspection of the

. plotted regression line one has been compelled to form his judg-

ment as to whether it was or was not a good fit.

Recently a Russian statistician, E. Slutsky,* has extended

1 Pearson, K., ‘‘On the Criterion that a Given System of Deviation from

the Probable in the Case of a Correlated System of Variables is Such that

it Can be Reasonably Supposed to Have Arisen from Random Sampling,’’

Phil. Mag., 5th Series, Vol. L, pp. 157-175, 1900. :

2 Biometrika, Vol. I, pp. 155-163.

3 Harris, J. A., ‘‘A Simple Test of the Goodness of Fit of Mendelian

Ratios,’? AMER. Nat., Vol. 46, 1912, pp. 741-745, 1912.

t Slutsky, E., ‘‘On the Criterion of Goodness of Fit of the Regression ion

and on the Best Method of Fitting Them to the Data,’’ Jour. Roy. Stat. e

Vol. LXXVII, Part I (December, 1913), issued 1914, pp. 78-84.

505

506 THE AMERICAN NATURALIST (Vor. XLVIII

Pearson’s theory to cover the class of curves, formerly not

amenable to such test. The result forms an extremely valuable

extension of biometric theory.

Briefly Slutsky’s essential result may be put as follows. He

finds (the complete proof is not given in this paper) that

nN, €

aeae)

N 2 ’

Tnxpy

where x” is the quantity denoted by the same letter in Pearson’s

original work, and is the argument in Elderton’s table; Nen is

the frequency in the zp array, i. e., the number of observations

on which each observed ordinate is based; ep is the difference

between the observed and the calculated mean y for each £p

array; ando,, is the standard deviation of each x» array; îi. e.,

the standard deviation of the group of observations from which

each particular y was calculated. S, as usual, denotes summa-

tion. Knowing x’, P is read directly from Elderton’s tables.

Slutsky gives a couple of examples of the application of the

method in his paper. For illustration here I have preferred to

take an example from my own unpublished data. The observa-

tions (y,,,) in this case are the mean butter productions of

American Jersey cattle, based on seven-day tests

The theoretical points Y,,, are calculated from the equation,

y = 14.21098 + 02500 — 003822 + 3.0104 log a,

the constants of which were determined from the observations

by the method of least squares.

The test for goodness of fit is carried out in Table I. It should

be said that, following the suggestion given by Slutsky in his

paper, I have used in the o,,, column the graduated rather than

the observed values. In the present case the scedastie curve is

hopelessly far from a straight line. It is, in point of fact,

logarithmic.

From this table we have x? = 32.115. This is beyond the range

of Elderton’s table. By a rough, but sufficiently noona, graph-

ical TOEA, I find for present values of n’ and x’,

P= .417 about.

In other words, if the butter production of Jersey cows changes

with age according to the curve given, we should expect to

5 For data see ‘‘ Jersey Sires and Their a Daughters,’’ published by

American Jersey Cattle Club, New York, 1

No. 572] > NOTES AND LITERATURE 507

get a worse agreement between observation and theory in 42 out

of every 100 random samples on which the point was tested. In

other words, the fit may be considered sufficiently good. As a

matter of fact, the fit is extraordinarily close over most of the

curve. Four (only) out of the 32 ordinates contribute more than

50 per cent. of the value of x°.

TABLE I

Age in Observed Cale, Butter Standard

Years |Butter Produc-| Production Errors | Frequency! Dev. of

tion in Lbs. in Lbs. Arrays

7 C= yyy

Tk Yy 5 “Pp (Yey z p ) ea Inep g "nrp

1.25 14.25 14.23 .02 2 500

1.75 15.15 15.15 00

2.25 15.57 15.69 AZ 273 1.49 1.771

2.75 15.96 16.06 10 312 on | 932

3.25 16.38 16.35 .03 545 2.07 | 114

3.75 16.72 16.57 -15 511 2.40 | R21

4.25 16.92 16.74 18 04 2.38 4.027

4.75 17.09 16.89 20 532 2.49 3.432

5.25 17.01 17.00 .01 556 2.56

5.75 17.07 17.09 .02 382 2.62 022

6.25 16.98 17.16 18 419 65 1.933

6.75 17.04 17.21 “ae 277 2.68 1.114

7.25 17.09 17.25 16 85 2.68 1.016

7.75 17.48 17.27 21 190 2.68 1.167

5 17.30 17.28 -02 66 2.67

8.75 Was ve 17.27 -10 121 174

9.25 17.56 17.25 31 109 2.61 1.515

16.67 17.21 54 95 2.57

10.25 17.05 17.17 12 2.52 143

10.75 17.42 17.11 39 619

11.25 16.95 17.05 10 2.40

11.75 .00 16.97 28 33 005

12.25 17.05 6.88 17 20 2.26 113

12.75 16.54 16.79 25 7 2.18 09:

13.25 16.34 6.68 34 11 2.09 .291

13.75 18.14 16.56 1.58 9 1.99 5.673

14.25 15.89 44 7 1.88 .599

14.75 16.15 16.30 15 5 1.77 036

15.25 16.37 16.16 21 4 1.65 065

15.75 15.75 00 2 1.53 053.

16.25 15.42 15.84 42 3 1.40 Boys

16.75 15.75 15.67 08 4 1.27 016

Totals... 5,781 32.115

It may be said, in conclusion, that Slutsky’s contribution is

one which will be highly valued by all investigators who have a

critical interest in the graduation of observational data, whatever

the field in which they may be working.

RAYMOND PEARL

08 THE AMERICAN NATURALIST . [Vou. XLVIII

A NEW MODE OF SEGREGATION IN GREGORY’S

TETRAPLOID PRIMULAS

IN a recent paper! Gregory reports a very interesting case in

which two different races of Primulas suddenly gave rise to giant

tetraploid forms, having double the usual number of chromo-

somes, and apparently having the factors doubled also (individu-

ally), for this was true of all the factors which could be followed

in his hybridization experiments. It is important to know how

segregation will take place in such individuals, as there are four

-allelomorphs of each gene present.

Let us suppose that a tetraploid form pure for the dominant

pure recessive giant (22). Gametes AA and aa will meet in

factor A (ana therefore of composition — =) is crossed with a

fertilization, forming the hybrid à = (the maternally derived

genes are represented on one line, say the upper, the paternally

derived genes on the other line).

Now, if this were an ordinary case of ‘‘multiple factors’’? in a

diploid organism, although the two dominant factors, which we

may again call A’s, may produce the same effect upon the organ-

ism, yet they are not interchangeable, and the same is true of the

recessive factors. That is, if we call both dominants A, we must

designate one of them as At, and the corresponding recessives

must also be designated as a and a’, for A will always segregate

into a different gamete from a, and At from a’, there being two

distinct allelomorphic pairs. On the chromosome view of he-

redity, we would say that A and a always lay opposed to

each other, in homologous chromosomes, on the spindle of the

reduction division, as did also A‘ and a‘, but neither A nor a

lay in chromosomes homologous to those of either A’ or a’,

and assorted independently of them. The line-up of factors on

the spindle in the reduction division in this case would be

equally likely to bes = or ai depending merely upon which

1R. P. Gregory, ‘‘On the Genetics of Tetraploid Plants in Primula

sinensis,’’ Proceedings of the Royal Society, 1914.

2%, e, a case where two (or more) independent pairs of factors produce

similar effects, upon the same character. Many examples of this are known,

e. g., the inheritance of red flower in flax.

No. 572] NOTES AND LITERATURE 509

way the pairs are turned with reference to each other. The first

alignment gives gametes AA! and aa’, the second gives Aat and

aA*. Thus three gametes with a dominant factor to one pure

recessive would on the average be produced, the ratio being

1AA:2Aa:1laa, omitting primes.

In a tetraploid form, however, A and A? are alike and inter-

changeable, as also are a and a‘. In the hybrid fara’ therefore,

there would be at least one other mode of pairing of allelomorphs

possible, giving two new modes of line-up on the reduction

spindle, and they would occur just as frequently as the two

previous kinds. The two new arrangements would be z3 x

giving gametes AA! = ata, like those in the first of the two

previous cases, and ai Do giving gametes Aa and atA1.° These

latter gametes would be indistinguishable from the Aat and aA?

gametes given by the second of the two usual arrangements un-

less A could be distinguished from A‘ and a from at. This could

happen only if the allelomorphs were of four different kinds or

if there were linkage of these genes with other genes for which

the plant was heterozygous. Unless, therefore, linkage or mul-

tiple allelomorphism were involved, we could not distinguish

between this mode of pairing of allelomorphs and the usual kind;

both would give three gametes containing at least one dominant,

to one pure recessive (i. e., LAA: 2Aa:laa, omitting primes).

Still a third type of pairing of allelomorphs is possible ina

tetraploid plant, however. There seems no a priori reason, on

the chromosome view, why, in a tetraploid plant, a gene should

have to segregate from one of the allelomorphs derived from the

opposite parent. That is, in a plant of composition — AE

paternally derived genes being indicated on, say, the upper hae

maternally derived ones on the lower, there is no apparent yy

why the line-up of chromosomes at reduction should not be -y At 5

a or as often as it is one of the other types, since all four

chromosomes are homologous. Thus we should get gametes Aa,

Atat, Aa! and Ata

3 If linkage with other genes could be followed, we should with this mode

of pairing obtain crossing over between the chromosomes containing A and

a‘, respectively, and between those containing At and a, respectively; this

would not occur on any other mode of pairing.

510 THE AMERICAN NATURALIST [Vou. XLVIII

We could distinguish such gametes individually from those

obtained by the ordinary arrangements only if linkage were in-

volved, for then we should sometimes obtain results indicating

that the chromosomes containing A and A? had crossed over

with one another, and so had probably behaved as homologous

chromosomes at the reduction division. However, we could also

determine whether this mode of pairing occurred or not merely

by determining the relative numbers of the different kinds of

gametes formed. For, if the third type of pairing occurred, we

should obtain 4Aa gametes in addition to the 2AA, 4Aa and 2aa

derived from the other two types of pairing. The ratio of

gametes would then be five containing a dominant to one reces-

sive, there being 1AA:4Aa:laa, as opposed to the ratio

1AA:2Aa: laa obtainable on either of the other modes of segre-

gation.* The latter or more usual ratio is the only one considered

by Gregory, who apparently takes it for granted that in so far

segregation must be of the same sort as in diploid forms.

Let us see which ratio is more in accord with his experimental

data. As the ratio of offspring in a back-cross is the same as the

gametic ratio, it will be seen that a back-cross of S by a reces-

sive should give 3A : 1a plant on Gregory’s view, the 3A’s consist-

ing of 1 = 25 ee On the other view, a back-cross should result

in 5A: 1a, the 5A’s consisting of 1 = = “ae —, On inbreeding an

4 A plant, however, owing to the random fertilization of

gametes, Gregory’s 3:1 gametie ratio would result in a 15:1

ratio among the offspring (which correspond to F,) and our

own 5:1 gametic ratio would give a 35:1 ratio of A toa among

the offspring.

A summary of his back-crosses of P, heterozygous thrum-eyed

plants of the type , <5 to recessive pin-eyed plants F gives

the result 61 thrum:6 pin (10:1, as compared to the two ex-

pectations 3:1 and 5:1). Among the F, thrums there should

4 Counts of chromosomes in the maturation divisions of the tetraploid

plants show that the chromosomes synapsed in pairs, not in groups of four.

Synapsis in fours would be, in effect, the same as pairing of the random sort

suggested in this paper, so far as any one set of allelomorphs are concerned,

but it might give different linkage results.

No. 572] NOTES AND LITERATURE 511

i A 7 A ,Aa

on Gregory ’s view be aa, a >: on the other view ‘= — a,

ie 85 ags 88

Tests of twenty-one F, thrums, by mating them to themselves

and also to recessives, showed that there was only one which was

: AA ¢

certainly eins and 15 which must have been â 5; (A few gave

numbers too small to be significant, and one or two were of doubt-

ful composition.) This result is within the limits of probable error

on the 4:1, but hardly on the 2:1 expectation. The one F,

thrum plant which was of composition AA gave, on back-

crossing, 67 thrums: 18 pins, a ratio of 3.7:1, to correspond with

Gregory’s 3:1 or my 5:1 expectation. On inbreeding it gave

44 thrums: 2 pins, a ratio of 22:1, to correspond with Gregory’s

15:1 or my 35:1 expectation. The other F, thrums, being of

composition 2 Z (aside from the few doubtful ones), gave, on

the average, 1 thrum: 1 pin on back-crossing, and 3 thrums: 1 pin

on inbreeding; these results would be expected on either view.

rosses were also made involving the character green versus

red stigma (green being dominant). Here the 2s forms, on

back-crossing, gave a total of 114 green: 30 red (3.8:1 instead of

3:1 or 5:1), and on inbreeding they gave 75 green:2 red

(37.5: 1 instead of 15:1, as on Gregory’s expectation, or 35:1,

on my own).

It will be seen that the numbers in the above crosses are too

small to be very significant, individually, for a settlement of the

question at issue, but if summed up they become more decisive.

Thus, a sunimary of the offspring of all back-crosses of the

AA ; : f

=r form to the recessive gives 242 dominants (A) :54 reces-

sives (a), or 4.5:1, as compared with the 3:1 expectation of

Gregory and the 5:1 of the view advocated in this paper.

Where the dominants among these offspring were tested they

were found to consist of 12 = and 15— a 3s compared with the

1:2 expectation of Gregory, and ours of 1:4. Finally, a sum-

mary of the cases where AA forms were inbred shows that 119 -

dominants:4 recessives resulted, a ratio of 30:1 where Greg-

512 THE AMERICAN NATURALIST [Vou. XLVIII

ory’s expectation would be 15:1 and our own 35:1. Moreover,

the individual records fluctuate in both directions about the

ratios to be expected upon our point of view, but practically all

vary in the same direction from the expectation of Gregory,

namely, in the direction of the other expectation.

here is reason, then, to believe that in these Primulas the

factors derived from the same parents may segregate from each

other as allelomorphs, while allelomorphs derived from opposite

parents meanwhile assort at random. For although the allelo-

morphs exist in sets of four they must pair’ two by two for segre-

gation, as do the chromosomes, and two derived from the same

parent may happen to pair with one another. The chance that

this should oceur is one third, since there are three possible

modes of pairing. Such a result is difficult to explain except on

the chromosome view of heredity. It would give ratios different

from those theoretically expected by Gregory, but more in

accord with his experimental data. The principle upon which

our own expectation is founded may be briefly summed up by

saying that where more than two factors which are normally

allelomorphie to each other are present, the pairing of these

allelomorphs with each other preparatory to segregation usually®

takes place at random.

HERMANN J. MULLER

5 That this is not always true is shown by Bridges’ case of ‘‘non-dis-

junctional’’ females of Drosophila, which contain one Y and two X chromo-

somes. Any two of these chromosomes normally act as homologues to each

other in the reduction division of the normal fly, which contains only two `

of them. But where all three are present together they do not pair at ran-

dom, for they oftener undergo the segregation X-XY than XX-~Y, pre-

sumably because the two X’s are much more like each other than like the

Y, and so more apt to act as homologues.

VOL. XLVIII, NO. 573 “ SEPTEMBER, 1914

THE

AMERICAN

NATURALIST

A MONTHLY JOURNAL

Devoted to the Advancement of the Biological Sciences with

Special Reference to the Factors of Evolution

CONTENTS

I. Studies on Inbreeding. Dr. RAYMOND PEAR - z 513

Il. The Chromosome Hypothesis of — sopied to Cases in Sweet Peas and

Primul ALVIN B. BRIDG 524

HI. The SER EEEE as E to P = A. H.

STURTEVANT - - -535

IV. Pattern dite in Mammals and Birds, Dr, GLOVER M. ALLEN - 550

V. Shorter Articles and Correspondence: The Bearing of the Selection Experi.

ments of Castle and = on the TANN of EEREN HERMANN

MULLER - = - 567

THE SCIENCE PRESS

LANCASTER, PA. GARRISON, N. T.

NEW YORK: SUB-STATION 84

The American

Naturalist

intended for ane soe and books, ete., intended for a should be

MSS

sent ago Editor of THE AMER

articles containing summaries of resea

CAN NATURALIST, —— -on-Huds n oe York.

wor

ing on the

problems of Sod cg evolution are especially welcome, ana will be gees sretervies

in pen

ea reprints of ne per foro are supplied to authors free of charge.

Further ‘eprint will be supplied a

Subscriptions and adv

subbertpiton oes Hed four

Canadian posta

ertisoments should be sent to the publishers.

ollars a year. Foreign postage is fifty cen

wenty- -five ‘eis additional.

The

nts and

The arie for single copies is

forty cents. Tho: i marefa rates are Four Dollars for a

THE SCIENCE PRESS

NEW YORK: Sub-Station 84

Lancaster, Pa.

Entered as second-class matter, age 2, 1908, at th

Garrison, N. Y.

Post Office at Lancaster, Pa., under the Act ot

Congress of aih 3, 1879.

FOR SALE

ARCTIC, ICELAND and GREENLAND

BIRDS’ SKINS,

Well Prepared Low Prices

Particulars of

G. DINESEN, Bird Collector

Husavik, North Iceland, Via Leidle, England

WANTED TO PURCHASE

a set of BIRDS OF AMERICA by J. J. Audubon,

7 or 8 volumes, please report, stating cash price, stat-

ing condition, binding and dates of volumes.

F. C. HARRIS,

Box 2244 Boston, Massachusetts

For Sale Entire

An important collection of Indian Birds’

Eggs, containing approximately 3,300

specimens of about 620 species, and in-

cluding among other varieties the Green-

ish Willow Warbler (Acanthopneuste

viridanus), Red-browed Finch (Calla-

canthis burtoni), Tibetan Sand-Grouse

(Syrrhaptes tibetanus), Spotted Sand-

Grouse (Pteroclurus senegallus), Ibis-

bill (Ibidorhynchus struthersi), White-

bellied Minivet (Pericrocotus erythro-

pygius), Bearded Vulture (Gypaëtus

barbatus), numerous Cuckoos with their

hosts, ete., ete.

Particulars and full list may be had

from

W. F. H. ROSENBERG

57 Haverstock Hill London, N. W.

Marine Biological Laboratory

Woods Hole, Mass.

INVESTIGATION

Entire Year

Facilities for research in Zoology,

nae ag an ee

for such a table

Courses labora’ instruota

INSTRUCTION Sack N RA om

July b E

1 to August 11,

1914

$50.0

Philosophical Aspects of a? aaee

Allied Sciences

-e so tn gator = oP

Open e ear

logi Botanical ace

epee ft at

information regarding material,

address

GEO. M. GRAY, Curator, Woods Hole, Maas

The annual announcement will be sent on ai =

The Director, Marine Biological Laboratory, Woods Hole

Mass-

THE

AMERICAN NATURALIST

VoL. XLVIII September, 1914 No. 573

STUDIES ON INBREEDING. V

INBREEDING AND RELATIONSHIP COEFFICIENTS !

Dr. RAYMOND PEARL

UNIVERSITY OF MAINE

In the discussion of inbreeding coefficients contained in

a series of recent papers from this laboratory” no mention

has been made of an important consideration which arises

in connection with such coefficients. The further problem,

to which we may now turn, may be stated in the follow-

ing way.

The pedigree of an individual consists of two halves.

One of these halves is made up of the sire and his an-

cestors; the other of the dam and her ancestors. Follow-

ing the conception of inbreeding set forth in detail in the

earlier papers of this series it is plain that the values of

the coefficients of inbreeding for a particular pedigree are

composed of the following elements.

1. The occurrence of the same individual animals more

than once on the sire’s side of the pedigree only.

2. The occurrence of the same individual animals more

than once on the dam’s side of the pedigree only.

1 Papers from the oo gaa Laboratory of the Maine Agricultural Ex-

periment Station, No. 6

2 Pearl, R., Paa on Inbreeding. I. A Contribution Towards an

Analysis of the Problem of Inbreeding,’? AMER. Nat., Vol. XLVII, pp.

577-614, 1913; ‘‘The Measurement of the Intensity of Inbreeding,’’ Me.

Agr. Expt. Sta. Bul., 215, pp. 123-138, 1913. Pearl, R., and Miner, J. R.,

‘‘ Studies on Inbreeding. III. Tables for Calculating Coefficients of In-

breeding,’’ Me. Agr. Expt. Sta. Ann. Rept. for 1913, pp. 191-202, 1913.

513

514 THE AMERICAN NATURALIST [Vou. XLVIII

3. The reappearance of animals which appear first on

one side of the pedigree (either the sire’s or the dam’s)

on the other side.

If only 1 and 2 are to be found in the pedigree it means

that the sire and the dam are totally unrelated (within the

limits covered by the pedigree in the particular case).

the other hand, the occurrence of 3 means that sire and

dam are in some degree related, and that a portion of the

observed inbreeding arises because of that fact. Now

the coefficients of inbreeding, in and of themselves, tell

nothing about what proportionate part has been played

by these three elements in reaching the final result. It is

a matter of great importance to have information on this

point, because of its genetic significance. It is the pur-

pose of this paper to describe a general method for ob-

taining this desired information.

The first step in the method, stated briefly, is to break

up the pedigree elimination table formed to get the suc-

cessive values Of pn; — Gnu, in our former notation, into

four different parts. One of these parts will include the

primary reappearance on the sire’s side of the pedigree of

such animals as appear first on the same side. This may

be called the ‘‘male only” table. The second part will

include the primary reappearance on the dam’s side of

such animals as first appear on the same side. This is

the ‘‘female only’’ table. The third part will include the

primary reappearance on the dam’s side of such animals

as first appear on the sire’s side. The fourth part is the

reverse of the third. These last two may be called the

‘‘eross tables.’’? The sums of the totals of these partial

tables will give the total pn. — qn. Values for the succes-

sive generations.

The formation of the tables on this plan may be illus-

trated with some examples. These examples will also

show the skeleton method of writing pedigree elimination

tables, which saves much labor. This was referred to,

but not significantly illustrated, in the earlier papers. It

consists simply in doubling the total of the column for

each generation rather than the separate items.

No. 573]

STUDIES ON INBREEDING

TABLE I

515

PARTIAL PEDIGREE La on eh TABLE FOR Kine MELIA Rioter 14TH SHOW-

IN RY REAPPEARANCES ON THE SIRE’S THE

Ponia OF as WHICH FIRST APPEAR ON THAT SIDE

E PRIM

SIDE OF

Generation

4| 5

Melia pond : Son

Melia Ann 8d... c.i

Lucy’s aoe. Pogis .

Melia Ann

St. oe tet

Letty Rio

Allie of es Ginter’:

mele Cine 6 68 6 6 eee Pe wpe ete

Oe ee ee 8S Ww OSs Pee tow es ve

ate

Huo’ s iran

Vi ae wie

Lord Lisg

Lucy of St. Tak

Diana of St. Lamberts:

Olof. of St. Lambert .

Oe eb ele ee kegs 6 ate WLS

e E a eee ee ee wey ee E ae

eee eee we ete

Oa sere eee se sate

a ee E Sa la D ee A E,

tht we ae se Gee ee we es

Cee ee ee ee ere ee Ged

eee ee rere tv ele

Ce eee ee ee er er)

ve we ele stare

Pee eee eee et ee we le epee fe wpe te fe see

` d

Ce wee ee a A a h E

r eo ie, ee a ee E E a

E E ae a E Oe ey eee E Rm LA Ga ee

EIE ea

Shae gone i). cera

few eee

Ps ee eee fo ae eee

ere thee sprees oa

Oe Se ee oe a ae Oe ee

Sew eine eel ee wee

Pe ee ee ee ee a ee

epee ee be ee

ee ee ee EAS EN eee ee ee ee ee

A Hee

ee ge tel ae

ee eee

ee ee eee ee

age e eRe molar ye

ns Fae th Ph

eee

tee

Pu es

i Fo 8 6

ee ee fae ee ae

pescao rw tas

Peer ee ee bee ee

se ee ede

oe oe ee en ae ie, Se ee 6

ee eee ey

fee ee ete Ss ee ete

see eee

P ee E

E TAE EAE S

see eee

s.s.s

+62 ee S

ee eee

s.es...

se ee ele

‘Te hee ee ieee ss

es ee a ed oe We eae |

447 | 898

2an

S this and au fale

Ltd

ng table the numb

pds the sum of the numbers in the

umulated ancestral oe up to the So in question.

ers in brackets ar

e in each case

g column. They represent, the

516 THE AMERICAN NATURALIST [Vou. XLVIII

The pedigree for 12 ancestral generations of the Jersey

bull King Melia Rioter 14th (103901) may be taken as the

first illustration.

TABLE II

PARTIAL PEDIGREE ELIMINATION TABLE FoR KING MELIA RIOTER 14TH SHOW-

GONO ecrini 2 3 4 goteo DA

King’s Bioter Lad sonh — — — 1 2 4 8 16 32 64 128

Table III is clearly the one which demands special

attention. As will shortly appear, it is the most important

for the theory of inbreeding. Let us attempt its analy-

sis. Just what does the first entry mean genetically ? Tt

states that King Melia Rioter, an animal which first ap-

peared on the sire’s side of the pedigree, reappeared in

the second ancestral generation on the dam’s side. What

this clearly means is that at least one half of all the dam’s

ancestors, in the third and higher ancestral generations,

are identically the same animals as are ancestors of the

—

cno P

ih eh ER:

“<7

bro

COEFFICIENTS

á 4 e 8 s0 a sa

GENERATIO

Fic. 1. Diagram a (a) Sees ge 1 inbreeing (heavy solid line) and (b)

the ehren (heavy broken line) for the Jersey bull, King Melia

Rioter To The KGN i order of oa taik and relationship between the sire

this case is evident by comparison with the lighter lines, which give

the rer it values for continued brother x sister, parent x offspring and cousin

reeding.

No. 573] STUDIES ON INBREEDING 517

sire. The next entry in Table III indicates that in the

fourth and higher ancestral generations at least 5/8 of

all the dam’s ancestors were the same individual animals

as were also ancestors of the sire. One half of them were

the same before the reappearance of St. Lambert’s Rioter

King. He makes up the additional 1/8 of the dam’s

ancestry.

TABLE III

PARTIAL PEDIGREE ELIMINATION TABLE FOR KiNG MELIA RIOTER 14TH SHOW-

ING THE PRIMARY REAPPEARANCES ON THE DAM’sS SIDE OF THE

PEDIGREE OF ANIMALS WHICH FIRST APPEAR ON THE SIRE’S SIDE

E a 6/0 sc1s es slala P@ 171 s | 9 (20114) 18

King Melia Rioter......... i1] B Dle E E epe

St. Teahibact’s Ristet King: Rs E a bwai E Ra S N. D

King of St. Lambert........ te see Oe ep Be oe Aa) oie Renee

e o; |

CORNO S SO o oo de EE S D N E E are ee

St. La: mbert’s Rioter King | oped egies ent GOw a ies Pica cule <> NEE Sota

St. Lambart: Boyce... AA E E Sern E Fae sechas

Tol eaa aea |1| 2 | 5 | 12 | 28 | 59 | 119 | 240 480/960 1,920

From these tables it is obvious that a very considerable

portion of the inbreeding shown in the pedigree of King

Melia Rioter 14th arises from the fact that his sire and

dam were closely related. Furthermore, both sire and

dam are closely inbred in their own lines. The curve of

. total inbreeding in this case is shown in Fig. 1, along with

the curves for continued brother X sister, parent by off-

spring, and cousin X cousin mating.

TABLE IV

SUMMARIZED PEDIGREE ELIMINATION TABLE FoR Kine MELIA RIOTER 14TH

Géneration 2. . 2.5 ii: i846 16) 7 & es Hi a

BF oniy. oai. urea 1| 3/16 41|105|219|447| 898| 1,796

S O aa a 3 2; ál Bi 16) S| i 128

Oross-over.. aasa 1/2|/5|12/28| 59|119 |240| 480; 960 1,920

Teste... 4, 2| 6 16 46 104 |232 |475| 959 1,922 | 3,844

From this we have, for = inbreeding coefficients,

518 THE AMERICAN NATURALIST [Vou. XLVIII

Ze ed

Z: == 25,00

Z, == 25.00

P So

2, >=> 50.00

T a OR Ee

Ze == S120

Z; = 90.63

D aP

Z; == 93.05

Ze == 95.80

Za == 93.85

These facts will possibly be made clearer to those not

actually working much with pedigrees by Table V, which

gives the first four ancestral generations‘ of the pedigree

of King Melia Rioter 14th.

Generalizing the above reasoning we get the following

result.

In A,, and: higher ancestral generations, 2/4 = 50.00 per

cent. of the dam’s ancestors are animals which are also

ancestors of the sire:

In A,, and higher ancestral generations, 5/8 = 62.50 per

cent. of the dam’s ancestors are animals which are also

ancestors of the sire.

In A,, and higher ancestral generations, 12/16 = 75.00 per

cent. of the dam’s ancestors are animals which are also

ancestors of the sire.

In A,, and higher ancestral generations, 28/32 = 87.50 per

cent. of the dam’s ancestors are animals which are also

ancestors of the sire.

In A., and higher ancestral generations, 59/64 = 92.19 per

cent. of the dam’s ancestors are animals which are also

ancestors of the sire.

4 In the study of pedigrees stress is naturally laid on the ancestral genera-

tions, rather than on the filial, as in breeding experiments. It becomes very

convenient to have a brief designation for ancestral generations, in the same

way that F,, F., ete., are used to denote filial generations. I would suggest

the use of the letter A with sub-numbers for this purpose. We then have A,

denoting the parental generation, A, the grandparental, A, the great-pa-

rental, ete.

No. 573] STUDIES ON INBREEDING 519

In Ag, and higher ancestral generations, 119/128 =

per cent. of the dam’s ancestors are animals AENG are

also ancestors of the sire.

In A,, and higher ancestral generations, 240/256 = 93.75

per cent. of the dam’s ancestors are animals which are

also ancestors of the sire.

In A,,, and higher ancestral generations, 93.75 per cent.

of the dam’s ancestors are animals which are also an-

cestors of the sire.

In A,,, and higher ancestral generations, 93.75 per cent.

of the dam’s ancestors afte animals which are also an-

cestors of the sire.

In A,» and higher ancestral generations, 93.75 per cent.

of the dam’s ancestors are animals which are also an-

cestors of the sire.

TABLE V

PEDIGREE FOR Four ANCESTRAL GENERATIONS OF KING MELIA RIOTER 14TH

|©. |© |No. 63200 *|No. 56581 No. 2 J

| x Melia yoa s Son.

b Melia Ann’s King. |No. 100775 9

Marjorie _ Lottie Melia Ann. _

x Melia Ann’s |No. 157263 Q No. 22041 Eg

= Son. Le : @ Melia Ann’s Son.

= Marjorie Melia

faa Ann. No. 9 05883 e Q

| S Mary Melia Ann

| = No. 181544 9 No. 58169 d'No. 54896 ea

| i _ St Lambert’s Rioter King.

| A King of All Kings. |No. 114804 Q

peoa 5 S Letty St. Lambert’s Letty.

$e Silver |No. 148456 Q No. 32559 F

Bik Hair. : j _ Exile of St. Anne’s.

gn Exile’s Silver

olo Hair. No. 60449 2

ie — ae 4th.

So |No.73104 "No. 63200 J No. 5658 g

~ ® k y ’s King.

= ® Marjorie Melia © Mel elerna oen

a Ann’s Son. No. 157263 à y

£ @ King Melia - & Marjorie Melia Ann.

m Riste. |No. 181544 Q|No. a

2 Q King of All Kings.

3 ® Letty Silver Hair. |No. 1 re)

= | & mae s a ive Hair.

% No. 219360 2 No: 62098 J |No. 54896 ae a)

St. Lambert’s Rioter King.

| E King Rioter’s @ St. Lam >

| 4 Lad 0. 14 2

mo he ' Dula i King’s Riotress Nora.

S iB Riotress No, 218796 Q\No. 57778 m

So 9 Maid. St. Lambert's Boy.

PE ins: St. Lambert’s No. 174761 9

Zz |e cine ON Rioter Lad’s First Daughter.

520 THE AMERICAN NATURALIST [Vou. XLVIII

These percentages are quantities of a good deal of

interest. They measure the degree in which King Melia

Rioter 14th’s sire and dam were related to each other.

Community of ancestry is the basis of kinship.

Percentages derived in the way shown above, from

cross pedigree elimination tables, I propose to call co-

efficients of relationship, and to designate by the letter

K, with appropriate sub-numbers referring to the genera-

tion. These relationship coefficients are, with some limita-

tions, independent of the inbreeding coefficients in the

values they may take, though the two will usually be cor-

related to some degree. It is, however, possible to have a

high value of Z with K—v.

TABLE VI

COMPARING THE MAXIMUM POSSIBLE VALUES OF THE COEFFICIENTS OF IN-

BREEDING (Z) WHEN THE COEFFICIENT OF RELATIONSHIP K

EQUALs (a) ZERO, AND (b) 100

Generation Maximum Possible Value Maximum Possible Value

of Z when K = 0 of Z when K = 100

i 0 0

A, 0 50.00

Ag 50.00 75.00

A, 75.00 87.50

A, 87.50 93.75

As 93.75 96.88

A, 96.88 98.44

As 98.44 99.22

Ay 99.22 99.61

Aw 99.61 99.80

The most important feature of the relationship coeffi-

cients is found in their genetic implications. This can be

indicated best by an illustration. Let us consider the case

of the maximum possible degree of inbreeding with K = 0.

This will be found when the sire and the dam are each

inbred to the highest possible degree (continued brother

X sister mating) but are in no way related to each other.

Such a case would be afforded, for example, if a Jersey

bull, the product of continued brother X sister mating,

was bred to a Holstein cow, which was also the product

No. 573] STUDIES ON INBREEDING 521

of a continued brother by sister breeding. Clearly K

would be 0, since no animal on one half of the pedigree

could even appear on the other. The values of the suc-

cessive coefficients of inbreeding (Z’s) in such a case are

shown in Table VI, where they are compared with the

coefficients of inbreeding in complete continued brother

X sister mating, where K = 100.5

From this it appears that an individual may be inbred

in 10 generations to within two tenths of one per cent. as

intensely, measured by the coefficients of inbreeding, if

his sire and dam are in no way related, as he would be if

his sire and dam were brother and sister. But clearly the

germinal constitution of the individual produced would,

except by the most remote chance, be quite different in the

two cases. This point is so evident as to need no elab-

oration. It has been brought out by East and Hayes.®

The values of the K’s for a particular pedigree evi-

dently furnish a rough index of the probability that the

two germ-plasms which unite to form an individual are

alike in their constitution. This will follow because of the

fact that the probability of likeness of germinal constitu-

tion in two individuals must tend to increase as the num-

ber of ancestors common to the two increases. Just what -

is the law of this increase in probability is a problem in

Mendelian mathematics which has not yet been worked

out. The general fact, however, seems quite sure.

From the above discussion it seems plain that in reach-

ing a numerical measure of the degree of inbreeding it is

not sufficient to consider coefficients of inbreeding alone.

The coefficients of relationship must also be taken into

account.

It is suggested that the two constants be written to-

gether for each generation, the coefficient of inbreeding

being followed by the coefficient of relationship in brackets.

Thus we have

ga of course, all of a sister’s ancestors are identical with her

brother’

tU. s. Dak Agr. Bur. Plant Industry, Bul. No. 243, pp. 1-58, 1912.

522 THE AMERICAN NATURALIST [Vot. XLVIII

INBREEDING AND RELATIONSHIP COEFFICIENTS OF Kina MELIA RIOTER 14TH

Zy (K,) = 0 (0)

Z, (Ky) = 25 (0)

Z, (K ) = 25.00 (50.00)

Za (K, ) = 37.50 (62.50)

Z, (K, ) == 50.00 (75.00)

Zs (Ky) ==71.88 (87.50)

Z. (K, ) = 81.25 (92.19)

Z, (K, ) =90.63 (92.97)

Z, (Ky) = 92.77 (93.75)

Zə (Ky) = 93.65 (93.75)

Zy(Ky) = 93.85 (93.75)

Zy,(K,2) = 93.85 (93.75)

The physical meaning of these expressions is simple

and straightforward. Z,(K,) tells us that in the 5th an-

cestral generation of King Melia Rioter 14th he had only

one half as many different ancestors as was possible for

that generation, and of his ancestors three fourths were

common to his sire and his dam. However one looks at

.the matter there can be no denial that King Melia Rioter

14th is a closely inbred animal.

In Fig. 1 the heavy broken line gives the relationship

coefficients for King Melia Rioter 14th. It will be instruct-

ive now to consider another example by way of contrast.

Again a Jersey bull, Blossom’s Glorene (102701), will be

taken. Only the final result need be given.

INBREEDING AND RELATIONSHIP COEFFICIENTS OF BLOSSOM’S GLORENE

Fel hy) =O (0)

ZAKS) y (0)

Z.(K,) 12.50 (0)

ZE) == 1256 (6)

Z,( Ks) = 25.00 (0)

Z,(Ke) == 29.69 (0)

Ze(K:) = 35.94. (0)

Z,(K,) = 40.23 (0)

The total inbreeding and the relationship curves are

given in Fig. 2.

The difference in the breeding of this bull and the one

considered in the former example is striking. In the 8th

ancestral generation Blossom’s Glorene has but 60 per

No. 573] STUDIES ON INBREEDING 523

cent. of the number of different ancestors possible in that

generation, but not one single animal in the ancestry of

his sire occurs in the ancestry of his dam (within the

limits A, to A,). The probability is that Blossom’s Glo-

rene is heterozygous in respect of most of his characters,

while King Melia Rioter 14th is homozygous.

——

—_——

COEFFICIENTS

i a Se Sa N

ae Tee maa i

s 4 é 8 40 te ta

GENERATIONS

Fig, 2. Diagram showing the total inbreeding (heavy solid line) and the

relationship (heavy broken line) curves for the Jersey bull Blossom’s Glorene,

a period of eight ancestral generations, Compare with Fig.

SuMMARY

The object of this paper is to call attention to the fact

that inbreeding of considerable degree may exist in the

entire absence of any kinship between the two individuals

bred together, and to bring forward a method of sepa-

rately measuring what proportion of the observed in-

breeding in a particular case is due to kinship of the pa-

rents, and what to earlier ancestral reduplication. A pro-

posed coefficient of relationship is described, and its ap-

plication illustrated by concrete cases.

THE CHROMOSOME HYPOTHESIS OF LINKAGE

APPLIED TO CASES IN SWEET PEAS

AND PRIMULA

From the Zoological Laboratory, Columbia University.

CALVIN B. BRIDGES

THERE are two views as to the nature of linkage. The

earlier view, developed by Bateson and his co-workers, is

that this phenomenon is an expression of symmetrical

reduplications in the germ tract. A more recent view,

developed by Morgan and his co-workers, treats linkage

on the basis of a linear arrangement of genes in the

chromosomes and of the history of these genes during

normal gametogenesis. The advocates of the reduplica-

tion view have rarely applied their principles to the re-

sults on Drosophila on the ground that the results for

Drosophila are complicated by sex-linkage. That sex-

linkage is simply an additional, but wholly independent,

phenomenon, is proven by the many cases in Drosophila

in which sex-linkage is not involved, yet in which the link-

age of the genes to each other is of the same type as the

linkage of sex-linked genes to each other.

In this paper I shall attempt to show. that the theory of

linkage which we have successfully applied to all cases in

Drosophila, whether involving sex-linked genes or genes

which show no sex-linkage, applies equally well to the

non-sex-linked cases occurring in sweet peas and primula.

The only serious drawback to such an application lies in

the nature of the data which have been collected for these

eases. The least satisfactory form of data from which to

determine a linkage value is that presented by F, results.

In eases in which two recessives enter from opposite par-

ents (‘‘repulsion’’), the excessive smallness of the double

recessive classin F, renders any calculation subject to great

error. Slightly better are the F, results from coupling,

524

No. 573] CHROMOSOME HYPOTHESIS OF LINKAGE 525

but here there is no direct parallelism between the

zygotic and gametic ratios. In determining what gametic

ratio underlies the F, results given by an experiment, the

practise has been to compare by the eye the given result

with a series of F, results calculated from selected gametic

ratios. Collins has shown’ that this practise has led to

serious error. In F, coupling cases in which there has

been no crossing over in one sex (autosome genes in

Drosophila), there is a direct relation between the gametic `

and zygotic series, but only in certain classes which com-

prise from one fourth to less than one half of the indi-

viduals of an experiment. While such data are more

accurate than the usual F, results, yet the percentage of

individuals which can be used directly is so low that we

avoid the use of such a method. In F, results involving

only sex-linked genes, the efficiency is at least 50 per cent.,

for here there is always a direct relation between the

gametic and zygotic ratios in one half the flies (the males).

However, half the total number of flies (the females) are

useless unless the cross is made in such a way that F,

becomes a back cross. These different kinds of F, results

(the two most advantageous of which are not generally

applicable) are separated in effectiveness by a wide gap

from the back cross which we use equally well in all cases,

which gives a zygotic ratio directly proportional to the

gametic ratio, and in which every individual occurs in the

most advantageous relations.

Perhaps the least unsatisfactory method of dealing

with such F, series as are available in the case of the

sweet peas, is by means of the coefficient of association as

derived by Yule. Yule’s coefficient of association is caleu-

lated from a zygotic series of the form AB:aB: Ab: ab by

the formula:

Coefficient of association = 4B corer on = i

To find the gametic ratio corresponding to this coeffi-

cient, use is made of a table which gives the coefficients

1 Am. Nart., ’12.

526 THE AMERICAN NATURALIST — [Vou. XLVIII

calculated from the zygotic series corresponding to such

gametic ratios as 2.5:1, 3:1, 3.5:1, ete. For the same

ratio in the coupling and repulsion series the coefficients

are slightly different, so that two tables should be made.

-Upon the chromosome basis the best method of express-

ing the amount of linkage is in terms of percentage of

crossing over. The gametie ratio n:1 found through the

coefficient Di association, when expressed as a percentage

becomes -

ot i.

According to the chromosome hypothesis, all genes

which are linked to each other lie in the same chromosome.

In sweet peas the first case in which linkage was observed

was that of round pollen? and red flower color. Later it

was found that hooded standard was linked to round and

to red. The genes for these three characters, then, may

be treated as though carried by the same chromosome,

which we may call chromosome I, of the sweet pea.

The relative distances of these genes from one another

in the chromosome can be determined from the degrees of

linkage. The farther apart in the chromosome any two

genes lie, the greater will be the amount of crossing over

between them. If two genes lie very close together, then.

the percentage of crossing-over will be very small (the

gametic ratio very large).

Fortunately Punnett has recently collected the data

upon these linkage cases in sweet peas. In the table which

follows, I have summarized the data given by the various

tables of Punnett. In the first column to the right of the

data appear the coefficients of association. In the next

column appear the corresponding gametic ratios caleu-

lated by interpolation to the nearest tenth. In the last

column are the equivalent percentages of crossing over,

found from the gametic ratios.

We may use one per cent. of crossing over as our unit

of distance in measuring the space between two genes.

2 T have used a terminology here like that used for the cases in Drosophila,

naming the gene after that member of the pair of allelomorphs which may

be considered as the mutant from the wild type of pea.

No. 573] CHROMOSOME HYPOTHESIS OF LINKAGE 527

The gene for red is then about eleven units from that for

round, and the gene for hooded is nearly one unit from

that for red.

TABLE I

CHROMOSOME I

Round Pollen and Red Color

Coefficient Percent-

of age of

Associa- Gametic Cross-

Wild Type Round Red Round Red tion Ratio overs

Coupling shed OT 583 614 2,197 9596 7.931 11%

Red Color and Hooded Standard

Wild Type Red Hooded Red Hooded

Coupling ...2,568 16 17 857 9998 125023 38

Round Pollen and Hooded Standard

Wild Type Round Hooded Round Hooded

Coupling ... 626 74 83 174 .8932 4.7:1 18.

Repulsion .,.3,140 1,413 1,438 14 9577 SASL 10.3

The order of arrangement of these genes in the chromo-

some can be discovered from a comparison of the linkage

values found above: The linkage value (11.2) for round

and red is the most accurately determined of those in-

volved, so that we may lay this down às our initial or

base line:

Ro R

0 11.2

DIAGRAM I. Rọ= round pollen, R=red flower.

The next most accurate value is that for red and

hooded, namely, 0.8. Hooded lies therefore only about

one unit from red, but if these two values only, namely,

round red and red hooded, were given, we should be un-

able to decide whether hooded lies between round and red

at a position near 10 (that is, 11.2 — .8) or beyond red in

a locus at 12 (that is, 11.2+.8). In order to determine

whether hooded lies to the left or to the right of red the

data for the third value, round hooded, need only be accu-

rate enough for us to decide between these values of 10

528 THE AMERICAN NATURALIST [Vou. XLVIII

and of 12 units. The data from the coupling experiments

(which even though less extensive then those from the

repulsion experiments are probably more accurate) give

a value of about 18 units. Since the repulsion data give

10 units, 18 is probably too high, and an intermediate

position correct. The higher (12) of the two possible

values is then the correct value. The position at 10 is not

excluded by these data, but is far less probable. In a case

in which one of the two first values is very small, as here,

the accuracy demanded of the remaining or third value is

much greater than in cases where neither of the values

are small, and one has only to decide between two very

different values by aid of the third. There are other ways

of arriving at this order of genes which are independent

of the size of the values. One of those methods, such for

example, as that of double crossing over, would definitely

settle the order of these three genes, but unfortunately

such data have not yet been published.

If hooded lies beyond red at 12, the complete first chro-

mosome diagram will be as follows:

Ro RH

0 2 2.

DiaGrAM II. Chromosome I, Sweet Pea. Rọ= round pollen, R= red flower,

H ooded,

In the above diagram R, indicates the locus of round

(and also of long). The symbols in the diagrams are

used to designate loci which may be occupied by either

allelomorph of the pair.

It has been observed that hooded flowers have always

a uniform color in standard and wings, instead of having

these two regions colored differently as in the normal or

bicolor type. Bateson assumed that this unicolorism was

only another somatic effect of the hooded gene. However,

an alternative explanation is that the unicolor is caused by

a specific gene which is very closely linked to hooded. If

this should be found to be the case, then this fourth gene

also will be located at about 12 units from round.

No. 573] CHROMOSOME HYPOTHESIS OF LINKAGE 529

There is one other gene which probably belongs in the

first chromosome, namely, the intensifier found in the

‘ black knight’’ race. The linkage data of red color and

intensity of color have been given in Report II to the

Evolution Committee, page 90.

TABLE II

Red Color and Intense Color

~~ Feoi

Associs- Gametic Cross-

Wild Type Red Intense Red Intense tion Ratio overs

Coupling eee |] 29 35 22 527 49:1

If these data are significant, then intense is in the first

chromosome at a locus about 35 to the right or left of red.

It should give about 24 (35 — 11) or 46 (35 + 11) per cent.

of crossing over with round, depending on whether it lies

about 24 to the left of round or 35 to the right of red.

THE SECOND CHROMOSOME OF SWEET PEAS

In the case of the second chromosome in sweet peas,

the linkage values are based on smaller numbers, but the

order of genes is more certain.

The first linkage case of this imon was that of

sterile anthers and light axils. Later the cretin form of

ower was found to belong to this linkage group. As in

the case of the first chromosome, I have summarized the

tables of Punnett in Table III.

TABLE III

CHROMOSOME II

Sterile Anthers and Light Axil

Coefficient Percent-

; Associs- Gametic Cross-

a d Type Sterile Light Sterile Light tion Ratio overs

Cansine p 170 41 30 379 .9945 22.: 1 4.4

Repulsion . 1,335 643 714 2 -988 20.:1 4.9

Light Axils and Cretin Flower

Wild Type Light Cretin Light Cretin

Coupling . . 282 49 52 59 .734 2.6:1 28.

Ropulkiok - so 22 27 3 .610 27:1 27.

530 THE AMERICAN NATURALIST [Vou. XLVIII

Sterile Anthers and Cretin Flower

Wild Type Sterile Cretin Sterile Cretin

Coupling mera! Ls 58 78 55 21 33.

Repulsion .. 764 355 345 25 .683 2.6:1 28.

The linkage value for sterile and light, namely, 4.4 units,

‘is the most accurately determined of those in the second

chromosome. The value for light and cretin is about 28

units. Using the distance 4.4 between sterile and light as

our base line, then, we should find that cretin lies at 4 + 28

or 32 from sterile if the order of genes is sterile, light,

cretin; but if the order is cretin, sterile, light, then cretin

should lie at 28 — 4 or 24 from sterile. The value for

sterile cretin should approximate either 24 or 32. There

is no very small value here as there was in the first chro-

mosome, and not such great accuracy is required of the

remaining value, since it should be easy to distinguish be-

tween 24 and 32. The coupling data for this value gives

33 units, which enables us to fix the order of genes as

sterile, light, cretin. The following diagram of chromo-

some IT expresses these relations more clearly.

L C

4.4 32.

D1aGRaAM III. Chromosome II, Sweat Pea. S= sterile, L = light, C = cretin.

orm

When crossing over is as free as in the case of sterile

and cretin and of light and cretin there should be some

double crossing over. That is, crossing over might occur

in the section of the chromosome near sterile and light

and at the same time another crossover could occur in the

section between light and cretin. This occurrence would

be readily seen if normal plants heterozygous in any com-

bination of these three genes were back-crossed to plants

purely recessive in all three. A relatively few plants from

such a test would give very valuable information on sev-

eral points, while an experiment of a few thousand indi-

viduals from such back-cross tests would enable one to

discover, through the phenomenon of interference, much

No. 573] CHROMOSOME HYPOTHESIS OF LINKAGE 531

as to the character of the chromosome, the average length

of the internode, and the percentage of chiasmas per node.

INDEPENDENCE OF CHROMOSOMES I AND II or Sweet Pras

If two groups of genes are carried by separate chromo-

somes, we may expect to obtain free assortment and

typical 9:3:3:1 ratios in F,, when any two genes from

different groups are involved. There are rather extensive

data for three such cases in sweet peas, and in each there

is practically complete independence. The data given in

able IV are summarized from Report III to the Evolu-

tion Committee (page 37) and Report IV (page 17).

TABLE IV

INDEPENDENCE OF THE FIRST AND SECOND CHROMOSOMES

Round Pollen (1st) and Light Axil (2d)

: wes apne Percent-

; Associ

Wild Type Round Light Round Light tion

1,246 341 399 142 ASL 225001 47.

age

a- Gametie Cross-

Ratio, Ts

Red Color (1st) and Light Axil (2d)

Wild Type Red Light Red Light

1,563 545 506 232 .136 LI6:1 47.

Red Color (1st) and Sterile Anthers (2d)

Wild Type Red Sterile Red Sterile

838 403 265 071 10i B

The greatest departure from the 50 per cent. of cross-

ing over expected from independent assortment is only

to 47 per cent.

There are several other characters whose genes seem to

be independent of those in the first and second chromo-

somes. This is interesting from the point of view that

each independent gene or group of linked genes requires

a distinct chromosome as a carrier.

508 THE AMERICAN NATURALIST — [Vou. XLVII

LinKAGE CASES IN PRIMULA

In the case of primula, linkage was first found between

red (versus green) stigma and red (versus magenta)

flower color. Long style (versus short) and dark stem

(versus light) were found to be linked with red stigma.

Indications were observed that still a fifth gene, a domi-

nant which reduces the color of the flower to a tinge in the

corolla tube, belonged to this group.

back cross involving the three genes, red stigma, red

flower and long style was made. Credit is due to Gregory

for the use of this method for obtaining linkage data.

Unfortunately many of the individuals were useless for

the linkage of red flower color, because of the occurrence

of white; and the numbers are small.

In Table V, I have summarized the data given by

Gregory.*

TABLE V

THE First CHROMOSOME OF PRIMULA

Red Stigma and Red Flower

Non-crossovers Crossovers

ah Q

z vo ES 2 àc i

Bo wf ei yi ees zs EB

od Fe ag mS egs ge 295

2a Os 7 a”

Coupling

DOCK eros oiio 28 39 17 18 1.9:1 34.6

Wild Red Red Red Stigma

Type Stigma Flower Red Flower

Coupling Fes... 1,174 305 289 232 O11 1.8:1 35.3

Red Flower and Long Style

Non-crossovers Crossovers

Red Long Wild Type Red Long

Coupling back cross. 40 53 6 5 8.4:1 10.9

Wild Type Red Long Red Long

Coupling fT ie oar 38 2 n o 12 .966 8.6:1 10.4

Red Stigma and Long, Style

Coupling back cross. 44 64 5 30 1.6:1 37.

4 Jour. Genetics, ’11, Vol. I; Proe. Roy. Soc., ’11, Vol. —, 84.

No. 573] CHROMOSOME HYPOTHESIS OF LINKAGE 533

Red Stigma and Dark Stem

Wild Type Red Stigma Dark Red Stigma Dark

mepulsion 34255 k 137 66 62 0 -— — —

The three values are—red stigma red flower 35, red

flower long style 11, and red stigma long style 37. Of

these, red stigma red flower is based upon the most data,

_ and may therefore be taken as our base line. The value

for red stigma long style should be 35 — 11 or 24, if the

order of genes is long, red stigma, red flower; but 35 + 11

or 46, if the order of genes is red stigma, red flower, long.

The value shown by the table is 37. This means that long

lies to the right of red at a locus 46.

R; R L

0: 35. 46.

DIAGRAM IV. Chromosome I, Primula. Rs= red stigma, R= red flower, L = long

t style.

58 59 > 60 61 62

The apparent discrepancy between the values 46 and

37 is due in most part to double crossing over, the effect

of which is always to lower large values disproportion-

ately more than short. When the discrepancy is known,

the amount of double crossing over can be calculated

approximately. Here the amount of double crossing over is

46 — 37

That is, 4.5 per cent. of all the gametes are the result of

double crossing over. A somewhat larger amount of data

from a back cross in which all the individuals are effective

would give by direct experiment a true value for the

amount of double crossing over.

A chromosome diagram should be built up of values

independent of double crossing over. According to our

experience with Drosophila, if there is not more than ten

per cent. of crossing over between two genes, the double

crossing over is negligible. Thus in the first chromosome

in sweet peas, the values obtained from the experiments

are not changed by double crossing over. However, in the

534 THE AMERICAN NATURALIST [Vou. XLVIII

case of the-second chromosome, where the total percentage

of crossing over is about 32, there is probably one or two

per cent. of double crossing over. The diagram of the

second chromosome is in this respect only tentative, and

the plotted position of cretin will be moved a little farther

to the right when the amount of double crossing over

between light and cretin has been found. The value 4.4

for sterile anther light axil is not affected by double

crossing over, since the section of chromosome between

these two loci is so short that a double break would prob-

ably not occur between them at all. The amount of double

crossing over between any two loci can only be found

when there is a gene between them. Thus if a gene

should be found which lies between light and cretin, either

by indirect calculation or, better, by direct experiment,

the amount of double crossing over could be found. The

more genes which can be worked with in the same chromo-

some, the more accurate becomes the diagram.

All the values found for these cases in sweet peas and

primula are based upon such small numbers that they

can be used only as illustrations of the way in which one

would apply to new cases certain principles worked out in

Drosophila. While they serve as examples in line with

these principles, they are entirely inadequate as proof.

A very interesting case of variation in linkage is pre-

sented by some of the families involving chromosome IT

of the sweet pea. In this article I have avoided such data

as far as I could, but it is possible that the order in which

I have aligned these genes will be found to be incorrect

when data upon all three genes in a back cross are ob-

tained. Such data would show, through the phenomenon

of double crossing over, what the order of genes is, even

though variations in the linkage should occur.

CoLUMBIA UNIVERSITY,

May, 1914

THE REDUPLICATION HYPOTHESIS AS

APPLIED TO DROSOPHILA

Dr. A. H. STURTEVANT

COLUMBIA UNIVERSITY

A NUMBER of papers developing the reduplication hy-

pothesis of linkage have recently appeared in the Journal

of Genetics. They are based almost entirely on the

experiments of Gregory (’11) on Primula and of Punnett

(713) on the sweet pea. The data are not entirely satis-

factory because of the relatively small number of genes

involved, and because in most cases the gametic ratios can

be only approximately determined. This is due to the

fact that most of the data concern F, counts, from which

gametic ratios can not be calculated directly. In Gregory’s

best case a much more satisfactory method was followed—

the heterozygous plants were tested, not by mating to

others of their kind, but by crossing with plants recessive

with respect to all the genes involved, which gives the

gametic ratio directly. In this case, however, we have

only a relatively small series of data involving as many as

three pairs of linked genes. It is obvious that from such

data no adequate test of the reduplication hypothesis can

be made.

The phenomena of linkage have been very extensively

studied, by Morgan and others, in the fly Drosophila. In

this animal there are many genes belonging to the same

linkage groups, and these have been studied on a large

scale. In the case of the sex-linked group there is never

any difficulty in calculating the gametice ratio from F,

results, since the F, males from any cross always give it

directly. I have recently published a paper (Sturtevant,

14) giving a complete summary of the published results

obtained from studies of the linkage of these genes. In that

paper I have adopted the chromosome explanation of link-

535

536 THE AMERICAN NATURALIST [Vow XLVIII

age proposed by Morgan (711). Here I shall use the same

data for a test of the reduplication theory. It may be of

value to contrast the two views by making a rigorous

application of them to the same facts. Since the data

concerning the sex-linked group of genes in Drosophila

form the simplest and most extensive series now avail-

able, I shall deal more especially with them. The reader

is referred to my other paper for the detailed data, for

references to original sources, and for a full treatment of

the chromosome hypothesis as applied to these and other

data.

It may be well to give first a brief catalogue of the

sex-linked genes discussed in this paper. The nomen-

clature is that suggested by Morgan (713). This may be

confusing to those accustomed to the ‘‘presence and ab-

sence’’ system, but this should not be a serious objection

here, since a clear conception of the somatic appearance

of the animals discussed is not essential for our present

purpose. The relations would be as clear if hieroglyphics

were used for symbols.

Y is the gene which differentiates the wild ‘‘gray’’

bodied fly from the yellow mutant, y.

V differentiates the wild red-eyed fly from the ver-

milion-eyed mutant, v. :

M differentiates the ‘‘long’’ wing of the wild fly from

that of the miniature-winged mutant, m.

R is another gene affecting the wings. The wild fly

has R, the rudimentary-winged mutant has r.

Br’ occurs in a dominant mutant form having a narrow

eye known as barred. The allelomorph present in the

wild fly is designated br’.

The other characters concerned bear such a relation to

one another that the genes involved are considered as

forming a system of quadruple allelomorphs. The alter-

native to this view is the assumption of complete linkage,

but I have given elsewhere (Sturtevant, ’13) my reasons

for preferring the multiple allelomorph interpretation.

The eye of the wild Drosophila is red in color. A single

No. 573] REDUPLICATION HYPOTHESIS 537

mutant obtained from it had white eyes (Morgan, 710),

and this character proved to be a simple sex-linked reces-

sive. From the white-eyed form arose a fly with eosin

eyes (Morgan, ’12). This new character was found to be

a sex-linked dominant to white, and a sex-linked recessive

to red. Finally, there arose a form with cherry eye color

(Safir, 713). This has the same relation to red and to

white as has eosin. Mated to eosin it gives an inter-

mediate color, which splits up into cherry, intermediate,

and eosin in F,. The nomenclature adopted in this case

is as follows:

Allelomorph present in the red-eyed fly, W.

Allelomorph present in the white-eyed fly, w.

Allelomorph present in the eosin-eyed fly, w°.

Allelomorph present in the cherry-eyed fly, w°.

Trow (713) has suggested the possibility of an asym-

metrical reduplication series, giving a gametic series of

WAB:xAb:yaB:zab, where w need not equal z, nor

x equal y. It should be noted that an actual demonstra-

tion of such a ratio, or of its non-existence, is almost ex-

cluded for the reason that it would be practically impos-

sible to be sure one was not dealing with a case involving

differential viability. However, perhaps the most stri-

king general fact brought out by the study of linkage is

that each pair of linked genes (allelomorphs), considered

separately, follows a perfectly regular Mendelian course.

I think we are, therefore, justified in assuming that the

number of gametes bearing A is always equal to the num-

ber bearing a, and similarly for B and b. Then, in Trow’s

asymmetrical series,

wtae=yte,

w+ y=sr +z.

Hence,

w=2 and t= y.

In all that follows I shall assume that the reduplication

series are always symmetrical. On this assumption it

becomes unnecessary to consider the two halves of the

538 THE AMERICAN NATURALIST [Vou. XLVIII

series separately, and I shall therefore use only two terms

in speaking of gametic ratios. By adding together the two

halves of the series larger numbers are obtained, so that

chance deviations are relatively smaller. Differential

viability is also partially overcome in this way. Of

course on the reduplication theory both terms of the

gametic ratio must be integers, since they represent num-

bers of cells, but nevertheless it has seemed to me more

convenient for purposes of calculation to express them

always in the form n:1. Thus a gametic ration of 3:2

may be written 1.5:1.

It was suggested by Bateson and Punnett (711) that

the intensity of coupling and of repulsion between the

same two pairs of genes may be identical. That this is

substantially the case has been shown again and again in

Drosophila, and has become a truism among those work-

ing on that form. Before presenting data on this point I

wish to bring up another matter on which the same data

have a bearing. Punnett (’13) has said, ‘‘ But where three

[pairs of] factors are concerned . . . the value of the

primary reduplications is evidently altered, and there

would seem to be some process whereby these reduplica-

tions react on one another.’’ Bailey (’14) has suggested

that the nature of this interaction may be such as to cause

the two primary series to be of equal intensity. It may be

categorically stated that there is no interaction effect in

Drosophila. The best data for a test of the relative inten-

sity of coupling and repulsion, and of ‘‘fundamental,’’

‘‘primary’’ and ‘‘secondary’’ reduplication series, in-

volving the same allelomorphic groups, is that furnished

by the relations of the various forms of W (W, w, w°, w°)

to the M pair of allelomorphs (M and m). Table I is a

summary of the data on this case. In computing the

fundamental series I have used only the data from such

of my own experiments as involve only two pairs of genes,

since that from other sources is for the most part made up

of primary series in which the other primary series in-

volved is masked.

No. 573] _ REDUPLICATION HYPOTHESIS 539

TABLE I

FUNDAMENTAL SERIES

Nature of Cross Actual Numbers Gametie Ratios

WM X wm 16+:1

Wm X wM 93: 221 1: 2.4 —

WM X wem 634: 348 1.8 +: 1

Wm X weM n Ge 120 1: 2.4 —

Wm X weM 46l: 855 1: 1.9 —

weM X wm . 4,171: 1,858 2.2 +:1

wem X wM 891: 1,898 1: 2.1 +

weM X wm woe 47 Fe: t

PRIMARY SERIES

; Other Primary

Nature of Cross Actual Numbers Gametic Ratio Series Involved

85 2.1—:1 MBr

wem X wM 69: -122 1:13 — MBr’

WM X wm 5,838: 2,911 2.0 +:1 YW

Wm X wM 1,111: 2,493 1: 2.2 + YW

WM X wm 2,261: 1,011 2.2 +:1 MR

Secondary Series Primary Series

WM X wem so) PIO: 407 1.8—:1 WF, VM

Wm X weM 227: 509 l: 2.2 — WV, VM

It will be noted that in all these cases the gametic ratio

approximates 2:1, or 1:2, according to the nature of the

cross. There are only four cases showing a noticeable

deviation from this value, and of these two involve only

small counts. The most serious is the first. In this case

there is a deviation of 54.3 from the 2:1 ratio, and the stand-

ard error is 16.7[V1/3 x 2/3 X (777 + 470) = + 16.7-].

Since the deviation is slightly over three times the stand-

ard error, it is perhaps significant, especially since there

is at least one other rather large deviation (the second

ratio in Table I). For our present purpose, however, it

is probably not significant, since similar deviations occur

in different experiments of exactly the same type. I have

recorded elsewhere (Sturtevant, ’14) the results of a num-

ber of tests of individual females heterozygous for these

two allelomorphic groups. Taking only those cultures

which produced 100 or more flies, we find the following

results: |

540 THE AMERICAN NATURALIST [Vou. XLVIII

Seven females of the constitution w°mwM gave gametic

ratios ranging from 1.5:1 to 2.7:1, with the modal class

at about 2.0:1.

Seventeen females w°Mwm gave ratios ranging from

1.5:1 to 3.4:1, with a single individual at 4.2:1. The

modal class was at about 2.2:1. -

It seems highly probable that all these deviations from

a 2:1 ratio, not due to insufficient numbers, may be satis-

factorily explained on the basis of differential viability,

which is known to occur here (for a discussion of the

vagaries of differential viability see Bridges and Sturte-

vant, 14). I do not wish to be understood as arguing

that the gametic ratio for any two pairs of genes is abso-

lutely constant, but only that it is in most cases uninflu-

enced by the way in which the genes are combined and by

heterozygosis for other genes. That it may sometimes

show marked differences is now well established. I have

myself studied two eases of this sort, and I have good

evidence (not yet published in detail) that there are defi-

nite genes which cause great differences in the gametic

ratios for whole linkage groups. In one case this gene

itself shows linkage to those in the group it affects. But

even here the intensity of coupling and of repulsion is

affected alike, and it makes no difference how few or how

many genes a fly is heterozygous for; the linkage is strong

or weak according to the form of the linkage-affecting

gene which the fly happens to carry. In each of these

cases I have been able to obtain about the same extreme

values both for coupling and for repulsion.

In what follows I shall assume that the intensity of the

reduplication series is not affected by the way in which

the genes are introduced, nor by the number of linked

genes involved in the cross. The obvious corollary of

this is that reduplication occurs even in homozygous indi-

viduals, and that the nature of the series of divisions is

in general independent of the constitution of the indi-

vidual. This conclusion is directly opposed to the point

of view expressed more especially by Punnett, in the

No. 573] REDUPLICATION HYPOTHESIS 541

passage quoted above and elsewhere. If reduplication

occurs at all it is the same in the wild fly as in the most

complex linkage experiment we have yet carried out.

If it is assumed that the intensity of coupling and re-

pulsion is identical, it becomes unnecessary to consider

them separately. I shall therefore lump together all the

data involving the same groups of allelomorphs, regard-

less of how they were put into the cross. When three

pairs of genes are involved there are eight possible com-

binations of them in F., but only four if we add together

the two halves of the reduplication diagram. There are

the two original combinations, which I shall designate

ABC. Then there are three combinations derived from

each of these by a shifting of one gene, which I shall

designate ABc, AbC and aBC, the small letters referring

to those pairs which have been shifted. Thus, to take an

imaginary case, if we cross LMn by lmN, the gametes

produced by the F, individuals will be classified as

follows:

ABC ABc AbC aBC

LMn LMN Lmn lMn

ImN lmn IMN LmN

In the following tables I shall reduce all data to this

form. In each case the genes will be arranged so that

AB and BC will be the primary reduplication series.

Table II contains such a summary of all the crosses in-

volving three pairs of sex-linked genes. Table III shows

the gametie ratios derived from these data, and also the

values for the secondary series calculated on the basis of

Trow’s ‘‘special’’ hypothesis. For the sake of brevity

only one term is used: a gametic ratio of 3:1 is written 3;

a ratio of 3:2 becomes 1.5, ete. With the simplifications

introduced here Trow’s formula becomes

ac — (48 x BO) +1

AP Eee.

1 As was pointed out by Punnett (’13), in a system of three reduplica-

tion series the one with the lowest intensity is to be regarded as the second-

542 THE AMERICAN NATURALIST [Vou. XLVII

TABLE II

Allelomorphic Groups ABC ABc | AbC aBC

BBO) Cp PE ORT CATES Se Sar 8,212 4,013 9 119

BOW fe i a ea oe 278 60 0

PVM SEU Ree 1,082 58 y 22 665

PE kee a SA 315 138 55 196

TIEF Sele Pee eS 93 34 10 54

WEM, ie ais 194 1 102

PUER ee. ee eet 1,726 535 139 872

EMED ee eee 20 73 25 129

TABLE III

Gametic ratios

Experiment Observed Calculated

AB BC AC AC

> A” Sree ee 6 as 95.5 2.0— 2.0+

FNE oe ee oe 438.0 1.74 1.72 1.74

VY Mero t.7 22.0 1b 1.6

FFR abet SiMe Segall 1.8 gE 1.3

PV Bi ee oc rae 2.0 3.4 1.3 1.4

AES ee eo 2.0 24.7 LZ 1.9

FNE eee 2.9 ; 1.3 1.6

FE ace oe ee ee 1.9 3.6 T2 1.4

It will be seen that in every case the calculated value

for the secondary reduplication is higher than the ob-

served value. The same relation comes out in two experi-

ments which I have done involving genes of another

group in Drosophila (see Table VIII, Sturtevant, ’14).

Punnett’s case is so involved that calculations accurate

enough for our present purpose can not be made. In

Gregory’s experiment one of the genes (M) could not be

followed in all the plants because masked by another gene.

We are not given the data for S and G in those plants in

which M was classified separately from those in which it

was not. The data are therefore not available for exact

calculations, since the numbers are too small to overcome

chance deviations. The data for my own two experi-

ments appear in Table IV.

The same relation comes out more strikingly in another

way. If we let m equal the intensity of the AB series and

n that of the BC series, then on Trow’s special hypothesis

No. 573]

REDUPLICATION HYPOTHESIS

543

the four kinds of gametes should occur in the following

proportions:

ABC — mn

ABc—m

aBC'—n

AbC —1

TABLE IV

Observed Calculated

Experiment

AB cll | AC AC

Big rey 3.4 11.6 | 2.4 2.7

CCR E N A 2.5 2:1 1.0 1.4

That is, 1/(m + 1) of the gametes should have A and B

interchanged. Of these, 1/(n +1) should have B and C

also interchanged. If N represents the total number of

gametes, then the size of the AbC class should be repre-

sented by the expression

AbC =

(m+ 1)(n+1)°

Table V shows the relation between the size of this class

as observed and as thus calculated, in the ten experiments.

Allelomorphiec

or eee

eee

TABLE V

ABC

“Observed Calculated

Era Ce 9 42

ERTE 0 0

pee eee a 22 30

FRN E 55 69

R E E 10 15

rem er l 4

E Rease 139 208

PAE A A 25 34

AA E 2 7

Dic ce a 12 20

Thus it appears that in all ten experiments Trow’s

formula gives values for the AC series and for the AbC

544 THE AMERICAN NATURALIST [Vou. XLVIII

term which are too large. Moreover, this feature appears

in a more complex cross which I have carried out, in-

volving four pairs of linked genes (YWV WM), and in each

separate part of all these experiments, regardless of how

the crosses were made. It may, then, be taken as a con-

stant relation. It can only mean that there is some rela-

tion between A and C besides that resulting from second-

ary reduplication. In other words, to use Bailey’s terms,

Trow’s ‘‘special’’ hypothesis is not valid.

Let us then examine what Bailey calls Trow’s ‘‘gen-

eral” hypothesis. Suppose the primary series to be of

the following values:

AB ==1:1,

BO een it,

AC 2 nz i.

Trow’s general formula for calculating what should be

the observed value of-the AC series is

The special formula is derived from this by assuming

n = 1, when the formula becomes

Be nae:

pgm.

Since this always gives a value which is too large, it

follows that n is always less than one. This means that

the AC primary series is reversed—that the combinations

present in the parents tend to be reproduced in fewer

numbers than the new combinations. I have worked this

out for the case of BCvSp (see Table IV), and find the

primary series there to be 0.6:1, though the observed

series is 1.0. The ‘“‘fundamental”’ AC series has been

. obtained for most of the eases in Table III, and has

always been found to be of the usual form (i. e., n:1,

No. 573] REDUPLICATION HYPOTHESIS 545

where n >1). (See Table I, Sturtevant, ’14.) In fact,

as stated above, the fundamental series always approxi-

mates the secondary (observed) series.

There are two hypotheses as to the mechanics of re-

duplication series where more than two pairs of genes are

involved. The first was suggested by Bateson and Pun-

nett (711), and consists in the assumption that when three

pairs are involved eight cells are formed by three succes-

sive divisions, each of which segregates one pair of genes.

The eight cells then represent the eight possible kinds of

gametes, and are supposed to reduplicate independently

until the proper proportions are reached. Bailey sup-

poses that if it be shown that two primary series do not

interact on each other this scheme will be more likely to be

correct than will Trow’s, which I shall discuss next. It

seems to me, however, that this hypothesis begs the ques-

tion. It is derived entirely by working backwards from

the observed results; it affords no basis for predictions;

and it does not offer a simple mechanical explanation of

any of the observed results. For pragmatic reasons I

believe we should adopt it only as a last resort. -

Trow supposes that two cell divisions occur, segregating

two pairs of genes. The four resulting cells then go

through with their reduplication, which is a primary one.

When this is finished there occur divisions which segre-

gate the other pair, and the other primary reduplication

is carried out. On Trow’s general hypothesis, which I

have tried to show is the only one which can hold, it is

supposed that the second series of reduplications is

affected by both of the first two pairs of genes. C is re-

duplicating more if with B than if with b, less if with A

than if with a. This scheme of Trow’s has one great

advantage in that it accounts for the fact that the class

which I have called AbC is always the smallest one.

Reference to Trow’s calculations will show that this rela-

tion should always occur, and Table IT shows that it does

occur. .On the octant scheme there is no explanation of

this relation—we oie have to assume that it does occur

somehow.

546 THE AMERICAN NATURALIST. [Vou XLVIII

It will be noted that several of the gametic ratios in-

volved here closely approach 2:1. YV, YM, WV and WM

are the most conspicuous examples. It may seem that

such a simple ratio is due to a very simple reduplication

series, but I do not think such an assumption can be suc-

cessfully maintained. The tables given above show that

YM and WM have approximately this same value when

they appear as secondary series, and the data for the

combination YWVM show the same thing for YV (see

Sturtevant, 714).

If, as I have maintained above, the same series of redu-

plications must occur in all flies, whether we can follow it

or not, then it follows that in these three cases the 2:1

ratio is never due to a simple series, but always to a long

and complicated one, since in all three one of the primary

series is of high intensity.

It was pointed out by Trow that the intensities of the

reduplication series afford a method of calculating the

number of cell divisions necessary to complete the series.

If we assume that approximately the same series is

occurring both in homozygous and in heterozygous flies,

we have the following series in Drosophila as a basis for

such calculations.

Sex-linked Group

EW 901

Wy =: 24

VM== 318

MR =: 30

RB x217

Second Group

BVg= 3.6

VgCv= 104

CuSp= 2.8

SpBa= 10+

Third Group

PEb=100+

No. 573] REDUPLICATION HYPOTHESIS 547

All of these series must be considered as either primary

or secondary and therefore involving primaries of higher

intensity. In fact there is unpublished evidence that

many of them can not be simple primaries. A num-

ber of series of very high intensity are known, and will

appear in future publications. Therefore all the calcula-

tions that follow give results which are far too small.

According to Trow, the minimal number of successive

cell divisions required to complete the series is given by

the expression mnp --- where m, n, p, ete., are the larger

terms of the primary series involved. In the present case

the value of that expression is something over 76,000,-

000,000. However, Trow’s formula seems to be wrong.

If a be the number of cell divisions required to produce

m cells, then 2*—m. If this expression gives a value of

a which is not an integer, then the next higher whole

number is to be taken. In the case of the first series two

divisions are necessary to segregate the genes, and in the

following series one is required. The number of succes-

sive cell divisions required then is (a+1)+(b+1)

+(e+1)+----+1, where b, c, ete., bear the same rela-

tion to n, p, etc., that a does to m. In the case of Droso-

phila the value of this expression is 56. As pointed out,

however, this value is certainly far too small.

The total number of cells required is given by the

aches a ee t d --- t2mn--- + 2mp

+ 2m - r -+ 2p: +--+ +2mnp + 2mn + 2mp

Eat ete REWI

This gives a value considerably above 600,000,000,000—a

manifest absurdity. However, it is not necessary that all

these cells should be produced, since the ratios would not

be appreciably affected by some lines becoming crowded

out. It is necessary, on the other hand, that all of the

series shall be completed in every line which does live,

since every female Drosophila, which is of the proper

constitution to be tested, shows n for every pair of

genes tested.

2 The results discussed here deal only with the linkage in female flies.

548 THE AMERICAN NATURALIST [Vou. XLVIII

Thus we are forced to assume an enormously complex

series of cell divisions, many of them differential, pro-

ceeding with mathematical regularity and precision, but in

a manner for which direct observation furnishes no basis.

It seems to me that it is not desirable to assume such

a complex series of events unless we have extremely

strong reasons for doing so. I can see no sound reason

for adopting the reduplication hypothesis. It apparently

rests on two discredited hypotheses: somatic segregation,

and the occurrence of members of the 3:1, 7:1, 15:1, etc.,

series of gametic ratios in more cases than would be ex-

pected from a chance distribution.

The chief advantage of the chromosome hypothesis of

linkage which has been proposed by Morgan (711), an

which I have followed elsewhere, seems to me to be its

simplicity. In addition it appeals to a known mechanism,

and a mechanism toward which the experiments of Boveri,

Herbst, Baltzer and others point as the correct one. It

explains everything that any of the forms of the redupli-

cation hypothesis does, and in addition offers a simple

mechanical explanation of the fact that ‘‘secondary

series’’ are always smaller than Trow’s ‘‘special hypoth-

esis’’ calls for them to be. On the reduplication hypoth-

esis this fact must merely be accepted, for, I think, it

can not be explained.

COLUMBIA UNIVERSITY,

LITERATURE CITED

Bailey, P. G.

714. Primary and Secondary Reduplication Series. Jour. Genet., II.

Bateson, W., and R. C. Punnett.

11, On Gametic Series Involving Reduplication of Certain Terms.

Jour. Genet., I.

cy a C. B., and A. H. Sturtevant,

A New Gene in the Second Chromosome of Drosophila, ete. Biol.

Bull., XXVI.

Sgor, A E

11. On Gametie Coupling and Repulsion in Primula sinensis. Proc.

Royal Soc., 84. B.

Morgan, T. H.

710. Sex Limited Inheritance in Drosophila, Science, XXXII.

No. 573] REDUPLICATION HYPOTHESIS 549

711. An Attempt to Analyze the Constitution of the Chromosomes on

the Basis of Sex-limited Inheritance in Drosophila. Jour. Exp.

712. Further Experiments with Mutations in Eye-color in Drosophila.

. Acad. Nat. Sci. Philadelphia

713. Atti and Unit Characters in Mendelian Heredity. AMER. NAT.,

XLVII.

Punnett, R. C.

13. Reduplication Series in Sweet Peas. Jour. Genet., III.

Safir, S. R.

’13. A New Eye-color Mutation in Drosophila. Biol. Bull, XXV.

Sturtevant, A.

”19: The Bitwlinyan Rabbit Case, with Some neers see on Mul-

tiple Allelomorphs. AMER. Nat., XLVII.

714, The Behavior of the Cheaniobniliee as Studied Through Linkage.

Zeits. f. ind. Abst.- u. Vererb.-Lehre

Trow, A. H.

713. Forms of Reduplication—Primary and Secondary. Jour.

Genet., II

PATTERN DEVELOPMENT IN MAMMALS AND

BIRDS. Ill

GLOVER M. ALLEN

Boston Museum or NATURAL HISTORY

PARTIAL ALBINISM IN Wp BIRDS

In birds under natural conditions of wild life partial

albinism is fairly common. Lists of species of which

- albinistic specimens are known were published by Ruth-

ven Deane (1876, 1880) some years ago, and by others.

Scattered instances are in all the bird journals or maga-

zines of general natural history. In most cases in which

the white markings are clearly defined against the pig-

mented parts of the plumage, these may be referred to

their particuiar primary breaks between the several

areas of pigment formation. In other cases the pigment

reduction is of the diffuse type, tending to form spots.

A few instances follow in which the several primary

patches have been observed in wild birds, either as acci-

dental marks or as permanent parts of the pattern.

The Crown Patch.—In 1908, a pair of robins nested

near Lowell Park, Cambridge, one of which showed a

partial separation of the crown patch, through the pres-

ence of a white band, as broad as the eye’s diameter,

passing from one eye around the back of the head to the

other eye. In the Wilson Bulletin (Vol. 2, p. 45, 1908)

W. E. Saunders records the capture of two robins each

with a white collar about the neck, probably marking the

separation of the neek patches from the shoulder patches.

Coues (1878) records a brood of black robins at St.

John’s, N. B., one of which was kept in captivity by the

late G. A. Boardman. In September, after moulting, it

was still pure black, except for white wings and tail,

which seems to indicate an areal restriction of the

shoulder and rump patches, though the pigment, where

550

No. 573] PATTERN DEVELOPMENT 551

produced, must have been superabundant. Ward (1908)

has described a case of a black robin becoming albinistic

and reviews a number of such cases. The ability of the

same feather follicles in different moults to produce

feathers with different sorts or amounts of pigment is

thus evidenced and has lately been carefully studied by

Pearl and Boring (1914) in the hen.

In addition to the case of the robin above mentioned,

the white line marking off the crown patch from the ear

patches is sometimes found abnormally in other birds.

Thus Sweet (1907) records two slate-colored juncos

(Junco hyemalis) taken in March, 1903, at Avon, Maine,

in which there was a white line above the eye, and the

black throat patch was absent, owing no doubt to the

ventral restriction of the neck patches, as often seen, for

example in pigeons. Maynard! figures the head of a

young female black-poll warbler (Dendroica striata) in

autumn, showing an inclination to assume a white super-

ciliary stripe. I am convinced that this mark so common

in many birds, is merely a development of the primary

break marking off the crown patch from the ear patches

so that it has become a permanent part of the pattern.

The failure of the crown patch to develop at all, as is

sometimes the case in the domestic pigeon, results in a

white-crowned bird. In the West Indian Columba leuco-

cephala, exactly this modification has taken place and the

entire top of the head is permanently white. The same

condition is found in sundry other genera, including a

humming bird, a heron, and others. It would be inter-

esting to discover by experiment if it were not easier to

produce a definite white marking through selecting for

the non-development of a certain patch or patches, than

to try to restrict a certain pigment patch to definite

bounds as in the experiments of Dr. MacCurdy and Pro-

fessor Castle (1907).

The crown patch as a separate unit in pigmentation, is

often of a different hue from the surrounding patches.

1‘‘ Birds of E. North America,’’ 1896, p. 585.

552 THE AMERICAN NATURALIST [VoL. XLVIII

Thus in the case of the terns, the black-eappéd chickadee,

the black-crowned night heron, and other birds, a black

crown patch is noticeably marked off.

The Ear Patches.—The ear patches in birds are small,

yet often specially marked out by white boundaries, which

are permanent parts of the pattern. Yet there is no

doubt but that the acquisition of such white boundaries

is a derived character. It is common for the ear patches

to be colored differently from the surrounding parts,

forming as in some species of tanagers a black auricular

area contrasted with the blue of the head and neck. Of

particular interest in the present connection, however,

are those cases in which a pigmented ear patch is more

or less clearly marked off by a white line above it or

below, or both. The superciliary stripe, so common in

birds, is of course a development of a primary break

above the patch, separating it from the crown patch.

Where the stripe is narrow it is hard to say which patch

has begun to be restricted, though often no doubt both

are more or less involved. Thus the Garganey teal has

a very wide white eye stripe, and in other species of

ducks the whole side of the head may be white, indicating

much greater restriction of pigment formation in con-

tiguous patches. A beautiful example of the develop-

ment of a white stripe at the lower border of the ear

patches is found in the Inca tern, in which a line of white

feathers runs from just above the gape along the lower

side of the auricular patch and separates it from the

dark throat. But not only is the white line developed,

but the feathers composing it are specially elongated and

recurved, as if the mark were one of particular decora-

tiveness. The dark ear patch is noticeable in many

hawks, separated above and below by white areas, as in

the duck hawk and the osprey, though differing in the

size of the white areas.

An instance in which the white line separating the

` crown patch from the ear patch, is even now in course of

becoming established as part of the permanent pattern,

No. 573] PATTERN DEVELOPMENT 553

is afforded by the common guillemot (Uria troille) of the

northern Atlantic. The other related species of the genus

have the head and neck uniformly pigmented, but in U.

troille a considerable proportion of specimens show a

narrow white eyebrow and a postorbital line, in exactly

the situation of the stripe in the albino robin previously

noted, though not so broad nor so extended. Birds so

marked were formerly considered a distinct species—the

ringed murre (Uria ‘‘ringvia’’)—or perhaps a plumage

of U. troille, and much effort has been made to determine

their exact status. Both plumages are found in the same

colonies and the two sorts of birds are known to have

mated together (Müller, 1862). Verrill estimated that

about 40 per cent. of the nesting birds he saw on the

Labrador coast were of this variety, but this is probably

a rather high estimate. I am convinced that the true

explanation of this puzzling variation is that incipient

albinism has gained a foothold, of such nature that areal

restriction of the ear or crown patches is developing, so

that a white line results between them. In the crested

auklet (Æthia) a member of the same family, of the

Pacific Coast, such a line has become fixed so that it

now forms a characteristic mark of the species. In the

case of the ‘‘ringed murre,’’ I should expect to see the

eye stripe in the young as well as in the adult stage of

those individuals which are to have the mark—in other

words it is a permanent trait. No doubt the heredity of

this white stripe is of some definite sort, and if a reces-

sive character, it may nevertheless in time become com-

mon to- an increasing number of birds, as this is a

colonial species and the possibility of inbreeding is thus

increased.

The Neck Patches.—In birds the neck patches extend

forward from the breast to meet the crown patch at the

occiput and the ear patches at the sides of the head,

thence ventrally to include the throat and chin. A study

of albinistie pigeons, as previously noted, indicates that

the neck patches are two separate areas of pigmentation,

554 THE AMERICAN NATURALIST [Vou. XLVIII

one on each half of the part covered, with an ultimate

center at the base of the neck, usually the last spot to re-

main when the area is much reduced.

In albinistic individuals, that is, those in which restric-

tion of the pigment areas has taken place, the neck

patches are usually first reduced at the upper part of the

throat, so that a white patch appears from the chin to

upper throat, as commonly seen in street pigeons; in

others, however, the restriction may be at the posterior

end of the patch, so that a white ring develops at the

base of the neck.

In many birds the neck patches have been much devel-

oped as characteristic pigmented areas. Two general

categories may be here distinguished: (1) those in which

the neck is rather uniformly colored all about, and (2)

those in which the ventral portion is heavily pigmented

and the dorsal portion much less so. In the latter belong

such birds as the black-capped ehickadee (Penthestes

atricapillus) with a black throat but a pale neck. So,

too, the golden-winged warbler (Vermivora chrysop-

tera). In this latter category it is probable that a sec-

ond factor is present, comparable to that producing a

centrifugal type of pigmentation in mammals, such for

example as in the Himalayan breed of rabbit, which has

the end of the nose and the feet black-pigmented, contrary

to the usual rule of normal areal reduction where the

extremities are the first to become white. That this is

a separate category from a physiological standpoint is

indicated by its behavior in heredity as worked out so

admirably by Faxon (1913) in the case of the Brewster’s

warbler. He discovered that the black throat as present

in the golden-winged warbler is recessive in the cross

with a related species, the blue-winged warbler (Ver-

mivora pinus), a yellow-throated bird. The offspring

of this cross have white throats,—the so-called V. leuco-

bronchialis. The black throat patch may be evidence of

‘‘centrifugal’’ pigmentation as defined farther on (p. 53).

The essential bilaterality of such a throat patch is

No. 573] PATTERN DEVELOPMENT 555

further shown by the fact that one half only may be pres-

ent as in the golden-winged warbler recorded by Dr. C.

W. Townsend (1908).

The first category, in which the neck is uniformly pig-

mented is illustrated by many of the duck tribe, and

probably involves the normal primary patches only.

The primary patches are usually restricted first antero-

ventrally producing a white throat. Often this is carried

dorsally so as to form a white ring around the upper

part of the neck by the separation of the neck patch from

the crown and the ear patches. Again, if the neck patches

are restricted posteriorly a white ring is formed at the

base of the neck, a common permanent character in many

species. The peculiar little goose-like bird— Nettapus, of

India—has developed this type of marking so that its

white neck is encircled by a narrow black ring, and the

Labrador duck (Camptorhynchus) has a nearly similar

mark (Fig. 57). Other ducks, e. g., the mallard, have the

white ring at the base of the neck, only.

In an interesting paper on the geese occurring in Cali-

fornia, Swarth (1913) has pointed out that in the cack-

ling goose (Branta c. minina) there is much variation in

the amount of white on the head and neck. Figs. 58 to

62 are traced from a series of photographs illustrating

this paper and show the throats of five specimens. The

wide range of variation in these specimens indicates to

my mind that this goose is in process of reducing the

neck patches, and thereby developing a white collar, such

as is present in the mallard, and perhaps also a white

throat. The usual condition seen in Branta canadensis

and in so-called normal specimens of B. c. minina is seen

in Fig. 58. The white cheeks have been developed long

ago in the history of the species, in part perhaps by the

depigmentation of the ear patches. Now a second change

is taking place in one of its subspecies. Thus in Figs. 59,

6l and 62, the neck patches have been reduced poste-

riorly, a varying amount in each case. In Figs. 60, 61 and

62 these patches have been restricted anteriorly pro-

556 THE AMERICAN NATURALIST [Vou. XLVIII

ducing a white throat, and as sometimes in the pigeon,

imperfectly, so that a little island of pigment is cut off

just at the chin. It is also obvious from these figures,

that reduction may take place either at one end or the

other, or at both ends in different individuals. The ulti-

mate development of this line of reduction will produce

i: a a | bo Gr b3

Fics. 58-62. VARIATIONS IN THE DEVELOPMENT OF THE NECK PATCHES IN THE

CACKLING Goose (after Swarth).

the narrow black collar seen in Nettapus previously men-

tioned. It is worth noting also that in this goose the

limits of the neck patch are by their black color sharply

defined posteriorly from the gray of the breast which is

pigmented from the shoulder patches.

The Shoulder Patches.—The shoulder patches appear

to center near the base of the wing, and in reduction

produce white remiges, such as appear in a domesticated

race of guinea fowl, as well as a white breast. The

domesticated guinea fowl often shows this white area in '

the midline of the breast as the pigment areas fail to

spread ventrally. In the normal pattern of wild birds,

however, white wings are seldom seen except among cer-

tain sea birds. White wing patches are often developed,

but these are frequently only bars on pigmented feathers

as in the goat-suckers. Probably among small land birds

much white in the large wing feathers is a disadvantage,

No. 573] PATTERN DEVELOPMENT 557

and so not much developed. It is noticeable that white

patches in the wing are often of such a nature that they

are concealed through the folding of the wings when the

bird is at rest. This accords with my belief that while

in flight the bird is unavoidably conspicuous by reason

of its motion, and that white patches showing at such

times add little or nothing to the disadvantage. In the

hairy and the downy woodpeckers (Dryobates), a white

stripe down the back is developed as part of the pattern,

and no doubt as in many mammals, marks the separation

between the pigment areas of opposite sides. Centrifugal

pigmentation is seen in some species as the kittiwake in

which the outer primaries are black.

The side patches are commonly continuous with those

of the shoulders, and when ventrally restricted, give a

white abdomen. Their median separation dorsally, is

seen in the hairy and downy woodpeckers as above noted.

I have not studied any special developments of these

areas, and they are commonly small.

The Rump Patches.—In birds as in. mammals the two

rump patches pigment the posterior extremity of the

body. Their ultimate centers are dorsal and so close to-

gether that it is much less common for them to be sepa-

rated medially than to be restricted laterally. With a

slight areal reduction, a separation takes place between

them and the side patches dorsally, so that a white area

on the rump results. Often this white area represents

doubtless a slight restriction of both sets of pigment

patches which by drawing farther apart increase the

white area along the lower part of the back. In the

domestic pigeon much variation may be found, from a

condition in which the lower back is wholly pigmented

to one in which it is mostly white. The primary break

which causes this white patch has been much developed

in many groups of birds as a particular mark in the

pattern. In many species it is simply of a paler hue than

the surrounding parts as in the yellow-rumped warbler

(Dendroica coronata) or the pine grosbeak (Pinicola).

558 THE AMERICAN NATURALIST (Vor. XLVII

In others the tendency to albinism thus expressed has

gone farther so that a pigmentless spot is formed. This

white rump patch is present in many unrelated groups

of birds in which it has independently arisen through

parallel development. Thus it is seen in many of the

smaller petrels, in the palm swift, the flicker woodpecker,

the white-rumped and other sandpipers, the white-

rumped shrike, the European house martin and others.

The tail feathers are pigmented by these patches, and

among various species show many steps in the process

of pigment reduction. As in the domestic pigeon, occa-

sional albinistic individuals show white outer tail

feathers, in accordance with the rule that the first pig-

ment reduction takes place at those parts of the primary

areas that are farthest removed from the pigment centers.

I have seen a white outer tail feather in wild specimens

of song sparrows and Lincoln’s sparrow and it is occa-

sional in other species. In others again this mark has

become developed and fixed as a species character. Thus

in the bay-winged bunting (Powcetes gramineus) there

is a single white outer feather on each side, in the junco

(Junco hyemalis) there are two. A white central tail

feather is much rarer, but a pure white tail is found occa-

sionally as in the hummingbird, Leucuria phalerata, the

bald eagle and certain gulls, due to the permanent reduc-

tion of the pigment area of the rump at this extremity.

I once examined an albino ruffed grouse (Bonasa) which

was entirely white except for a single feather among the

upper tail coverts at the left side of the rump. This

blemish in the otherwise pure white bird seemed inexpli-

eable to those who examined it with me, but it merely

represents the last remnant of the left-hand rump patch,

still persisting though all the other pigment centers were

inactive.

It is very interesting that the white rump mark, so

commonly found in unrelated ‘groups of birds, is one

which is conspicuous in flight only, and the same is true

of many of the white tail marks, such as outer white

No. 573] PATTERN DEVELOPMENT 559

feathers that disappear when the tail is shut. This points

to the conclusion that the development of a white mark

which is ever conspicuous is allowed in nature in such

cases only where it may be no detriment to the species

through rendering it too conspicuous by contrast. Thus

the bald eagle or the black-backed gull have nothing to

fear from such a banner mark. For small weak-flying

birds, however, the case may well be different. Yet even

these often show much white and I believe that it would

be possible for a species in its phylogeny to develop more

and more white if at the same time its habits of watchful-

ness or other actions developed equally to counteract any

disadvantageous result that might accompany the in-

crease. No doubt also a psychic factor is involved, com-

parable to what among ourselves we call ‘‘fashion.’’

Thus a change in action or dress which departs too far

from the accustomed appearance is apt to be disliked at

first, though in time it may if persisted in, be tolerated

and at length accepted. In the development of white

markings, for example in the feathers of the tail, it

seems likely that a series of small steps must have been

made rather than too great and sudden changes. So in

the rock pigeon the white of the tail is limited to the outer

vane of the outer tail feather. In the turtle dove the

outer vane of the outer feather, and the entire tips of the

four outer feathers are white. The next step would be

to develop an entirely white outer feather and then two

(as in the passenger pigeon) and so on. In the sparrows

similar steps are shown by the lark sparrow (Chon-

destes) in which the tips only of the outer feathers are

white, the bay-winged bunting which has practically all

the outer feather white, and a little of the tip of the

second, the junco with two outer feathers and part of a

third white. No doubt steps such as these must have been

passed through by many white-tailed species.

It is difficult to say how disagreeable to their normally

colored neighbors, albino birds may be. I have seen an

albino robin in the fall of the year with a flock of other

560 THE AMERICAN NATURALIST [Vou. XLVIII

robins and a white-spotted bee-eater with a flock of its

brethren, in both cases wholly at peace. This of course

was in flocking time when the social spirit is strong. The

song sparrow (Melospiza) with white outer tail feathers,

previously mentioned, was attacked and driven off by

another song sparrow. In the Journal of the Maine

Ornithological Society (Vol. 6, p. 48, 1904), C. H. Clark

writes of a pair of albino eave swallows (Petrochelidon

lunifrons), at Lubee, Maine,

among a large colony of the common ones who seemed greatly annoyed

at the albinos’ presence and fought with them until they finally killed

one ... or rather injured it so badly that it died soon after.

I also have a note of a white robin at Montclair, N. J.,

which in early July, 1909, was seen to be much beaten and

driven about by another robin and eventually flew at full

speed against a tree and was killed.

CENTRIFUGAL COLORATION '

In addition to the primary pigment patches which I

have discussed at some length, and the speckled condition

r ‘‘Einglish’’ marking, there is, as I have already inti-

mated, a third condition in which pigment is developed

at the extremities or points. Itmay be called a centrifugal

type and is almost the reverse of the centripetal or ‘‘pri-

mary-patch’’ class.

The two latter types of pigmentation may both be

found in the same individual, but ordinarily this is not

evident except in cases where the primary patches are

somewhat restricted in area. It then may become appar-

ent that pigment is present at exactly those points where,

in the centripetal type of coloring, it is first to be lacking.

Moreover it persists strongly, even though the primary

areas are much reduced or largely absent. Curiously

this sort of pigment seems almost always to be black.

Apparently centrifugal pigmentation does not occur in

all species. I have never seen any trace of it in dogs.

In the house cat it is frequent, however. Thus in Figs.

No. 573] PATTERN DEVELOPMENT 561

18 and 19 it appears at the end of the tail. In the former

figure the sacral patches are much reduced, though pres-

ent, and together spread nearly half the length of the tail.

The terminal half, or less, of the tail, however, is dark-

pigmented, and a break occurs between the two sorts of

markings, due to the failure of the centripetal patch to

spread so as to unite with the centrifugal area. In Fig.

19 the sacral patches have wholly failed to develop but

the centrifugal patch still covers the distal half of the tail.

Possibly the dark heel marks in Fig. 16 are patches devel-

oped in the same way. In the house cat, a dark or

‘*smutty’’ nose is often present in contrast to an other-

wise. white face, or with the ear patches only slightly

reduced. In the breed of rabbits known as ‘‘ Himalayan,”

the centrifugal pigmentation remains, though the centri-

petal markings have di ed, so that it is pure white

except for the black nose, ear tips and toes. No doubt,

however, it would be possible for the two types of pig-

mentation to appear in a single individual. This is sug-

gestive of the winter phase of the Arctic hares, in which

the black ear tips contrast strongly with the otherwise

white pelage. The physiology of the process whereby

certain animals acquire a white winter coat is not yet

fully worked ‘out. It is curious that in occasional melan-

istic individuals of the eastern varying hare, the black

color is retained throughout the winter, instead of being

replaced by white—again a persistence of black pigment.

In dappled gray horses a black patch sometimes appears

on the bridge of the muzzle, usually the first place to show

white in the restriction of centipetal pigmentation. The

feet may also be black. Among certain antelopes a black

muzzle mark is similarly present, and in Hunter’s ante-

lope (Damaliscus hunteri) a white border partly sur-

rounds such a mark, This, I believe, is due to a slight

‘restriction of the ear patches, sufficient to prevent them

from reaching the muzzle, and of about the same nature

as seen in the blesbok (Damaliscus albifrons) in which,

through the absence of a centrifugal nose patch, the entire —

a, a

562 THE AMERICAN NATURALIST [Vou. XLVIII

front of the muzzle is white. The white chevron on the

muzzle of several antelope (Strepsiceros, Taurotragus)

is probably the result of a similar restriction of ear

patches combined with a centrifugal nose patch, leaving

a white line between. The black dorsal stripe seen in

many mammals and the black tail tip are probably mani-

festations of centrifugal pigmentation. The latter mark

is common in stoats (Mustela) and among those that

change to a white coat in winter, as the ermine, the tail

tip still remains black. In sundry other genera, as

Genetta, a black tail tip is part of the normal pattern.

In their paper on albinistic negroes, Simpson and

Castle (1913) published some highly interesting photo-

graphs of ‘‘piebald’’ individuals. In four persons of one

negro family the hair over the median part of the head

from the occiput to forehead is pure white, as though due

to a restriction of the aural pigment patches. In addi-

tion, more or less of the median area of the back, as well

as the hands (including much of the forearms) and feet

(including the lower part of the ankle) are pigmented.

These latter areas may represent centrifugal pigmenta-

tion, but it should be noted that this is present in the

dermis. Possibly there is a close relation between dermal

pigment and that produced in the centrifugal style of

pigmentation.

Among birds, the black of the outer tail feathers of the

ptarmigan (Lagopus) may be comparable. A black area

is also sometimes present on the middle of the throat, or

as in certain gulls the outer primaries may be black.

This form of pigmentation is not found universally and

the conditions governing its appearance are unknown,

though its heredity in the ‘‘Himalayan’’ rabbit has been

somewhat studied by Professor Castle.

SuMMARY

The principal points of this paper may be summed up

as follows: .

1. In mammals and birds that normally are com-

No..573] PATTERN DEVELOPMENT 563

pletely pigmented, there are certain definite points of

the body from which as centers the tendency to develop

pigment in the epidermal structures may become less

and less. Outward from each of these centers pigment

formation spreads to include very definite areas which in

wholly pigmented animals overlap slightly at their

borders or are at least contiguous.

2. A reduction in the area covered by any of these

primary patches results in a white mark at the line of

junction of two contiguous color patches, where no pig-

ment is produced. These white marks between the pri-

mary patches are spoken of as primary breaks.

3. Through a study of the breaks in pied individuals

of domesticated species of mammals and birds, the

boundaries of the primary patches have been determined.

These are homologous in the two groups and subject to

- a certain amount of variation in different types. They

are: a median crown patch unpaired, and five paired

patches on the opposite sides of the body, which are

named from the general areas they co ver, the ear, neck,

shoulder, side and rump patches. Their limits are more

precisely defined under the different species treated.

4. These patches are physiologically independent of

each other and may be differently colored in the same

individual.

5. Pied patterns among many wild species have been

brought about through the areal reduction of these pig-

ment patches in a definite way so that the white markings

resulting as breaks between the reduced patches have

become fixed and form a permanent part of the normal

pattern.

6. In several wild species this development of white

markings is shown to be even now taking place, but the

amount of pigment reduction is still fluctuating so that

the white markings vary much in extent with different

individuals.

7. The development of such white markings takes place

probably by little and little, so that the departure from

564 THE AMERICAN NATURALIST [Vou. XLVIII

type is not so great as to arouse antagonism against the

varying individual on the part of others of its species.

Also, the gradualness of the change allows the species to

become accommodated to any ESN | that might

concomitantly arise. ,

8. The converse of this centripetal style of pigmenta-

tion is present in many species, and results in pigmenta-

tion (commonly black) at the extremities or along lines

where primary breaks occur in the centripetal form,

namely at the tip of the nose, ears, tip of the tail or the

toes; possibly the black dorsal stripe is due also to centri-

fugal pigmentation. Patterns may develop as in certain

antelopes by a white break between patches of the two

types. 2

In conclusion, I wish to express my indebtedness to

Professor W. E. Castle for much helpful criticism and `

advice, and to the Museum of Comparative Zoology for

permission to make record of specimens in its study

collection.

REFERENCES

Allen, G. M.

1904. The Heredity of Coat Color in Mice. Proc. Amer. Acad. Arts and

_ Sei., Vol. 40, pp. 61-163.

Brewer, W.

On ‘the Disposition of Color-markings of Domestic Animals.

mer. Assoc. Adv. Sci., Vol. 30, pp. 246-251.

Butler, A. x

1888. Notes Concerning Albinism among Birds. Jour. Cincinnati Soc.

Nat. Hist., Vol. 10, pp.

1888a. pro in the Cuvier Club Collection. Jour. Cincinnati Soc. Nat.

+ Vol. 10, pp. 216-217.

Castle, W. k p MacCurdy, H.; also Simpson, Q. I.

Cory, C. B.

1912. The Mammals of Ilinois and Wisconsin. Field Mus. Nat. Hist.,

Zoo t 1L

TARTERA :

Some a of Albinism. Ornithologist and Oologist, Vol. 9, p. 48.

ee, E.

1878. Melanism of Turdus migratorius. Bull. Nuttall Orn. Club, Vol. 3,

pp. 47-48.

No. 573] PATTERN DEVELOPMENT ; 565

Deane, R.

1876. Sr te and Melanism among North American Birds. Bull.

Nuttall Orn. Club, Vol. 1, pp. 20-24.

1880. Additional Cases of Albinism and Melanism in North American

Birds. Bull. Nuttall Orn. Club, Vol. 5, p. 25 (also 1879, pp. 26-

)

30, Vol. 4

Faxon, W.

1913. Brewster’s Warbler i ge leucobronchialis) a Hybrid

between the Golden-winged Warbler (Helminthophila chrysop-

a) and the Blue- pept Warbler "read dag pinus).

. Mus. Comp. Zeol., Vol. 40, pp. 316,

Hoffman, ap Hi

1878. Remarks upon Albinism in Several of Our Birds. AMER. NAT.,

Vol. 12, pp. 474—476

Keller, C. A.

1893. Evolution of the Colors of North American Land Birds. Occa-

sional Papers Calif. Acad. Sci., No. 3, xii + 361 pp., 19 pls.

Lawrence, G. N.

1889. Remarks upon Abnormal Coloring of Plumage Observed in Several

Species of Birds. Auk, Vol. 6, pp. 46-50.

ES C. C.

1914. ‘‘Dominant’’ and ‘‘Recessive’’ Spotting in Mice. AMER. NAT.,

Vol. 48, pp. 74-82.

sare . A,

1885 Albiniem. Auk, Vol. 2, pp. 113-114.

MacCurdy, H., and Castle, W. E.

1907. S élection and Groes-broading in Relation to the Inheritance of

oat-pigments and Coat Patterns in Rats and Guinea-pigs.

Carnegie Inst. Washington, Publ. 70, iii + 50 pp., plate.

T 8. H. ©.

1862, Faerornes fuglefauna med bemaerkninger om _ fuglefangsten.

Jidensk. Meddelels, Copenhagen, 1862, pp. 1-78.

Pearl, R.

1914. On Aah Results of Inbreeding a Mendelian Population: a Correc-

and Extension of Previous Conclusions, AMER, NAT., Vol.

i pp. 57

Pearl, R., and Boring, Alice M.

1914. ‘Bonk paoe KX Observations Regarding Plumage Patterns.

Science, New Ser., Vol. 39, pp. 143-144.

Pocock, R. I.

1907. On the Black- Fa Tan Pattern of Domestic Dogs (Canis fa-

miliaris), Ann, Mag. Nat. Hist., Ser. 7, Vol. 19, pp. 192-194.

1909. On the Colors of Horses, Coes and Tapirs, Ann. Mag. Nat.

Hist., Ser. 8, Vol. 4, pp. 404-415.

Ramaley, F, -

1912. Mendelian Proportions and the Increase of Recessives. AMER.

Nar., Vol. 46, pp. 344-351. a

Simpson, Q. I., and Castle, W. E. ,

1913. A Family o of a Negroes. AMER. NAT., Vol. 47, pp. 50-56,

Figs. i z

566 THE AMERICAN NATURALIST [Vou. XLVIII

Stone, W.

1912, The Phylogenetic Value of Color Characters in Birds. Jour. Acad.

Nat. Sci. Phila., Ser. 2, Vol. 15, pp. 311-319, pl. 27.

Strong, R. M.

1904. The Metallic Colors of Feathers from the Sides of the Neck of the

Domestie Pigeon. Mark Anniv. Vol., New York, pp. 263-277,

lL 2

pl. 20.

1905. Causes of Blue and Green in Feathers. Biol. Bull., Vol. 8, pp.

237-238

Swarth, H. S.

1913. A Study of a Collection of Geese of the Branta canadensis Group

from the San Joaquin Valley, comic Univ. of Calif. Publ.,

Zool., Vol. 12, pp. 1-24, pl. 1-2, 8 text-figs.

Sweet, D. A.

1907. Notes from Avon PETEN juncos from Maine]. Jour. Maine

Ornith. Soc., Vol. 9, p.

i G., and A. H.

909. Coneealing Coloration in the Animal Kingdom. New York.

=a tee C: H

1883. Some Albinos in the Museum of the Philadelphia Academy. Bull.

Nuttall Orn. Club, Vol. 8, p. 126.

Townsend, C. W.

1908. On the Status of Brewster’s biegi PESIS leuco-

bronchialis). Auk, Vol. 25, pp. 65-6:

Ward, H. L.

1908. A Rapid Melanistic and Subsequent Partial Albinistic Change in

a Caged Robin. Bull. Wisconsin Nat. Hist. Soc., Vol. 6, pp.

43—47.

Worthen, C. K.

1897. Albinism, Melanism and Hybridism. Osprey, Vol. 1, pp. 23-24.

SHORTER ARTICLES AND CORRESPONDENCE

THE BEARING OF THE SELECTION EXPERIMENTS

OF CASTLE AND PHILLIPS ON THE

VARIABILITY OF GENES

CAsTLE and Phillips have recently reviewed the results of six

years’ work in which they selected for and against ‘‘hoodedness’’

in rats.! In ‘‘hooded”’ or ‘‘piebald’’ rats only part of the coat

is pigmented ; the area of dark (versus white) coat varies greatly

in different animals, but tends, in those of medium grade, to

cover the head, shoulders and middle of the back, like a hood.

Starting with a strain which was probably hybrid, although of

unknown ancestry, and selecting during thirteen generations

for a larger extent of colored coat (‘‘plus’’ selection), they suc-

ceeded in obtaining animals with a greater and greater area of

pigmentation. The average, the mode, and the extremes were

raised. Conversely, selection for less pigmentation (‘‘minus”’

selection) was accompanied by a gradual but decided and

continual dimunition in the dark area. ‘‘Return’’ selection

also succeeded ; that is, plus selection was effective even in a line

which was already lighter than the average on account of a

previous minus selection, and, vice versa, minus selection caused

a lightening of a strain that had been made exceptionally dark

by a prior plus selection.

Certain crosses proved that more than one factor affecting

hoodedness is involved in the difference between the different

races. Therefore the production of animals of desired grade by

selection may perhaps be explained as a mere sorting out, into

different lines of descent, of different combinations of the various

factors for hoodedness originally present in the heterozygous

ancestors. It is the opinion of Castle and Phillips, however, that

this explanation will not suffice to account fully for the continued

efficacy of selection in their experiments, and they believe it

probable that a factor or factors for hoodedness are undergoing

variation of a fluctuating nature.

1 Castle and Phillips, ‘‘Piebald Rats and Selection, An experimental test

of the effectiveness of selection and of the theory of gametic purity in

Mendelian crosses.’? Published by the Carnegie Institution of Washing-

ton. See also Castle’s ‘‘ Pure Lines and Selection’’ in American Breeders’

Magazine, 1914.

567

568 THE AMERICAN NATURALIST [Vot. XLVIII

A conelusion so radical and so opposed to previous work

should not be accepted, however, as long as it remains at all

reasonably possible to use instead an explanation in harmony

with the results of Johannsen and other investigators. Johann-

sen dealt with a character—dimensions of seed—which must be-

yond any doubt have been partially dependent upon a very great

many factors, yet he found that selection had no effect whatever

after he had separated the different genotypes from one another.

Thus he proved the constancy of a great many genes ‘‘at one

blow’’—namely, of all the genes appreciably concerned in seed

size. Of course, if there had been a chance for cross-fertilization

in his experiments, he, like Castle, would have obtained a result

from selection, but this would have been due to recombination,

not variation, of genes. All our evidence points to the conclusion

that the vast majority of genes are extremely constant, although

they differ somewhat in that very slight amount of variation

which they do show. For example, in Drosophila, although in

the case of most genes not more than one mutation has been

found, yet in one case (possibly in two or three cases) a locus has

mutated three times, each time in a different way, thus giving

rise to a system of multiple allelomorphs containing four mem-

bers. This gene evidently is more subject to mutation than the

others, yet this formation of a series of multiple allelomorphs can

not even remotely be compared to fluctuating variability, for the

three mutations were all large steps (much smaller could easily

have been detected), and they were found only during the exami-

nation of some millions of individuals in the rest of which the

locus was not observed to mutate at all. Some few genes are

known, however, which really do change frequently (e. g., that

for ‘ Gapiaghted?? corn), but these cases are extremely rare;

moreover, here the degree and nature of the change are fixed,

and also, after the change has once occurred the instability of the

gene is lost. Thus, in no known ease do the variations of a gene

among, let us say, several thousand immediate descendants of the

individual possessing it, form a probability curve, as neo-Dar-

winians might perhaps suppose, nor even are any cases known

where genes can undergo frequent changes that may vary at all

in kind or amount or occur successively.

Let us then inquire into the probability and gdeanay” of that

explanation of Castle and Phillips’s results which does not require

the assumption that a gene or genes involved change compara-

No.573] SHORTER ARTICLES AND CORRESPONDENCE 569

tively frequently and suecessively, but which assumes a sort-

ing out of numerous factors. It is now pretty generally ac-

cepted by Mendelians that the germ plasm of any of the higher

organisms contains a large number of genes, which play vari-

ous rôles in the numberless processes and reactions of devel-

opment whereby the egg is transformed into the adult indi-

vidual. The exact nature and intensity of any one characteristic of

this adult organism (e. g., hoodedness in rats) is dependent upon

the nature of each of the various reactions which were involved

in producing this character, and thus dependent upon all the

genes (and environmental factors also) involved in any of

those reactions. Now, in an ordinary Mendelian cross, all the

individuals are usually homozygous and alike in respect to all

but one of the pairs of genes that noticeably affect the character

concerned. In such a case, then (so far as differences in environ-

mental influences do not obseure the outcome), one obtains the

simple Mendelian results derived from the segregation, at reduc-

tion, and recombination, at fertilization, of but this one pair of

allelomorphs.

The strain of hooded rats, however, was probably a hybrid

between two races of rather remote relationship. When two such

races are crossed, the individuals often differ in more than one

pair of those factors that affect the character studied, especially

if the character is such as to be influenced. by a relatively large

number of genes, It can not be questioned that some characters

are thus determined or influenced by a much larger number of

developmental reactions than are others, and such characters

will therefore vary more in inheritance, since if a difference

exists between two individuals in respect to any given gene,

these characters are more likely to be affected than others. Gross

size, for example, is a character dependent in this way upon an

exceptionally large number of genes, for any gene which influ-

ences the size of any organ must affect to some extent the total

size. In some other cases in which characters are found to be

influenced by relatively many genes, the reason for this is not-

so evident, e. g., in the case of the red flower-color of flax, or the

truncated condition of the wing in some races of Drosophila..

Here the production of the character may be conceived to be |

dependent upon some reaction that can be easily modified by

various means.? For our present purpose we must assume that

2It is conceivable that differences in respect to numerous genes hav ve:

sometimes arisen even in the ¢ ease of characters not naturally very B ;

570 THE AMERICAN NATURALIST [Vou XLVIII

the character ‘‘hoodedness’’ belongs in this class and that the

ancestral hooded rats used by Castle and Phillips were the de-

scendants of a cross involving many genes for that character.

The results of such a cross are of course complicated, for the

different pairs of allelomorphs generally can undergo recombina-

tion at the reduction division of the hybrid, so that in F, or

subsequent generations as many different genetic types of indi-

viduals are formed as there are possible different combinations

of those factors wherein the ancestors differed. Not all these

genetic types, of course, will fall into different phenotypes, yet

generally there will be a large number of overlapping pheno-

types among the progeny.

The larger the number of factors in which the two ancestral

lines differed, the larger will be the number of different possible

combinations of these factors, and accordingly the smaller will

be the chance of any individual having one of those particular

combinations necessary to a relatively high or a relatively low

intensity of the character. In other words, the larger the num-

ber of factors (for one character) for which a population is

heterogeneous, the more numerous are the possible different

grades of intensity of this character among the different indi-

viduals, but the fewer will be the individuals which approach the

more extreme grades theoretically possible in such a population.®

Suppose, for example, that two parents differ in five pairs of

factors for hoodedness, which are partially dominant* to their

allelomorphs and summative in their action. Then in F, not one

influenced by diverse means, merely because one of ae two races had been

subjected to a very long and drastic selection, so that any of those rare

mutations which affected that character in the desired es had in thi

race been preserved. Selection in such a case, however, would have to in-

volve many millions of orn ual

3 One extreme, e. g., the ‘plan,’ ’ will be rather frequent, however, if all the

‘*plus’’ factors dominate completely. But in the case of the hooded rats

we must assume either that dominance is generally incomplete or that in

the case of some factors the ‘‘minus’’ allelomorph dominates in the case

of others the ‘‘plus,’’ since F, rats from a cross of the plus by the minus

strain are on the average intermediate in type between these two extremes

#It is of course by no means necessary to assume incomplete dominance

of the factors. If dominance is complete (in some cases the ‘‘minus’’ fac-

r may dominate, in others the ‘‘plus’’), the rigor of selection will be di-

minished, since heterozygous forms can not be distinguished from homozy-

us. Therefore, although a somewhat greater number of individuals will be

found having the limiting values, it will take longer to bring the average up

to the limit.

No.573] SHORTER ARTICLES AND CORRESPONDENCE 571

individual in a thousand will have the most extreme dark or

light grade of hoodedness possible. However, by selecting the

more extreme individuals, and mating them together, a still more

extreme grade of hoodedness may be obtained in F, (both as to

average and limiting values), and the same process may be con-

tinued for a good many generations. The number of generations

during which effective selection is possible depends on the num-

ber of factors concerned, the rigor of selection, and the amount

of inbreeding of brother to sister.

In regard to the latter point, since brother and sister are much

more apt to be alike in their genetic constitution than are other

individuals, offspring from such a mating are more apt to be

homozygous and alike, or, we may say, such offspring will tend to

be homozygous and alike in a larger number of factors; then,

mating two individuals homozygous for these factors together,

there will be much less variation and so less opportunity to con-

tinue selection among their progeny. In the case of Castle and

Phillips’s experiments, however, no such attempt at inbreeding

was reported. Here, then, the individuals mated together would

be more apt to differ genetically, even though they looked alike

(thus, one might be AA bb, the other aA bB), and their

descendants would therefore present a larger number of different

combinations of factors for the selector. Often a greater effect

may be eventually. produced in this manner than by inbreeding,

for a larger number of combinations of factors are thus pro-

duced, some of which may be of more extreme type. The effect

would usually be slower, however, since such matings tend to

keep the strain heterozygous and are often steps backwards.

Cross-breeding, then, will help to explain the relatively slow but

long-continued and eventually large effect of selection in Castle

and Phillips’s experiments, although such a result could also be

obtained without cross-breeding if the factors were numerous

enough,

The ‘‘return selections’? also are easily explicable on the

multiple factor view. Due to the original difference in so many

factors, and the fact that cross-breeding diminishes the tendency

to homozygosis which selection favors, the rats were presumably

heterozygous even after generations of selection. They would

not be as heterozygous as before, of course, and, correspondingly,

Castle and Phillips did find less variation in the rats after selec-

tion. Yet there would still be a good chance for recombination,

572 THE AMERICAN NATURALIST — [Vou. XLVIII

and an alteration in the race could therefore be produced by

further selection or by return selection. As we have seen, this is

especially true if certain factors are completely dominant, al-

though dominance is by no means a necessary condition.

As a very simple illustration, let us suppose that the ‘‘plus’’

factors A and B dominate over the ‘‘minus’’ factors ‘‘a’’ and

‘*b,’’ respectively, and each increase the pigmented area to about

the same extent. To begin with, two moderately hooded indi-

viduals, Aa bb and aa Bb, were mated together. They produced

laa bb—light-hooded, laa Bb and 1Aa bb—both moderate, and

lAa Bb—dark. We first select for dark; mating the dark

rats together, 9 darks, 6 moderates, and 1 light, would be pro-

duced (F,). The average color of the offspring has thus been

increased by selection (the limiting color, too, if dominance is

incomplete). It can be still further increased in subsequent

generations. On the other hand, the color can be made lighter

again by a ‘‘return selection,’’ for if, instead of mating the F, or

F, darks together, we mate the moderates or mate darks with

moderates, many of the matings will give offspring lighter, on the

average, than in the preceding generation; e. g., Aa Bb by Aa bb

- gives 3 dark, 4 moderate, 1 light, as compared with the previous

9 dark, 6 moderate, 1 light. In subsequent generations, the

average could be brought still lower.

Let us now see whether there is any experimental evidence

in support of the multiple factor explanation of Castle and

Phillips’s results, aside from the fact that it is adequate and is

the only one consistent with other work. One point of evidence

we have noted—the variability of the rats continued to decrease

as a result of selection in either direction. This we should of

course expect on the multiple factor view, for selection gradually

tends towards homogeneity in a population, even though it may

require a long time to produce complete homogeneity. The

second and strongest evidence is from crosses.

The crosses show that one of the factors concerned in differ-

entiating hooded rats from wild rats, which are pigmented all

over, or from ‘‘Irish’’ rats, which are almost completely pig-

mented, is ‘‘hypostatie.’’ In other words, a rat having the

normal allelomorphs of this factor will always be self-colored, or

nearly so; one having the other allelomorphs will always be

distinctly hooded, although the amount of the hoodedness varies.

“Self, ”’ as it happens, is dominant, in this case, over hooded.

No.573] SHORTER ARTICLES AND CORRESPONDENCE 513

Thus, on crossing a hooded to a wild or Irish rat, all the F, are

self (or nearly so) ; in F, there are three selfs to one hooded, but

the hoodeds vary in intensity. The question then is, does this

variation (so far as it is not due to ‘‘environmental’’ differences)

depend upon what other ‘‘epistatic’’ or ‘‘modifying’’ factors for

hoodedness may or may not be present, or is there evidence that

it depends instead, or in addition, upon a variability of one or

more of the factors for hoodedness? As will be shown below, it

ean be proved that different combinations of modifying factors

do occur in the different hooded indiviuals: this being true, there

can be no ground for making the unusual postulate that in this

case or in the selection experiments a factor or factors concerned

undergo variation. |

The proof is that when light hooded rats from the minus

strain are crossed to wild or Irish rats the hooded rats in F, vary

` much more than did the original strain of hooded rats and aver-

age much darker. Obviously, the P, hooded rats differed from

the wild or Irish in a number of modifiers as well as in the hypo-

static factor; moreover, as we should have expected, this differ-

ence consisted chiefly in the fact that the wild or Irish rats con-

tained ‘‘plus’’ allelomorphs in place of some of the ‘‘minus’”’

modifiers present in the P, strain that had undergone minus

selection. Thus the F, hooded rats, containing various combina-

tions of these modifying factors wherein the two strains differed,

varied much more than did the parental strain of hooded rats,

and were on the average much darker.

In order to escape this conclusion that modifying factors were

involved, Castle and Phillips at first postulated that the reason

that the F, hooded were darker than the original ‘‘minus’’ strain

was because the factor for hooded had in many cases become con-

taminated by its allelomorph (the factor for self) in the F, rats.

This is violating one of the most fundamental principles of

genetics—the non-mixing of factors—in order to support a vio-

lation of another fundamental prineiple—the constancy of fac-

tors. The refutation of their supposition came unexpectedly

soon. It would be expected, on the view of multiple factors, that

the wild or Irish rats (containing the allelomorph for self in

place of the hypostatie factor for hooded) would not possess as

many ‘‘minus’’ modifiers as the hooded strain which had been

Specially selected to contain as many of these as possible; neither

would these ‘‘self’’ rats contain as many ‘‘plus’’ modifiers as the

574 THE AMERICAN NATURALIST [Vou. XLVII

hooded strain which had undergone plus selection (and which so

contained nearly all of the plus modifiers originally present in

either the self or the hooded ancestors). Thus it was to be ex-

pected that, just as a cross of self with the minus race gave F,

hooded rats darker than the original minus strain, so a cross of

wild or Irish rats with hoodeds resulting from the plus selection

would give F, hooded rats lighter than those of the plus strain.

This result was actually obtained. It was fatal to the idea that

the difference between the P, strain of hooded rats and the F,

hoodeds was due to contamination of the allelomorph for hooded

with that for self, since such contamination should have resulted

in F, hooded rats darker than those of P,, not lighter. For wild

and Irish rats are both much more extensively pigmented than

hoodeds even of the plus strain.

The change in hoodedness from P, to F, was therefore due

to recombinations of the modifying factors wherein the two '

strains differed. That many such modifiers were concerned is

indicated by the evenly distributed variability of the F, hoodeds

and the fact that very few were as extreme as the hooded grand-

parents. The same fact is brought out in a cross of the minus

with the plus race; here no clear-cut ratios were obtainable, the

classification into different genotypes being rendered impossible

by the multiplicity of factors (no one of which was hypostatic

as in the other crosses). Of course, this knowledge of so many

factors being concerned in the crosses helps our interpretation

of the selection results decidedly, for the more numerous are the

factors concerned, the longer would it be possible to continue an

effective selection on the progeny of the hybrids, and the orig-

inal hooded rats of the selection experiments were admittedly in

all likelihood descended from just such hybrids. The exact num-

ber and effect of the different factors can not be determined from

Castle and Phillip’s data, since to do this very special crosses

must be made and individual pedigrees kept. Selection experi-

ments can be of little value so long as there are factors for which

the individuals may be heterozygous, unless these factors can be

accurately followed in inheritance.

Of course, it is quite possible that in the course of these long-

continued experiments mutations affecting the hoodedness occa-

sionally happened to arise, especially since it seems likely that

this character is dependent upon an unusually large number of

genes, for then, as a matter of mere chance, any mutation which

No.573] SHORTER ARTICLES AND CORRESPONDENCE 95175

occurred would be more likely to affect it than it would be to

affect most characters. It is interesting to note that one such

mutation, of a very marked and unquestionable character, was

in fact observed. The mutant factor proved to be a strong

““plus’’ modifier, which was almost completely dominant, and

itself showed no contamination or variation, so far as could

be determined. It arose, as it happened, in the plus strain.

A part of the effectiveness of selection may therefore have been

due to the occurrence and sorting out of such occasional muta-

tions, but there is no way of telling how many of these took place,

or any need for assuming them at all in explaining the result.

These rare mutations, however, would form a very different phe-

nomenon from such fluctuating or frequent and progressive vari-

ation of a gene or genes concerned as Castle postulates. Although

the academic possibility of variation of the latter type can not

be denied, there is no experimental evidence which can be used

to support it, and there is good evidence against it in many

individual eases.

It is difficult to believe that this suggestion of Castle and

Phillips was not made in a spirit of mysticism, when we con-

sider also their suggestion that the genes may undergo contami-

nation, and especially when we consider the following passage,

with which their paper concludes:

It seems to us quite improbable that the plus mutation could have

arisen in the minus selection series. We believe that the repeated se-

lection which was practised had something to do with inducing this

change in the plus direction. If one ean increase at will the “ modi-

fiers” which make the pigmentation more extensive, it does not seem

strange that after a time a readjustment should oceur within the cell

which should incorporate modifiers in that part of the cell which is re-

sponsible for the unit-character behavior of the hooded pattern. This

would amount to a quantitative change in the unit-charaeter for hooded

pigmentation.

To thus suppose that independent genes fuse or induce changes

in one another, merely because they happen to produce similar

end effects upon the organism, and in spite of the fact that they

usually lie in different chromosomes and are apt to differ from

each other as much as do other genes, is utterly teleological.

A paper by A. L. and A. C. Hagedoorn criticizing Castle’s work

and conclusions, appeared at the same time as the paper of

576 THE AMERICAN NATURALIST (VoL. XLVIII

Castle and Phillips.» The Hagedoorns champion the multiple

factor hypothesis as an explanation of Castle’s results, and also

cite certain rather inconclusive experiments of their own to sup-

port this point of view. They err, however, in supposing that

the factors concerned must be incompletely dominant; as we

have seen, this is not a necessary assumption, if we admit that

in the case of some modifiers the ‘‘minus’’ allelomorph dominates,

in others the ‘‘plus.’’ They also err in denying the poaki,

on the multiple factor view, of successful ‘‘return selection,” if

inbreeding be strictly followed. In fact they offer this as a test

of their point of. view. As we have seen, ‘‘return selection’?

would be possible in some cases, even if the animals were inbred;

and in Castle and Phillips’s experiments, where inbreeding was

not followed, ‘‘return selection’’ was certainly very effective.

Finally, papers have recently appeared by MacDowell,® in

which he gives evidence that certain other cases of inheritance

(e. g., head size in rabbits), formerly considered by Castle to

support the idea of genic variation and contamination, are

probably best interpreted on the view of multiple factors instead.

His evidence consists in the fact that the characters concerned

are somewhat more variable in the offspring of. back-crosses than

in F,, as we should expect on the basis of recombination of

multiple factors, but which he believes could not plausibly be

explained otherwise.

HERMANN J. MULLER

SA. L. & A. C. gee Le ‘í Studies on Variation and Selection,’’ Zeit. f.

ind. Abst. u. Verab.,

6 E. C. MacDowell, aiei Factors in Mendelian Inheritance,’’ Jour.

Exp. Zool., 1914, and Carnegie Inst. of Wash., 1914.

VOL. XLVIII, NO, 574 OCTOBER, 1914

pend

THE

AMERICAN

NATURALIST

A MONTHLY JOURNAL

Devoted to the Advancement of the Biological Sciences with

Special Reference to the Factors of Evolution

CONTENTS

Page

Sex-limited and Sex-linked Inheritance. Professor T. H. MORGAN =- — 677

. Inheritance of Endosperm at in Sweet x hse oo of Maize. G. N.

COLLINS and J. H ~ 684

2 G Res

A Study of Variation in the Apple. W. J. YOUNG z

Shorter Articles and Discussion: Variation and Correlation in the Mean Age

at Marriage of Men and Women. RTHUR a Roxana H.

VIVIAN. Duplicate Genes. SEWALL WRIGHT = - 635

Notes and Literature: A Study of Desert Vegetation, Professor CHARLES E.

Bessey - = > a i a e | CE

THE SOIENCE PRESS

LANCASTER, PA. GARRISON, N. ¥.

NEW YORK: SUB-STATION 84

The American

Naturalist

intended for piui and books, etc., intended for review should be

MSS

sent to Ba Editor of THE AM

RICAN NATURAL

rt articles saatat summaries of rese

IST, Garrison-on-Hudson, New York.

earch work bearing on the

igor -i Sear evolution are especially welcome, and will be given preference

in pos

rea reprints of eaae are supplied to authors free of charge.

con

Further ‘eprint will be supplied a

ons and

E rer Sabi is four ees a

aerea postage twenty-

ents.

adver visdaens should be sent to the publishers.

ear.

e ce as additional.

The aiivertinine: rates are Four Dollars for a page.

The

Foreign postage is fifty cents pe

The charge for single copies

THE SCIENCE PRESS

Lancaster, Pa.

Garrison, N. Y.

NEW YORK: Sub-Station 84

Entered as second-class matter, ipa 2, 1908, at the Post pery at Lancaster, Pa., under the Act of

ngress of Mareh 3, 1

F E

ARCTIC, ICELAND and GREENLAND

BIRDS’ SKINS,

Well Prepared Low Prices

Particulars of

G. DINESEN, Bird Collecto

Husavik, North Iceland, Via aiie koai

WANTED TO PURCHASE

a set of BIRDS OF AMERICA by J. J. Audubon,

7 or 8 volumes, please report, stating cash price, stat-

ing condition, binding and dates of volumes.

F. C. HARRIS,

Box 2244 Boston, Massachusetts

For Sale Entire

An important collection of Indian Birds’

Eggs, containing approximately 3,300

specimens of about 620 species, and in-

| Marine Biological Laboratory

cluding among other varieties the Green- |

ish Willow Warbler (Acanthopneuste

viridanus), Red-browed Finch (Calla- |

canthis burtoni), Tibetan Sand-Grouse |

(Syrrhaptes tibetanus), Spotted Sand-

Grouse (Pteroclurus senegallus), Ibis- |

bill (Ibidorhynchus struthersi), White-

bellied Minivet (Pericrocotus erythro-

pygius), Bearded Vulture (Gypaétus

barbatus), numerous Cuckoos with their

hosts, etec., etc.

Particulars and full list may be had

from

W. F. H. ROSENBERG

57 Haverstock Hill London, N. W.

Woods Hole, Mass.

INVESTIGA Facilities for research in Zoology,

TION Embryology, Physiology and Bot-

Entire Year any. Seventy-six private labora-

Leste ges oe a for not over three

tables se sre

ae AA ia te eat

Ses — wek ce werden the pe

of m rs of the maf. The fee

for su $50.

urses of laboratory instru

truction

INSTRUCTION t laboratory ta tnverte-

1 to August 11,

$50.00. A lecture course

pe ag wg A ing E

Animals ts, a

pfeil Hiatt Shere

Open the Entire Year

THE

AMERICAN NATURALIST

Vout. XLVIII : October, 1914 No. 574

. SEX--LIMITED AND SEX-LINKED INHERITANCE

PROFESSOR T. H. MORGAN

CoLUMBIA UNIVERSITY

Darwin used the expression ‘‘inheritance as limited by

sex’’ to include all cases in which a character is peculiar

to one sex. His list of such cases covers in the main the

group of secondary sexual characters. Darwin’s expres-

sion has been contracted to sex-limited inheritance, and is

widely employed to-day in the same general sense in which

Darwin used the expression. For instance, Bateson in

his book ‘‘Mendel’s Principles of Heredity’’ includes

both horns in sheep and color blindness in man as sex-

limited characters.

Now that the inheritance of several of these cases has

been definitely worked out, it has become increasingly evi-

dent that such characters as color blindness, and hemophi-

lia in man, the twenty-five ‘‘sex-linked’’ characters in Dro-

sophila, and certain characters in birds and in butterflies

follow a law of inheritance that is essentially different

from that followed by some of the other cases. It has

become necessary, therefore, to recognize two groups of

cases that differ fundamentally in regard to their heredity.

To one of these groups I have applied the term sex-linked

inheritance, and, for the present at least, we may still make

use of the older expression sex-limited inheritance (and

1 See pp. 169-174 in section headed ‘‘ Heredity Limited by Sex; the Horns

of Sheep,’’ where the term sex inheritance limited descent (p. 172) also

appears.

577

578 THE AMERICAN NATURALIST [Vou. XLVIII

sex-limited character) to cover that class of cases (obvi-

ously a very mixed one which will be broken up as our

knowledge regarding it becomes more certain) that in-

cludes largely, as originally intended, the secondary sexual

characters.? In those cases of sex-linked inheritance, in

which the male is heterozygous for the sex factor, the

grandfather transmits his peculiarity, through his daugh-

ters, to half of his grandsons only; and reciprocally an

affected female transmits her peculiarity to all her sons,

and, through her sons bred to her daughters, to half of

her granddaughters and to half of her grandsons’. More-

over the appearance of the character in the female is not

exceptional or abnormal, as is sometimes implied in cases

like color blindness in man, for, the character can always

be transferred from the male to the female by suitable

crosses,

On the other hand, there are cases in which a character

appears in one sex only—the character is limited, there-

fore, to the male or to the female. Such cases may be

properly called sex-limited, and were so called by Darwin.

As typical examples I may cite the horns of certain races

of sheep that are present in the ram and absent in the

2G. H. Shull has recently said (Zeit. Ind. Abst. und Vererb., XII, 1914,

p. 160) that, in his opinion, it would be better to retain the term sex-limited

for those cases that I call sex-linked and call other cases secondary sexual

characters. This view is not historically in accord with Darwin’s usage 0

e term ‘‘limited by sex.’’ This fact, in itself would be a sufficient argu-

ment for rejecting Shull’s suggestion, but, in addition, the term sex limited

is an actual misnomer for the class of cases to which he proposes to apply it.

There are cases like the eosin eye of Drosophila that differ in the male and

female in the same way as do many secondary sexual characters (in fact they

are such in a descriptive sense) but nevertheless show sex-linked inhert-

tance. Since a new name is required to express our fuller information

regard to some of the characters that were mapa included under the

older term, why not begin by adopting suita n

3 In those cases in which the female is paranda us for a sex factor, as

in birds and in butterflies, the same principle is involved but the sequence is,

in a sense, reversed; thus the grandmother transmits, through her sons, her

peculiarity to half of her granddaughters; and reciprocally, the affected

male transmits his peculiarity to all of his daughters, and, through his

daughters bred to his sons, to half of his grandsons and to half of his grand-

daughters,

No. 574] INHERITANCE 579

ewe (or else more developed in the ram than in the ewe);

the color of butterflies like Papilio Memnon, with three

types of females; and the dark spot on the abdomen of the

male of the bug Euchistus variolarius. These characters

can not be transferred through the gametes to the female

of their own race by any known combination.

Whether one likes or does not like the particular terms

used to denote these two classes of cases, the fact remains

that there are two such categories, and to ignore their

existence is only to make obscure a distinction that is per-

fectly plain.

Concerning the mechanism involved there is something

more that may be said. It has been sufficiently shown in

the case of sex-linked inheritance that the sex-linked char-

acter follows the known distribution of the sex chromo-

somes. It is unnecessary to repeat here the abundant

evidence jn support of this statement. The simplest inter-

pretation of this known relation is that the character is

dependent for its realization on the sex chromosomes. I

do not mean, of course, that the sex chromosomes alone-

produce the character but that something in these chromo-.

somes, some ‘‘factor,’’ acting in conjunction with the rest

of the cell, conditions the character.

On the other hand, in the ease of sex-limited characters

the facts can not be explained on the assumption that the

characters follow the sex chromosomes. It is clear that

they do not do so. But we can give a consistent interpre-

tation of the facts if we assume that sex-limited characters

follow the distribution of the ordinary chromosomes.

Since this relation has recently been not understood

and misinterpreted I may be pardoned, I hope, for taking

up the question once more.

Wood crossed horned Dorset sheep with hornless Suf-

folks. The sons had horns, the daughters lacked them.

Inbred these gave in the F, generation—horned J, 3;

hornless 3, 1; horned 9, 1; hornless ?, 3. Bateson and

Punnett have shown that the results are explicable on the

basis that one factor for horns in the male produces

580 THE AMERICAN NATURALIST [Vou XLVII

horns but one factor is insufficient in the females. This

conclusion was put to the test by breeding an F, hornless

ewe to a hornless ram. The F, ewe should be hetero-

zygous for the factor for horns, and, therefore, when she

is bred to a homozygous hornless ram, half of her off-

spring should be heterozygous for hornlessness and half

homozygous for hornlessness. Since half of her sons

should have a factor for horns they are expected to

develop horns, and this is what occurred. Half of the

daughters also should have a factor for horns, but should

not develop horns, and this also was true.

It has been recognized for several years that this and

related cases can not be explained on the assumption that

the factors involved are carried by the X or by the Y

chromosomes. But we can interpret the statement that

one factor for horns is sufficient in the males to call forth

horns, but not sufficient in the female ‘‘in terms of chromo-

somes,’’ if a factor for horns is carried by one of the

chromosomes other than the sex chromosome. In other

words we need only appeal to a mechanism with which we

are familiar to cover the results.

The second illustration is furnished by the recent

experiments of Foot and Strobell, and since the authors

have rejected the chromosome hypothesis as inapplicable

to their results, and since in the case of insects the condi-

tions are simplified because castration experiments have

shown that the sex glands are not themselves responsible

for the secondary sexual characters, we may profitably

consider this case even more fully.

In one of the bugs, Euchistus variolarius, the male has

a black spot on the abdomen. The female lacks the spot.

A female of this species was crossed to a male of another

species, viz., Huchistus servus, having no spot in either

sex. The daughters had no spot, the sons had a spot

fainter than that of variolarius. Inbred these gave, in F,,

249 females without a spot, 107 males with a spot (devel-

oped to different degrees) and 84 males without a spot.

The F, results show that one factor for spot in the male

No. 574] INHERITANCE 581

suffices to call forth in some degree the spot in the hybrid.

Its intensity varies from a condition approaching that in

pure variolarwus to a faint spot (possibly even to no spot

at all). The F, results show also that a single factor in

the female fails to cause the spot to develop in that sex.

In the F, male the failure of the spot to reach in most

cases its full development shows obviously that the same

conditions that produce a male that is perfect so far as

his sex gonad is concerned, do not suffice to cause the full

development of the spot, although the factor for the spot

is present in one dose at least. The only confusion that

is liable to arise is that in none of the F, females did the

spot appear, although in some of them there must have

been a double dose of spot. But the difficulty is imaginary

as a little thought will show. In the first place the female

of E. variolarius herself does not show the spot, yet this

female must have a double dose of spot if spot is in the

X chromosome or in any other chromosome (except the Y).

Foot and Strobell by an elaborate analysis of the case

show that the factor can not be carried by either the X or

the Y chromosome. It is unnecessary to repeat their

argument; for, if the factor were carried by the X chromo-

some, only half of the grandsons should show it, while, in

' fact, many more than half of them show it; and it could

not be carried by the Y chromosome because the Y chromo-

some of variolarius is not present in the female, hence

could not have entered the cross as made. We are con-

cerned then only with a third possibility, viz., that there is

something in the female condition itself that is inimical to

the development of the spot. Since neither X nor Y

carries the factor in question it must be present in duplex

in the female of variolarius (if every gamete must have tt

in simplex and the experiment shows that this is the case),

and since the spot does not show in the female of vario-

larius, it is obvious that it can not appear in that sex even |

in duplex. If it be granted that the character is like other

Mendelian characters, and the authors’ evidence show that

it is inherited as are Mendelian characters, the conclusion

582 THE AMERICAN NATURALIST [Vou. XLVIII

is self evident; for, in demonstrating that all of the

gametes of variolarius carry spot the authors actually

destroy their own argument.

It only remains to point out some of the different ways

in which a factor being present in duplex both in the male

and in the female produces its effect only in the male. In

some cases it has been shown that the ovary produces

some substance that is inimical to the production of cer-

tain characters. For instance in fowls and in ducks the

presence of the ovary suppresses the development of

the male plumage. That the factors for the male plumage

are present is shown by its development when the ovary

is removed. But in some insects it has been found that

neither the ovary nor the testis produces these kinds of

substances; for, when the testis or the ovary is removed

the secondary sexual characters are not affected. Here

the mode of explanation must be different. But the con-

ditions, or complex, or factors that produce the ovary in

the female are acting in every cell of the body, and con-

sequently an effect, that is indirectly caused in the fowl

or duck, might be directly caused in the insect. For, each

cell is a chemical factory. Such a factory may help to

produce an ovary and the ovary produce a substance that

demonstrably suppresses the male plumage, or the same '

kind of factory may do similar work through the activity -

of some other part of the body, or conceivably it may do

its work in every cell of the body. This it seems to me is

the most reasonable view to take of the matter in the case

of the variolarius-servus cross. We can express the same

thought in symbols by representing:the female of vario-

larius by XXAABBCCDDSS, etc., and the male by

XYAABBCCDDSS, ete. The chemical interaction be-

tween two X’s and the rest of the cell is of such kind that

it produces a female, and the female complex, as such, is

inimical to the development of a spot and favorable for

the development of the accessory organs of reproduction

and of all secondary sexual characters of the female, while

XY and the rest of the cell is inimical to the development

No. 574] INHERITANCE 583

of the acessory organs and of the secondary sexual char-

acters of the female, and favorable for the development of

the accessory sexual organs and of the secondary sexual

organs of the male. This view is of course compatible

with the idea that there may be special factors for these

organs in chromosomes other than the sex chromosomes,

and the view holds both in a general way and on the

special chromosome hypothesis as well.

To assume that all the factors for characters that are

shown by the male or by the female must be carried by a

sex chromosome of some kind, if carried at all by chromo-

somes, is a travesty of the point of view of those who hold

to the chromosome hypothesis as a reasonable working

hypothesis to account for Mendelian inheritance. Just

as it has been shown that there are factors in the sex

chromosomes that affect many parts of the body, that are

not concerned with differences of sex; so, on the other

hand, the evidence shows that there are factors in other

chromosomes that are influential in producing secondary

sexual characters.

INHERITANCE OF ENDOSPERM TEXTURE IN

SWEET x WAXY HYBRIDS OF MAIZE

G. N. COLLINS anp J. H. KEMPTON

U. S. DEPARTMENT OF AGRICULTURE

INTRODUCTION

In a previous publication,’ the first and second genera-

tion of crosses between sweet and waxy varieties of maize

were reported and a tentative explanation of their be-

havior was suggested. It is now possible to add the

results of the third season, which to some extent afford a

test of the explanation proposed in our first publication.

The immediate result of crosses between the Chinese

variety of maize having a waxy endosperm and varieties

with sweet endosperm was the production of seeds having

a horny endosperm indistinguishable from that of ordi-

nary field varieties of maize. In the second xenia genera-

tion all three kinds of endosperm reappeared in the pro-

portion of 9.20 horny, 3.95 sweet, and 2.85 waxy. This

ratio was accepted as a 9:4:3 dihybrid ratio. For al-

though the deviations of the individual ears, individual

families and the totals were too large to be ascribed to

chance, the deviations were not consistently in one direc-

tion and to predicate more complicated formule would

have necessitated different assumptions for different ears.

The only interest in treating the problem in this way would

be that of solving a mathematical puzzle, for it would be

practically impossible to secure individuals enough to test

adequately the validity of the assumptions which it would

have been necessary to make.

Admitting, then, that the ratios were only an approxi-

mation representing a general tendency, it became of

1 Collins, G. N. and Kempton, J. H., ‘‘Inheritance of Waxy Endosperm in

Hybrids with Sweet Corn,’’ Circular 120, U. S. Department of Agriculture,

Bureau of Plant Industry, 1913.

584

No. 574] INHERITANCE 585

interest to learn whether predictions were still possible.

For the purpose of making comparisons easy, the

original diagram representing the second xenia genera-

tion is here repeated. (See Fig. 1.) The meaning of the

symbols is as follows: S is the factor for sweet, and X the

factor for waxy. When both S and X are present the seed

is expected to be horny. Small letters indicate the absence

or latency of the factors.

K SX a

ETT

KKA DEE SX”

TKE Sr Ko

=. e gig L Steel he aed ve a

ONK HORNY XY WAXY

KIX ATATA

HORNY SWEET WAY | SVE’

Sx sx SX SX SX

Fic. 1. Diagram showing the gametic composition of second-generation hybrids

between waxy and sweet varieties of maize

Since in both sweet and waxy the alternative factor

necessary to produce horny is assumed to be lacking, the

gametes produced by sweet varieties are represented by

Sa and the gametes produced by varieties with waxy endo-

sperm by sX. The synthetic horny produced by crossing

waxy and sweet is then represented by a combination of

586 THE AMERICAN NATURALIST [Vow. XLVIII

these, or SxsX. Assuming a chance recombination of

these factors in the gametes derived from these synthetic

horny seeds, the gametes will be of four kinds. Both the

sweet and the waxy may be present (SX) or the sweet

may be present without the waxy (Sa), or the waxy with-

out the sweet (sX), or both may be absent (sx). At ferti-

lization each of these kinds of gametes may unite with

any one of the four corresponding kinds derived from the

other parent, producing 16 zygotic combinations. In the

diagram the four classes of gametes from one parent are

given in the horizontal row at the top, and the same four

classes from the other parent in the vertical row at the

left. Each gametic combination from the top is repeated

four times in the squares below, while each combination

at the side occurs four times in the corresponding hori-

zontal row of squares. Thus each of the squares repre-

sents the result obtained by combining the gametes repre-

senting the horizontal and vertical rows that intersect at

that point. In all cases where both S and X occur together

the seed should be horny, where only S occours the-seed

should be sweet, when only X occurs it.should be waxy,

and in one square (No. 16), where neither S nor X occurs

there is a new combination which the results have shown

to be a new type of sweet seed, indistinguishable from

ordinary sweet seed but behaving differently when crossed

with other types of endosperm.

In accordance with the above analysis the expected re-

sults in the third xenia generation were as follows:

Proportion- Proportions

f Seed

ate No. of

Ears. Classes.

Self-pollinated horny.

i 1 All horny

2 3 horny: 1 sweet

2 3 horny: 1 waxy

4 9 horny: 4 sweet: 3 waxy

Self-pollinated sweet. :

All sweet

Self-pollinated waxy.

1 All waxy

2 3 waxy: 1 sweet

No. 574] INHERITANCE 587

Crosses between different plants

eds.

from horny se

25 All horny

20 3 horny: 1 sweet

20 3 horny: 1 waxy

16 9 horny: 4 sweet: 3 waxy

Crosses between different plants

from sweet seeds.

All sweet

Crosses between different plants

from waxy seeds

5 All waxy

4 3 waxy: 1 sweet

Crosses between horny and sweet.

3 All horny

6 1 horny: 1 sweet

1 1 horny: 1 waxy

2 3 horny: 1 waxy

2 1 horny: 2 sweet: 1 waxy

? 4 3 horny: 4 sweet: 1 waxy

Crosses between horny and waxy.

5 All horn

4 3 horny: 1 sweet

10 1 horny: 1 waxy

8 3 horny: 2 sweet: 3 waxy

Crosses between sweet and waxy.

All horny

All waxy

Ae bo po bd HS HY

bat

1 horny: 2 sweet: 1 waxy

THIRD XENIA GENERATION

Four of the ears bearing second xenia generation seed

were selected for planting in 1913, one self- and one cross-

pollinated ear from each of the two hybrid families Dh 216

and Dh 221. These families were selected because in 1913

the family Dh 221 showed the greatest deficiency of sweet

seeds and Dh 216 was the only family that showed sweet

seeds in excess of the expected. The three classes of

seeds from each of the ears were planted separately.

Unfortunately as the result of an accident crosses were

not made between the plants grown from the different

classes, but a total of 77 selfed ears were obtained, a num-

588 THE AMERICAN NATURALIST [Vou. XLVIII

ber sufficient to indicate whether the initial assumption |

regarding the gametic compositions was of value in

arranging the observed facts.

PROGENY or SWEET SEEDS

Sweet seeds were assumed to result from squares 6, 8,

14 and 16. It will be seen that in none of these is there

any factor other than S and since the absence of both

factors, as in square 16, is also assumed to produce sweet,

we should expect nothing but all sweet ears from self-

pollinated plants grown from sweet seeds.

Seventeen self-pollinated ears were secured from plants

grown from sweet seeds. All the seeds of these ears were

sweet with the exception of one waxy seed. This one waxy

seed was colored and since it occurred on an ear from a

white sweet seed that otherwise produced only white

sweet seeds, the exception may reasonably be ascribed to

accidental foreign pollen.

Progeny or Waxy SEEDS

Waxy seeds were assumed to have resulted from the

combinations shown in squares 11, 12 and 15. Seeds from

square 11 should produce only waxy seeds. Squares 12

and 15 should produce ears with waxy and sweet seeds

in proportion of 3 waxy to 1 sweet. There should, there-

fore, be one all waxy ear to two with both waxy and sweet

seeds. There were in all 29 ears from waxy seeds, 11 of

which were all waxy and 18 with both waxy and sweet

seeds. The numbers are small but at least both kinds of

ears were secured and the proportion does not violate the

original assumption. The 18 ears with both waxy and

sweet seeds all produced them in approximately the 3:1-

ratio. The numbers are given in Table I. The totals with

3,154 seeds indicate that if there is a deviation, it is almost

certainly less than 2 per cent.

All the sweet seeds that occur on ears grown from waxy

seeds are assumed to belong to the new class of sweet

seeds corresponding to that represented in square 16.

No. 574] INHERITANCE 589

Plantings of such seeds are being made for comparison

with the ordinary class of sweet seeds having the same

ancestry. These are represented by the sweet seeds occur-

ring on ears having horny and sweet seeds.

TABLE I

WAXY SEEDS SELF-POLLINATED, EARS SHOWING WAXY AND SWEET SEEDS.

EXPECTED: 25 PER CENT. SWEET

P tE | Pedigree a rey | ho Bab | Per Cent. of | Devia-

a r | n | is t

hi okt ti | Number is | Secd, | Seeds | Sweet Seeds Ereb.

1938 | 301| 216| 85 | 28.241.7 | +1.9

1939 | 112] 85| 27 +27 |= 3

- | {|1940 | 264! 202) 62 | 23.5 +18 |- 8

aay eee {1942 | 18| 14| 4 | 222466 | — 14

| [1948 | 349| 258| 91 | 261416 | + .7

1949 | 149| 100 49 | 329426 +3.0

(1950 | 138| 103| 35 | 254425 | - 2

| {1972 | g0 87 | 224414 | -19

973 | 187| 136, 51 | 27.3422 | +1.0

othe | |1974 | 174| 138| 36 | 2072421 | —2.0

, | 4 1975 | | 18 | 212430 | -13

penal EE | [1976 | 34] 2t) 13 | 382456 | +24

“| $1977 | 813} 232]. 81 | 2692417 | + 5

| (1978 | 109|- 79| 30 | 29 + 9

| (1904 | 136| 105| 31 | 228424 | -— 9

Dh 221-2 | $1905 |. 155| 116| 39 | 25.2423 | + .1

(Self-Pollinated) | } 1996 51] 31) 2% | 39.7 £46 | +3.1

| (1997 190 | 146 | 44 | 22.2421 | — 9

| Total... 3,154 | 2,351, 803 255+ 5 +1.0

Progeny oF Horny SEEDS

From the horny seeds the expected results are more

complicated. They may be tabulated as follows:

1 ear (Square 1) with seeds all horny

2 ears (Squares 2 and 5) with seeds 3 horny: 1 sweet

2 ears (Squares 3 and 9) with seeds 3 horny: 1 w

4 ears (Squares 4, 7, 10 and 13) with seeds 9 horny: 4 sweet: 3 waxy.

Ears were, therefore, expected in the proportion of 1 all

horny ear, 2 with horny and sweet seeds, 2 with horny

and waxy seeds and 4 with all three classes. Thirty ears

were secured from seed classed as horny. These ears

were distributed as follows: 1 all horny, 5 with horny and

590 THE AMERICAN NATURALIST [Vou XLVIII

sweet seeds, 3 with horny and waxy seeds, 19 with horny,

sweet and waxy seeds and 2 all sweet.

The two all sweet ears are entirely outside the expected.

Their appearance may be explained on the assumption

that seeds classed as horny in 1912 were in reality sweet.

No microscopical examination of the starch was made

and the seeds .were classified on their appearance,

wrinkled seeds being classed as sweet and smooth seeds

as horny. The separation of horny from sweet seeds is

more difficult to make than waxy from either horny or

sweet.’

There were, however, very few doubtful seeds in the

second xenia generation and in suggesting this interpre-

tation, we may with some propriety be accused of attempt-

ing to explain away ‘‘green balls.’’ 3

The two all sweet ears were descendants of an ear Dh

221-2, which showed an excess of horny seeds and a

deficiency of sweet. The expected number of sweet seeds

in Dh 221-2, which had a total of 493 seeds, was 123 and

only 106 were classified as sweet. If this deviation re-

sulted from a faulty classification, that is, if some of the

sweet seeds failed to show the characteristic wrinkled

exterior, we might expect that about 17 of the 300 seeds

classed as horny would produce ears with all sweet seeds.

Eleven of the ears secured from horny seeds in 1913 were

descendents of this ear.

The remaining 28 ears from horny seeds are distributed

among the 3 classes in reasonably close agreement to the

expected. Measured by Pearson’s formula for the good-

ness of fit, it appears that such a deviation might be

expected once in about twenty times.

2The difficulty of distinguishing between sweet and starchy seeds in

crosses where the starchy variety has small seeds has been pointed out by

East and Hays, ‘‘Inheritance in Maize,’’ Bull. 167, Conn. Ag. Exp. Sta.,

1911, p. 40.

3 Pearson, K., and Heron, D., ‘‘On Theories of Association,’’ Biometrika,

IX, pp. 309-314,

4 Phil. Mag., Vol. L, 1900, pp. 157-175. The application of Pearson’s

formula to data of this kind was called to our attention by Mr. G. Udney

Yule.

No. 574] INHERITANCE 591

The three ears with horny and waxy seeds produced

these classes in the expected 3:1 ratio. The numbers are

given in Table II.

TABLE II

Horny SEEDS SELF-POLLINATED. Ears SHOWING HORNY AND WAXY SEEDS.

EXPECTED: 25 PER p ai, WAXY

Per Cent. of tion +

: x Total R No | No.

Pedigree | N ‘Horny | Waxy

Parent E

arent Kar Number Moke | Beedle | Sead | Waxy Seeds Pian

h 216-2 | | |

(Self-Pollinated) 1962 | 327 | 247 | 80 | M6456 | 3

Dh 221-2 2000 | 312 | 285 | 77 | 24.7216 2

(Self-Pollinated) 2007 | 121 | 82 39 | 322429 25

Total...| 760 | 564 | 196 | 258411 | 7

Four of the five ears that produced horny and sweet

seeds were also as close as could be expected to the 3:1

ratio. The fifth, however, Ped. 1965, with 249 seeds, had

only 19 sweet seeds or 7.6 per cent. The numbers are

given in Table III. The only explanation that can be

TABLE III

HORNY SEEDS SELF-POLLINATED. EARS SHOWING HORNY AND SWEET SEEDs.

EXPECTED: 25 PER CENT. SWEET

Devia-

| Total | No. | No. Per Cent. of | Devia

reer | ee ay ay | rr Bro

Dh 216-1 |

(Cross-Pollinated) 1965 | 249) 230 | w 76 #L1 | 16.0

Dh 216-2

(Self-Pollinated) 1979 | 442| 344 | 98 | 222413 | 22

Dh 221-1 |

(Cross-Pollinated) 1988 | 160| 121 39 | 244423 .3

221-2 2003 175 | 184 | ü | 28.442.1 8

(Self-Pollinated) 2008 179 141 38 21.2 + 2.1 1.8

Total. ..| 1,205 | 970 | 235 | 195+ 8 | 6.9

offered in connection with this exceptional ear is that

suggested for the occurrence of the two all sweet ears

among those grown from seeds classed as horny, namely,

the existence of sweet seeds which failed to show a

592 THE AMERICAN NATURALIST [Vou XLVIII

wrinkled surface. This explanation is rendered less prob-

able, however, by the unusual behavior of the aleurone

color in this same ear. In the previous discussion the

aleurone color has not been considered. To treat of the

aleurone color would naturally lead to the question of

correlation between that character and endosperm texture,

a subject which in these crosses is very complicated and

for the treatment of which the results thus far obtained

are inadequate. It may be said, however, that with the

exception of Ped. 1965 the proportions of colored to white

seeds in all the ears bear out the assumption that the in-

heritance of the aleurone color is governed by two factors,

both of which must be present to produce color. In Ped.

1965, however, which was grown from a colored seed, only

23 of the 249 seeds were white. The colored and white

seeds are beautifully distinct with no intermediate or

doubtful seeds. The ratio of 9.2 per cent. white might be

- explained as an approximation to the dihybrid ratio of

6.25 per cent. but we must then admit that instead of both

factors being necessary for the development of color

either factor alone may produce color.

The 19 ears from horny seeds that showed all three

classes are assumed to have the same gametic composi-

tion as the original second xenia generation, previously

reported. The numbers are given in Table IV. The last

column of the table gives the odds in 1,000 that deviations

equal to those observed are not chance deviations from the

expected proportions, as calculated by Pearson’s formula.

Thus in Pedigree 1953 the odds are 809 to 191, or practi-

cally 4 to 1, that the deviation is not the result of chance.

As in the original ears, the approximation is sufficiently

close to render futile any attempt to predicate a different

arrangement of factors, but many of the deviations are

too large to be ascribed to chance. In the totals the sweet

class is too low and the waxy too high, in fact there is no

significant difference between the totals for these two

classes. The deviation from the expected is, however,

No. 574] INHERITANCE 593

largely the result of two ears Ped. 1954 and 1967, and if

the explanation suggested for the two all sweet ears from

horny seeds is admitted, it may also account for the devia-

tion in these two ears. In both ears the deficiency of

sweet seeds is accompanied by an excess of horny seeds,

while in neither ear is there a significant excess of waxy

seeds.

TABLE IV

HORNY SEEDS SELF-POLLINATED. EARS SHOWING ALL THREE CLASSES.

EXPECTED; 56.25 PER CENT, Horny, 25 Ses CENT. SWEET

18.75 PER CENT. W

i : nees

Horny Seeds Sweet Seeds Waxy Seeds pe 1000

Pedi- | Total ` ‘ in evia

Parent .Ear No. |

g Seeds

197 | 236 87 36 66 | 78 999+

1955) 198! 111| 103| 50 37 | 40 |470

1956 42| 17 | 11 | 13 16 |777

B gs hang 1957| 148) g3| 74| 37| 42 | 28| 32

Pollinated) ) 1958| -176| , 99| 100 41 35 |133

963) 540 289| 135 | 133 | 101 | 118 | 832

1 70 4| 18 13 | 18 | 983

1966 170 101| 42 | 35 | 32 | 34 | 524

1967, 158. s9| 108| 39 | 23 | 30 27 |995

Dh 216-2 |

Pollinated)| 1980 35; 20| 16| 9 |. 12 7 | 8 |486

Dh 221-1 1985) 258| 145 131| 65 73 | 48 | 54 |784

Cra. 1986, 120) 67| 72, 30| 25 | 23| 23 |443

Pollinated) | 1987) 375 211| 201| 94 | 92 | 70 82 |716

1909] 77) 43| 42| w] 16 | .14 | 19 |676

Dh 221-2 oh oF <b as) eee wt 8 8

s 2002| 118| 66| 71| 30| 29 | 22 | 18 |426

Pollinated)| | 2004! 134) 75|) 81| 34| 29| 25 | 24 |455

2009| 79) 44| 48| 20 | 10 | 15 21 |979

Total 3,141| 1,767 | 1,800 | 785 | 681 | 589 | 660 | 999.99

With these two ears excluded the deviation in the total

for the remaining 17 ears may be ascribed to chance.

Tested by Pearson’s formula such deviations might be

expected once in about 50 times.

594 THE AMERICAN NATURALIST [Vou. XLVIII

CONCLUSIONS

The immediate (xenia) result of crossing varieties of

maize having sweet and waxy endosperm was the produc-

tion of seeds with a horny endosperm resembling that of

ordinary field varieties. In the second xenia generation

all the ears contained seeds of the three classes, sweet,

waxy and horny, in fairly definite ratios. The data were

arranged in accordance with the Mendelian formula corre-

sponding most nearly to the observed numbers.

The third generation, like the second, gave results suff-

ciently close to dihybrid ratios to render unprofitable the

assumption of more complicated ratios. There are, how-

ever, deviations from the expected numbers of too great

magnitude to be ascribed to chance.

The ratios of waxy to non-waxy seeds were regular as

far as the conditions of the experiment could determine,

except for a slight excess in the number of waxy seeds in

nearly all the ears in which all three classes appeared

(Table IV). A deviation in number of waxy seeds as

large as that shown in the total would not be expected to

occur as the result of chance more often than once in one

thousand times.

The ratios between sweet and horny, while approxi-

mating the predicted ratios, show numerous irregularities.

Wherever there is a significant deviation in the number

of sweet seeds, the observed number is below the expected.

Reasons are advanced for believing that the deficiency of

the sweet class may result from a failure of some sweet

seeds to develop a wrinkled exterior rather than from any

irregularities in segregation.

The results show the value of representing the char-

acters by gametic factors. This method provides an

orderly arrangement of the facts of heredity thus far

observed with respect to these characters and makes pos-

sible fairly accurate predictions regarding the genetic be-

havior of the various seed classes.

WASHINETON, D. C.,

July, 1914

A STUDY OF VARIATION IN THE APPLE

W. J. YOUNG

Assistant HORTICULTURIST; WASHINGTON EXPERIMENT STATION

As a rule the subject of variation in the several char-

acters of the apple has been given but incidental attention,

and that usually in connection with the study of other

problems. As a result the literature on the subject is of

a fragmentary character consisting usually of a few ob-

servations here and there in papers dealing with other

subjects.

It is perhaps worth while to note a few of the investi-

gations which have thrown some light in an incidental

way upon the causes of variation in apples. In fertilizer

tests which were made at the New York Geneva station!

and elsewhere, no well-defined and uniform influence of

the various elements of plant food upon the color could

be detected, though the New York station reports more

decided results in seasons when the natural conditions

were unfavorable to the development of highly colored

fruit. In the comparison of tillage and sod mulch in an

apple orchard, also conducted by the New York Geneva

station,? it was found that the fruit from an orchard in

sod was more highly colored and matured one to three

weeks earlier than that from the tilled plot, though the

latter was better in quality and kept four weeks longer in

common storage. The influence of the stock upon the

character of the fruit is a matter of much obscurity, the

investigation of which presents such difficulties that it

has received little attention. The effect of pollination

also is still far from settled. It was thought at one time

that the characters of the fruit were profoundly modified

by the pollen received by the blossom. The data on this

1 Bull. 289.

2 Bull. 314.

595

596 THE AMERICAN NATURALIST [Vou. XLVIII

subject, have been collected by Munson,? who found that

evidence that the pollen has any direct effect upon the

fruit is largely lacking. Aside, then, from indirectly

modifying the size of the fruit, the influence of the pollen,

in so far as our present knowledge goes, may be left out

of account in a study of apple variation.

Without doubt the most noteworthy contributions to

the knowledge of apple variation are the recent papers

by Shaw, of the Massachusetts station, and Stewart of the

Pennsylvania station. Shaw’s first paper, which ap-

peared in the Massachusetts station report for 1910, deals

entirely with the variation of the Ben Davis apple. In

comparing specimens grown in a number of widely sepa-

rated localities it was noted that variations due to cli-

matic condition were strongly marked and affected prac-

tically all characters of the fruit. Modifications of form

were especially noticeable The depth of coloration was

looked upon as correlated with latitude, being pink in the

specimens from Arkansas and deep crimson in those

grown farther north. The amount of overeolor seemed

to be controlled by local conditions. The color was espe-

cially good in the apples from the Pacific coast and those

from Colorado, Pennsylvania, and Indiana. In a given

orchard temperature appears to be the most influential

factor governing size. The flesh was notably white in the

fruit from Colorado. The apples from Colorado and

California were less firm than those from other localities.

The southern-grown specimens were more juicy and of

better quality than those from the north, which were apt

to be dry, hard, flat, and sometimes astringent. It ap-

pears that a mean temperature of at least 60° F. for the

growing season is required for the satisfactory produc-

tion of the Ben Davis. The poor quality of the northern-

grown specimens is apparently due to a lack of sufficient

heat to properly develop the fruit.

In Shaw’s second paper in the Massachusetts station

report for 1911 the fact is emphasized that the grower

3 Me. Sta. Rept. (1892), pp. 29-32.

No. 574] VARIATION IN THE APPLE 597

should choose those varieties which he can grow to the

highest degree of perfection under his conditions of soil

and climate. The causes of variation are summarized,

giving special attention to the influence of temperature as

a factor in the distribution of apple varieties. The north-

ern limit is regarded as fixed by the lowest temperature

which the tree will stand, while the effect of summer

heat upon the development of the fruit is looked upon as

limiting the distribution southward. The elongation

of the fruit was found to be correlated with a low tem-

perature for two or three weeks after blooming. A low

summer temperature produces greater acidity, higher

content of insoluble solids, greater astringency, smaller

size, and scalding in storage. The extent of coloration

was regarded as decreasing from the center of distribu-

tion in passing either north or south, while the intensity of

coloration was considered greatest in high latitudes and

altitudes. Excessive summer heat results in uneven

ripening, premature dropping, rotting on the tree, poor -

keeping quality, lack of flavor, mealiness, less intense

color, and smaller size. For each variety there is a mean

summer temperature at which it reaches its highest de-

velopment.

It will be noted that Shaw’s method of investigating

the problem consisted in securing fruit for comparison

from widely separated localities and attempting to corre-

late the various characters with the conditions of produc-

tion. Stewart, on the contrary, confined his study to

apples grown in one locality and noted the effect of modi-

fying one at a time those factors within his control. This

is the more scientific method of procedure, but has the dis-

advantage that the variations are far less striking and a

smaller number of factors can be studied. An account of

Stewart’s experiments and the results so far attained is

found in the reports of the Pennsylvania station since

1907. These papers deal largely with the effect of fer-

tilizers and different cultural methods on the yield, color,

size and growth of the apple. The various factors influ-

598 THE AMERICAN NATURALIST [Vou XLVIII

encing these characters are enumerated and the results

are given of the studies made of them. It is noted that

the factors are so interrelated that the best conditions for

producing one effect are often injurious in some other

direction and that the chief problem in orchard manage-

ment is a proper balance of the various factors. An

‘¢ optimum principle ’’ is recognized, according to which

plant growth and development increase as the most dis-

tant essential factors approach the optimum. The factor

farthest from the optimum, therefore, whether below ot

above, may control the results from a crop

OUTLINE OF THE EXPERIMENT

Since the season of 1912 was one of full.crop in nearly

all centers of apple production, conditions were especially

favorable for the study of variation in this fruit. The

writer accordingly obtained specimens for study and com-

parison grown in a number of localities under quite dis-

similar conditions. The method employed was therefore

that of Shaw, as pointed out in the last paragraph, rather

than that of Stewart. The study has been pretty largely

confined to Washington-grown apples, though a few have

been obtained for purposes of comparison from the east

and middle west. The formal investigation of the prob-

lem has been carried on but a single season, which is en-

tirely too brief a study to demonstrate conclusively all

points touched upon. The conclusions reached, however,

are strongly supported by many observations in various

localities extending over a number of seasons, and are so

suggestive of further lines of study as to justify a report

at this time.

In carrying on this investigation the aim has been to

secure as much information as possible regarding the con-

ditions under which the fruit was grown. The endeavor

has been to get into communication with the growers and

obtain from them through correspondence data regarding

the character of the soil, rainfall, irrigation, elevation,

exposure, temperature, age of trees, fertilization and

No. 574] VARIATION IN THE APPLE 599

cultivation. The chief line of observation had to do with

the variations which occur in the different samples of the

same variety as obtained from different sources. To get

at this side of the problem, careful observations were

made as to the condition of the apples, and their various

characters were recorded in a complete technical descrip-

tion of each sample for the purpose of making a compara-

tive study of the samples of the several varieties. In ad-

dition to this written description, photographs were made

showing typical specimens in various positions and when

cut in cross and longitudinal sections. In general it may

be said that variations are found in the form, size, color,

internal structure, texture, flavor, quality, specific grav-

ity, chemical composition, time of ripening, and keeping

quality. The attempt is made to correlate these char-

acters with the conditions of growth in so far as they are

known and to work out the law of the relation of environ-

mental factors to the characters of the fruit.

The following apples were made use of in the study:

Arkansas, one sample; Arkansas Black, one sample;

Baldwin, eight samples; Ben Davis, nine samples; Deli-

cious, three samples; Esopus, seven samples; Gano, seven

samples; Grimes, seven samples; Jonathan, eight sam-

ples; Lawver, one sample; McIntosh, one sample; North-

ern Spy, seven samples; Rhode Island Greening, three

samples; Rome, eleven samples; Stayman, five samples;

Tompkins King, four samples; Wagener, six samples;

White Peamain, three samples; Willow, one sample;

Winesap, ten samples; Winter Banana, one sample; Yel-

low Bellflower, four samples; Yellow Newtown, seven

samples; and York Imperial, two samples, making a total

of 117 samples embracing 24 varieties. These apples

were obtained from fourteen localities in the state of

Washington and also from one locality in each of the fol-

lowing states: New Hampshire, Indiana, Missouri, New

York, and West Virginia.

Before leaving the preliminary portion of this paper

the writer wishes to express his appreciation of the aid

600 THE AMERICAN NATURALIST [Vou. XLVIII

received from those who have helped in various ways in

the investigation. Thanks are due to the members of the

staff of the department of horticulture for suggestions

and encouragement, to the members of the library staff

who have rendered aid in the study of the literature of the

subject, to Mr. Geo. A. Olson, chemist of the experiment

station, who has analyzed the various samples of Grimes,

Jonathan, Yellow Bellflower, and Winesap, and finally

to the various fruit growers and others who have cooper-

ated in securing the fruit and have furnished notes on the

conditions of production. To all these the writer takes

pleasure in acknowledging his gratitude and indebted-

ness.

ENVIRONMENTAL FACTORS

Aside from small individual differences, better called

fluctuations than variations, and other more striking

modifications of comparatively infrequent occurrence

and obscure origin, which it is customary to explain as

bud variations, if, indeed, the application of a name to a

phenomenon can pass as an explanation, it is quite gener-

ally recognized that variation in any variety of fruit is

due to the operation of external influences. A knowledge

of the various factors which make up the environment

and their influence upon plant life is necessary to an in-

telligent study of variation. It should be noted, however,

that this influence is not necessarily the same with plants

propagated vegetatively as with those grown from seed.

In the latter case certain modifications of an adaptive

nature which enable the plant to fit in more perfectly with

modifications tend to disappear by the elimination of the

individuals possessing them. In the former case, on the

other hand, the modifications observed are the direct re-

sult of the conditions, unaffected by selection, and whether

desirable or not they persist as long as the environment is

unchanged and the vegetative propagation is continued,

unless, indeed, the environment is so unfavorable that the

No. 574] VARIATION IN THE APPLE 601

changes induced are pathological in nature and the plant

can not survive.

Perhaps the most important factor to which plant life

is subjected is the moisture relation. This may be deter.

mined by the amount of moisture actually present or by

the modifying influence of other coexistent factors which

interfere with the availability of the moisture and the

capacity of the plant to make use of it. Among such in-

direct influences may be noted the modifying effect of

temperature upon the rate of absorption and transfer of

moisture, the presence in the soil of certain salts or humic

acids which interfere with the osmotic activity of the

roots, and certain atmospheric conditions favorable to

rapid transpiration. In such cases care is necessary to

determine which is the direct and which the indirect cause

of the modifications. If it is borne in mind that many

factors cause variation through their influence on the

moisture supply confusion may often be avoided.

The temperature relation is much more obscure than

the moisture relation in its effect upon plant growth.

Heat, being a molecular phenomenon, acts directly upon

the protoplasm and its effects are therefore physiological.

It is now pretty well understood that heat alone is in-

capable of modifying plant structure, but acts indirectly

through other factors and the functions of the plant. The

direct effect of temperature is limited very largely to its

influence upon the rate and amount of development. A

slight difference in the average temperature of the grow-

ing season influences greatly the relative development of

apple varieties. The accompanying table gives the mean

monthly temperature during the growing season at

Geneva, N. Y., and Pullman, Wash., since the establish-

ment of the érperiment stations at those points, as well as

the mean for two years at White Salmon, Wash.

Locality. April | y | Jano July | Aug. | Sept. | Oct. | Aver.

Geneva, N. Y........ aa | 87. i a | 69 | 63 | 50 | 60

Pullman, Wash....... 47 52 | 59 | 66 58 | 48 57

White Salmon, Wash..| 50 | 56 | 63 70 | 67 60 | 53 60

602 THE AMERICAN NATURALIST | Vou. XLVIII

The difference in the development of certain varieties

of apples at these places will be noted later. It will be

seen that the season opens slightly earlier in Pullman

than in Geneva and closes at about the same time. It

would appear, therefore, that the better development of

most varieties at the latter station is due rather to the

higher temperature than to a difference in the length of

season. At White Salmon the season is considerably

longer than at either of the other stations, while the tem-

perature from May to September is intermediate.

Latitude and altitude are frequently mentioned as im-

portant factors in the modification of varieties. These,

however, are not primarily factors, but depend for their

influence upon the effect of other factors, which in turn are

influenced by the location. Differences in altitude espe-

cially result in marked changes in climate often in places

geographically near together.

The light relation is of much importance to the fruit

grower. It is clearly evident that the development of

color in apples is largely dependent upon the sunshine,

and quality also may be affected through the production

of sugars. Both intensity of insolation and duration of

the daylight must receive consideration. In general, trop-

ical, arid or alpine situations are characterized by high

insolation, while a long period of daylight during the sum-

mer months is a factor in northern latitudes.

The effect of atmospheric influences is largely indirect.

It has already been noted that the condition of the air

may modify the moisture relation through its effect upon

transpiration, thus dryness, high temperature, and rar-

ification all favor evaportion, and this effect may be in-

creased in windy situations. Atmospheric pressure is a

factor of importance in high altitudes.

The soil may be of importance as a factor in causing

variation through either its chemical composition or its

physical properties. The former leads to a consideration

of the influence of fertilization, the latter to the effect of

different methods of culture. Here again other factors,

No. 574] VARIATION IN THE APPLE 603

and especially the moisture relation, have an important

bearing, since one of the primary results of cultivation is

the conservation of the soil moisture. There is no doubt

that the nature of the soil greatly affects the crop and the

matter has been given much study. The intimate associa-

tion of other factors, however, makes it somewhat difficult

to pick out those influences for which the nature of the

soil is directly responsible.

The influence of other organisms includes not only a

consideration of the effect of insect and fungus pests but

in the broad sense embraces such items as pollination,

pruning and thinning, intercrops, cover-crops and plant-

ing distance. Human agencies, including all operations

of orchard management, might properly be included here.

Many of these are, of course, indirect, exerting an influ-

ence through their effect upon some other factor.

THe Law OF THE OPTIMUM

Having enumerated the chief external influences to

which plants are subjected during their period of develop-

ment and to which variation is largely due, the question

naturally occurs whether there can be formulated any

basic principle or law which will express the manner in

which plants react with the environment. Such a law

would be of use not only in the study of variation, but

would shed much light on the adaptation of plants to new

environments. It would constitute a unifying principle

whereby isolated facts and disconnected observations ap-

pear in proper relation and perspective. Though a dis-

cussion of this subject might logically be delayed until

after the characters of the several varieties and their

modifications have been noted, it is thought most fitting to

introduce the statement at this point and examine the

fruit in the light of such generalizations as it has been

possible to make.

A plant can live and perform its functions only within

certain intensities of the various factors of the environ-

ment. The degrees of intensity beyond which activity

604 THE AMERICAN NATURALIST [Vou. XLVIII

ceases are known as the zero points. The plant does not

necessarily die at once, but passes into a dormant state.

If the intensity becomes still more unfavorable a point is

finally reached at which death occurs. The minimum de-

gree of intensity of a factor at which the plant may re-

main active is known as the lower zero point, while the

greatest intensity is called the upper zero point. With

some factors these points are wide apart, so that, other

conditions being favorable, the plant will continue to de-

velop after a fashion at any but the most extreme intensi-

ties of such factors. With other factors the limits are

comparatively narrow. A plant will reach that degree of

development only which is permitted by that factor which

is in the least favorable degree of intensity. Such factors

are called limiting factors. In passing from one zero

point toward the other, a point is finally reached at which

any given function of a plant reaches its highest state of

activity. This point is known as the absolute optimum for

that function and may not correspond to the most favor-

able intensity of that factor for the performance of the

other functions of the plant. The point of intensity of a

factor at which all the functions of the plant are per-

formed to the best advantage is termed the harmonic

optimum. If each factor is of an intensity corresponding

to the harmonic optimum, the plant is in a condition of

equilibrium known as the ecological optimum and will

reach the highest state of activity of which it is capable.*

As the life of a plant is made up of various functions,

so its structure is made up of a number of organs having

various characters. These characters are the result of |

development, which in turn is dependent upon the per-

formance of the several functions of the plant under the

influence of those external conditions which make up the

environment. If a factor of the environment is modified

in its intensity, the balance of the functions of the plant

is disturbed and the plant reacts to its changed environ-

ment by a modification of its functions which may result -

4 Schimper, A. F. W., ‘‘Plant Geography.’

No. 574] VARIATION IN THE APPLE 605

in a different kind of development, or in other words a

variation. Having observed the close connection between

the characters and the functions of the plant, we may now

inquire whether the former maintain a relation to the

environment similar to that maintained by the latter.

Putting aside generalizations for the present and confin-

ing attention to the apple, it is to be noted that both Shaw

and Stewart foreshadowed such a relationship in the

papers already noted. Neither, however, carried the

analysis far enough to formulate a rule of general appli-

cation, though Stewart came near doing so. Shaw recog-

nized that the highest perfection in any given variety

could be attained only under the most favorable summer

temperature. Stewart applied this idea to other factors

than temperature in his ‘‘optimum principle,’’ which is

‘ that plant growth and development increase as the most

distant essential factors approach the optimum.’’ His

failure to recognize the connection between the various

factors of the environment, on the one hand, and the sepa-

rate characters of the apple, on the other, may be ac-

counted for by the fact that his investigations dealt only

with fruit grown under slightly modified conditions, which

resulted only in such slight variations that the independent

modification of the separate characters escaped notice.

In examining various samples of apples produced under

the influence of quite dissimilar combinations of environ-

mental factors, the writer has many times noted the modi-

fication of certain characters more or less independently

of others. It is true that characters are often found to

vary together through a relationship of direct or inverse

correlation. Such cases, however, are possibly as often

due to the response of the various characters to the same

factor of environment as to any direct connection between

the characters, though the latter no doubt exists in many

cases. Keeping in mind these facts and also the close

relationship of function and character, the writer has

formulated a principle which he believes is of general ap-

plication not only to apples but to other horticultural

606 THE AMERICAN NATURALIST [Vou. XLVIII

crops and perhaps in a degree to all plant life. For this

generalization, which expresses the relationship of char-

acters to environmental factors the name ‘‘ Law of the

Optimum ’’ is proposed.

This law may be stated as follows: For any given

variety there is for each character a certain intensity of

each essential factor of the environment at which, other

conditions remaining the same, that character reaches its

highest development. When all essential factors are in a

condition of optimum intensity for any character, that

character will reach the most perfect development of

which it is capable. A modification of the intensity of any

such factor either above or below the optimum will be ac-

companied by a less perfect condition of the character

concerned. The optimum intensity of a factor may be

wide or narrow in its limits and the optimum for one char-

acter may or may not overlap the optimum for others.

A variety will be at its best when grown in an environ-

ment the factors of which are as near as may be to the

optimum intensity for all characters. Under such cir-

cumstances the variety is in a state of balanced adapta

tion to its environment. If removed from such an en-

vironment to one in which certain factors are distant from

this state of average optimum intensity for all characters,

the equilibrium is destroyed and the variety is thrown

into a state of unbalanced adaptation, in which those

characters farthest removed from their respective optima

are injuriously affected, while others may be bettered by

being placed in a combination of factors of an intensity

nearer their optima. A discussion of the practical appli-

cation of this law and its bearing upon apple culture in

the northwest will be deferred for the present and taken

up in a later section.

A COMPARATIVE STUDY OF THE SAMPLES

A close study of the various lots of apples used in this

experiment brings to light variations in practically all

characters. Many, however, are modifications of charac-

No. 574] VARIATION IN THE APPLE 607

ters inconspicuous in themselves or are slight in amount

and so do not attract attention. A complete account of all

variations noted would comprise a full technical descrip-

tion of each sample which would far exceed the limits of

this paper. For this reason it is thought best to append

only some brief comparative notes regarding the more

conspicuous variations noted in each variety. In this

connection it is well to note the origin so far as known of

the varieties included in this study. Arkansas and Arkan-

sas Black, Arkansas; Baldwin, Massachusetts; Ben

Davis, probably Kentucky or Tennessee; Delicious, Iowa;

Esopus, New York; Gano, probably Kentucky or Mis-

souri; Grimes, West Virginia; Jonathan, New York;

Lawver, possibly Kansas; McIntosh, Ontario, Canada;

Northern Spy, New York; Rhode Island Greening, Rhode

Island; Rome, Ohio; Stayman, Kansas; Tompkins King,

New York; Wagener, New York; White Pearmain, prob-

ably Eastern States; Willow, Virginia; Winesap, New

Jersey; Winter Banana, Indiana; Yellow Bellflower, New

Jersey; Yellow Newtown, New York; York Imperial,

Pennsylvania. It will be observed that all originated in

the east or middle west. Most no doubt appeared as seed-

lings and were selected and propagated because of their

excellence and value when grown under those conditions

of environment which prevail at their places of origin; in

other words they were individuals which happened to be

in a condition of balanced adaptation to that environment.

Their behavior under other environments could be deter-

mined only by actual tests, and some notes on the subject

are included in the following paragraphs.

Arkansas (Mammoth Black Twig).—As only one sam-

ple of this variety was examined its behavior can be com-

pared only with what is known of the variety in other

localities. The fruit was more elongated and conical in

shape, smaller in size and less highly colored than that

produced in the warmer apple-growing sections of the

east. The flesh was inferior in texture, indicating poor

development. The variety seems not at all adapted to

608 THE AMERICAN NATURALIST [Vou. XLVII

the location where grown, but might do better at lower

altitudes and in warmer situations in the state. Never-

theless, the quality is not good enough to recommend the

variety for dessert, and it is to be hoped that it will not be

planted extensively in the northwest. The keeping qual-

ity was excellent. ;

Arkansas Black—This variety of the Winesap group

attains a deeper color than the Winesap and equals that

variety in size and quality. The specimens examined

were not especially well colored though, it is known to

color well in the irrigated valleys. It seems to be better

adapted to the conditions of the state than the Arkansas.

In keeping quality it was among the best.

Baldwin.—The Baldwin attains its: highest perfection

in New York and New England, where it is a great favor-

ite in the markets and is produced more largely than any

other variety. As grown in this state the fruit is smaller

and more elongated than the eastern product and has a

more deeply furrowed basin. As grown at Pullman the

color lacks intensity, though the fruit.is well covered. In

the western part of the state the fruit is well colored,

especially in the northern part of the Puget Sound Basin.

The lots from White Salmon show a good many poorly

colored fruits mixed with those of better color, while the

quality is rather better than in those examined from other

parts of the state. It is, however, inferior to the eastern-

grown Baldwin and is evidently poorly adapted to the

conditions of the northwest. All of the Washington-

grown fruit displayed a tendency to wilt in storage and

some of the lots from the western part of the state rotted

seriously as a result of fungous infections not apparent on

the fruit at the time of storage.

Ben Davis.—Though displaying considerable lack of

balance in the adaptation of the different characters to

conditions in certain parts of the state, this variety seems

on the whole to reach a good degree of development in the

warmer valleys. In quality the lot from Missouri was

superior to those from any part of Washington, though

N .574] VARIATION IN THE APPLE 609

many of the Washington-grown apples of the variety were

equal to those from most sections of the east. Striking

variations in form were displayed by the fruit from dif-

ferent localities. Those lots from the more elevated and

cooler sections of the state were of an oblong, conic form

and usually had shallow irregular basins, while those

from the warm valleys were less elongated and had deep

and usually quite regular basins, being more like the fruit

from the Ben Davis belt of the east. The fruit developed

better texture and quality also in the valleys though it was

coarser and more spongy than the eastern fruit. Most of

the Washington grown samples of Ben Davis were more

decidedly striped than those from the east. This effect is

produced by the clearer yellow ground color, which in the

eastern-grown fruit is more or less suffused with red.

The apples from the elevated localities of Pullman,

Cloverland and White Salmon were relatively small in

size and poorly colored.. Because of its low dessert qual-

ity, the planting of this variety for shipment to the east

can not be recommended. The most desirable feature of

the Ben Davis fruit is its good keeping quality. A tend-

ency to mealiness late in the season was observed in

some of the fruit from the irrigated valleys, while those

grown at Pullman and Cloverland wilted badly toward

the close of the season.

Delicious.—This is one of the newer varieties and when >

well grown is a dessert apple of fine appearance and high

quality. In many of its characters, but especially in flavor

and aroma, Delicious resembles the White Pearmain,

though in color it bears a likeness to the Winesap group.

In moderately elevated situations in some parts of the

state it displays a well-balanced adaptation and attains

excellent size, color, texture and quality, though none of

those examined were quite equal in quality to the Deli-

cious from New York. When grown in too low and warm

a location the fruit has a tendency to become overripe and

when stored tends to soften in the center, after which it

loses greatly in quality. The sample from Clarkston had

610 THE AMERICAN NATURALIST [Vou XLVIII

a beautiful dark red color, while that from Cloverland

was dull in color and poor in texture.

Esopus (Spitzenburg).—This is almost the only variety

which the writer has examined that attains the first rank

as a dessert apple in this state. In certain sections it dis-

plays a better balance of adaptation so far as flesh charac-

ters are concerned than any other variety. The samples

obtained from White Salmon and the irrigated valleys

were of excellent quality as dessert apples, though of

scarcely as good texture as the variety attains in the east.

Overgrown apples are especially coarse in texture. West

of the Cascades and in the more elevated locations the

Esopus does not reach as high quality as elsewhere. This

variety is inclined to wilt in storage unless well grown.

Gano.—This is an apple of the Ben Davis type, but of a

‘more uniform red color. Practically all the remarks in-

cluded under Ben Davis, aside from those dealing with the

distribution of color, apply equally well to the Gano. At

its best, the Gano is of slightly better quality than the Ben

Davis, which fact, together with its more handsome ap-

pearance, renders it a more desirable variety to plant, yet

neither can be recommended in a section desirous of build-

ing up a reputation and market for dessert apples. It is

interesting that both the highest color and the best as well

as the poorest quality was attained by apples from the

east and middle west.

Grimes (Grimes Golden).—This variety, like the Ben

Davis, displays considerable variation in form, depending

on the locality of production. The specimens from the

middle west were roundish to decidedly oblate, while those

grown in Washington were all more or less elongated.

Those grown west of the Cascades displayed a greater

tendency to a conical shape than those from the eastern

part of the state, and were also poorer in quality. When

grown in the more elevated sections, as at Pullman,

Grimes appears poorly developed and immature and is

inferior in size and quality. Those from Grandview dis-

played the best balance of characters and it seems prob-

No. 574] VARIATION IN THE APPLE 611

able that this variety is better adapted to the irrigated

valleys than to other sections of the state. All samples

were more or less wilted by midwinter, except the fruit

from Grandview, which remained firm but showed some

tendency to rot. Scald was very bad in the latter part of

the season.

Jonathan.—Although rather extensively grown in a

number of localities in Washington, none of the fruit

which the writer has examined gave evidence of a well-

balanced adaptation to the conditions of growth which

prevail in the state. All were inferior in color to the fruit

_ obtained from the east and middle west. The apples from

Clarkston and the Yakima Valley were of good size but

lacked both richness of flavor and aroma. The same lack

was evident in the fruit from the western part of the state.

At Pullman a pretty good quality is attained, but the

_ fruit does not come up to the requirements as to size and

gives other evidence of imperfect development. At

Cloverland and in other elevated locations fruit of a poor

texture and deficient coloring is produced. Jonathan

seems to reach its highest development in certain sections

tributary to the Ohio valley and the Washington-grown

Jonathans can not compete with fruit from that section

when well grown. The samples from Morgantown, West

Virginia, were of a beautiful clear dark red color, good

size, fine tender flesh, and very high quality. In storage

these specimens remained firm and retained their flavor

until April. The others wilted considerably after mid-

winter.

Lawver.—tThis variety attains good size and fine color

‘in the irrigated valleys, but the quality is not good enough

to recommend it to the fruit growers of the northwest.

The variety ordinarily keeps well but the specimens

stored proved to have poor keeping quality—owing to

fungous infection.

McIntosh_—The McIntosh is deserving of attention as a

variety of high quality which appears to have a fairly

well-balanced adaptation to certain sections of the north-

612 THE AMERICAN NATURALIST [Vou XLVIII

west. At Pullman the elevation is too great for the best

development of the variety, but the Spokane Valley pro-

~ duces MeIntoshes of a high degree of excellence. There

is good reason to believe that the valley of the northern `

and northeastern sections of the state can rival the Bitter

Root valley of Montana in the production of this variety.

The fruit stored wilted badly by midwinter and lost much

of its flavor soon after.

Northern Spy.—Of all the varieties examined the

Northern Spy seems least adapted to the conditions of

growth in this state. As produced in New York and New

England this fruit is a dessert apple of the highest quality

when well grown and properly colored. In Washington

east of the Cascades the color fails to develop and the

quality is much inferior to that of the eastern-grown fruit.

In the western part of the state the color develops as well

as in the eastern states, but the quality is no better than .

elsewhere in the state. The unsurpassed cooking quality

of this variety seems to be largely retained, however,

which is its only redeeming feature. It may be worth

planting to a limited extent as a culinary fruit for home

use, but can not compete in the markets with the eastern-

grown Northern Spys. The specimens from the western

part of the state were largely infected with fungi, result-

ing in much decay early in the season. Those from Pull-

man and Clarkston kept fairly well, though the former

wilted badly late in the season.

Rhode Island Greening.—This variety, together with

Baldwin and Northern Spy, constitutes the most promi-

nent and successful apples in the orchards of New York

and New England. They are also among the varieties least’

adapted to the conditions found in this state. Their per-

fect balance of adaptation to eastern conditions is prob-

ably to a large degree responsible for their popularity in

the east and may also account for the lack of balance

which they display in the northwest. As grown at White

Salmon and at Pullman the Greening reached a good size,

but was decidedly inferior in quality to the specimens

No. 574] VARIATION IN THE APPLE 613

from New Hampshire. At Pullman the fruit was rather

flat and strongly ribbed, while at White Salmon the apples

were oblong in shape and had, as a rule, rather small

cavities. It can not be recommended for Washington,

except possibly for local use as a culinary fruit. This

variety is a fairly good keeper. Those grown at Pullman

wilted badly late in the season, while the lot from White

Salmon gave evidence of considerable fungus infection.

Rome (Rome Beauty).—This is one of the most popu-

lar varieties grown in the state east of the Cascade Moun-

tains and is about the only commercial variety which

reaches good marketable size in the high uplands of the

Inland Empire. The Rome reaches its highest develop-

ment in the Jonathan belt of the middle west. The best

specimens examined, all characters considered, came from

Morgantown, West Virginia. They were of a nearly uni-

form deep red color, of good size and attractive form, and

of pretty good quality for the variety. In many parts of

Washington the Rome fails to color well. The specimens

from White Salmon and Grandview were especially poor

in color. The latter were overgrown and of poor quality,

while the former were among the best of the variety. The

usual form of the variety is round or nearly so, varying

to somewhat roundish conic or roundish ovate. The form

of the cavity is subject to quite a little variation. As pro-

duced at Pullman and other elevated sections of the state

the cavity is very shallow, but becomes deeper in the val-

leys. The specimens from West Virginia had fairly deep

cavities. Indeed it seems probable that those localities

which produce Ben Davis of the elongated type also pro-

duce Romes with the shallow cavities. The Rome is by

nature a culinary apple. In quality it is but little better

than Ben Davis. It seems unfortunate, therefore, for the

lasting reputation of the industry, that it should have be-

come so firmly established in northwestern horticulture.

It is to be earnestly hoped that it may in time be replaced

by a variety of better quality. In its adaptations to the

conditions of the state, the Rome seems to be fairly well

614 THE AMERICAN NATURALIST [Vou. XLVIII

balanced in most of its characters. The balance, however,

is not the same in all sections and is nowhere quite so per-

fect as in certain localities in the middle states. Most

samples kept well until the latter part of the season and

then became mealy. The overgrown specimens from

Grandview were the first to break down in this way.

Those grown at a greater elevation showed a slight tend-

ency to wilt late in the season. None of the samples dis-

played an inclination to rot until late in the season.

Stayman Winesap.—In both size and quality the Stay-

man is the best of the Winesap group. Its most serious

fault is a rather dull color which often fails to cover the

fruit well. The samples obtained from the middle west

were of better color and texture than those grown in

Washington, though the lot from Indiana were very coarse

in texture. Those grown at Pullman were small and in-

ferior in every way. The fruit from Grandview was

especially large, flat, and fairly well colored, while that

from White Salmon was more elongated, slightly less

colored, and rather more aromatic in flavor. These two

lots retained their firmness in storage much longer than

the others and those from White Salmon scalded badly

late in the season. It is very similar to the Winesap in

its adaptations.

Tompkins King.—This variety is popular in the west-

ern part of the state, where it attains a large size and good

color, though the latter character develops well at Pull-

man. None of the samples equaled in quality the variety

as grown in New York. Those grown at Pullman had a

very good flavor, though the flesh characters were those of

poorly matured fruit. The fruit from the western part of

the state was of a fairly elongated conic form, while that

grown at Pullman was shorter and strongly ribbed. This

variety appears to be but poorly adapted to Washington

conditions. The fruit grown at Pullman wilted badly late

in the season, while that from western Washington rotted

considerably owing to fungus infections.

Wagener—Though of the Northern Spy class, the

No. 574] VARIATION IN THE APPLE 615

Wagener displays a much better balance of adaptation to

the conditions of the state than the Northern Spy. It

seems to reach its best development in the cooler regions

of the state. The specimens from Grandview were of

good size and very juicy, but were poor in color, coarse

in texture, and deficient in flavor. Wagener develops

especially well in the Spokane Valley. The specimens

from Opportunity were large, well colored, and of excel-

lent quality, though somewhat coarse in texture. Those

grown at Pullman were more aromatic but possibly not so

rich in flavor and did not develop sufficient size. This

variety does well west of the Cascades and especially in

the northern part of the Puget Sound Basin. The speci-

mens from Eastsound were large, highly colored, and fine

in texture, but less aromatic than the eastern Washington

fruit. The samples obtained from West Virginia gave

evidence of having been grown too far south. They were

poorly colored and of rather poor texture, but of good size

and excellent flavor. In form the fruit from Opportunity

was roundish, that from Eastsound roundish conic, while

the remainder was decidedly flattened and all samples

were more or less strongly ribbed. This variety shows

very little tendency to wilt in storage. The fruit from the

highlands keeps well, but that from the irrigated valleys

shows a tendency to physiological decay. Scald is serious

after midwinter. `

White Pearmain (White Winter Pearmain).—In gen-

eral appearance this variety often closely resembles the

Yellow Newtown, but is usually more elongated and more

largely blushed. Moreover, it is quite different in flavor

and is remarkable for its fine aroma. It is a variety of

high quality and attractive for a yellow apple, moreover,

it attains its good qualities in the irrigated valleys better

than on the highlands, the specimens from Cloverland be-

ing dull and green in color and poor in texture, but well

blushed and highly aromatic. Its worst fault is suscepti-

bility to the apple scab. It would seem to be better

adapted to growing in the state than some of the more

616 THE AMERICAN NATURALIST [Vou. XLVIII

popular varieties. The fruit from the Yakima Valley

retained its firmness much better than that from Clover-

land, but lost somewhat in flavor toward the close of the

season.

Willow (Willow-Twig).—The writer has examined this

variety only as grown in the elevated portions of eastern

Washington. In such locations it does not develop espe-

cially well in either size or color and is of too poor quality

to be worthy of consideration. Moreover, it wilts badly

in storage, though when well grown the fruit has excellent

keeping quality. It is evidently poorly adapted to this

section.

Winesap.—In some of the irrigated valleys this variety

is one of the most popular apples grown. It attains a

good marketable size and an attractive color, though none

of the samples examined were equal in color or quality to

the Winesaps from Indiana and West Virginia. In ele-

vated localities, as at Pullman, Cloverland and White

Salmon, the fruit is small and poorly colored and has flesh

characters indicating imperfect development and matur-

ity. As grown in the irrigated valleys the fruit is apt to

be deficient in flavor, and, if large, coarse in texture. The

lot from Cashmere showed the best balance of characters

of any Washington, grown specimens, but these were in no

way superior to the Winesaps from West Virginia. It is

probable that the better grown fruit from the eastern

Winesap districts is equal to that grown in Washington in

all respects, with the possible exception of size, which, if

large, is, as noted, apt to be accompanied by deterioration

in quality. It is evident then, that the balance of adapta-

tion of this variety to northwestern conditions is imper-

fect at best and that the planting of Winesaps in Wash-

ington may easily be overdone. This variety proved to be

one of the best in keeping quality. Those from Pullman

and Cloverland wilted late in the season, though most of

the other lots were in excellent condition in April and a

few were held in storage until July.

Winter Banana.—As only a single lot of this variety

No. 574] VARIATION IN THE APPLE 617

was examined in detail, it is difficult to make very positive

statements regarding its behavior in the state. Though

less desirable than a red apple, it is a variety of handsome

appearance and is fairly good in quality. It is perhaps

rather better adapted than the averageto certain sections

of the state and appears to develop best in fairly elevated

situations. It is especially well liked in the Spokane Val-

ley, and fruit grown there is said to have good keeping

quality, though the specimens from western Washington

were past season by midwinter. They wilted badly and

showed much scald.

Yellow Bellflower—This variety appears to be better

adapted to the western part of the state than to the irri-

gated valleys. The apples from Clarkston were coarser

in texture, milder in flavor and poorer in quality than

the samples received from the east. There were no very

striking differences in form, structure or appearance ex-

cept that the eastern Bellflowers were more often blushed

than those from Clarkston. The apples from Puyallup

were overgrown specimens from young trees, were coarse

and spongy in texture, and inferior in quality. As this is

a tender fruit, easily injured by careless handling, and

does not appear to be especially well balanced in its

adaptations, it is not desirable to plant extensively for

shipping. Moreover, it is not a good keeper. The speci-

mens from Puyallup were practically past season when

received and those obtained from the east were more or

less injured and such specimens decayed quickly. Some

of the lot from Clarkston, however, kept sound and firm

until past midwinter, but deteriorated in flavor toward the

last.

Yellow Newtown.—When at its best, this variety has

few equals. It is narrow in the limits of its adaptations

and its successful culture in the eastern states is confined

to small areas, where, however, it is in nearly perfect

equilibrium with its environment. In many places in the

northwest it is grown successfully, though it scarcely

equals in quality the best eastern product. The fruit

from White Salmon and some of the irrigated districts

618 THE AMERICAN NATURALIST [Vou XLVII

was of excellent quality, but coarser and less delicate in

texture and of not quite so good flavor as the apples from

West Virginia. The specimens from Cloverland were

hard and green and gave evidence of imperfect maturity.

Evidently the elevation is too great for its proper devel-

opment. The single sample from western Washington

consisted of well-colored, extensively blushed fruit, but

was inferior in quality. Owing to its limited area of suc-

cessful production in the east, it is worth planting in

Washington wherever its characters give evidence of a

fair degree of balance of adaptation with the environ-

ment. This variety is perhaps a better keeper than Wine-

sap. Some of the fruit from White Salmon kept in good

condition until July, though overgrown fruit and that

which has been exposed to heat before storage showed

signs of physiological decay late in the season. Under-

developed specimens wilted in storage.

York Imperial_—In sections of Virginia and neighbor-

ing states the York Imperial occupies the place of su-

premacy held by the Baldwin farther north. This is

doubtless due to its perfect balance with the environmental

conditions of that region, and, like the Baldwin and other

sorts perfectly adapted to their eastern habitat, this

variety finds itself out of equilibrium when moved to the

northwest. The apples from western Washington were

of good size and color, but were coarse and undesirable

in texture and poor in quality. The specimens grown

at Pullman were smaller, more elongated, and less com-

pressed than the others, and the axes were less oblique.

They were somewhat better in quality, though not good

enough to justify more extensive planting. The fruit

wilted in storage, and that from western Washington

gave evidence of fungous infection and scalded badly after

midwinter.

DISCUSSION OF THE EFFECT OF ENVIRONMENT Upon APPLE

CHARACTERS

Size—Size is the direct result of development. An

apple will reach its maximum in growth when all factors

No. 574] VARIATION IN THE APPLE 619

are at the variety optimum for the physiological proc-

esses upon which development depends. A departure

from this optimum, whether toward a greater or less in-

tensity, means a decrease in size, as is observed in ap-

proaching either the northern or southern range of a

variety. It has been frequently noted, however, that the

optimum for growth is not the best combination of fact-

ors for the development of certain other desirable char-

acters, so that it is well to choose an environment having

certain factors in a somewhat less degree of intensity,

being content with fruit of fair size but superior in other

respects. Since the apple contains about 85 per cent. of

moisture it is evident that the water supply is a factor of

prime importance in determining size. It is possible by

excessive irrigation to force an abnormal growth of the

fruit, though always apparently at the expense of text-

ure, flavor, and keeping quality. It is evident, then, that

if fruit of good quality is expected, irrigation must be

moderate in amount, especially with vigorous young

trees. Thinning may result in increased size owing to

the larger amount of moisture available for each fruit.

Temperature and length of season are of importance in

determining, respectively, the rapidity of growth and de-

gree of development attained.

Form.—One of the striking features revealed by the ~

study of a number of varieties from several localities is

the fact that the modification in shape due to the differ-

ence in environment is by no means uniform for the

several varieties. Some varieties are quite constant in

shape while others are much more plastic in this respect.

Moreover, certain varieties are much more easily in-

fluenced than others which respond in the same way,

while still others respond differently to the same factors.

One of the most frequently observed and conspicuous

modifications of form consists of the elongation of the

axis of the fruit relative to the horizontal diameter. This

character has been especially studied, in the case of the

Ben Davis, by Shaw, who found the elongation most

noticeable in fruit from the northeastern states, the mari-

620 THE AMERICAN NATURALIST [Vov. XLVIII

time provinces of Canada, and the Pacific coast. Shaw’s

papers dealing with this subject have already been noted.

Upon studying the climate in these localities, it was found

that the temperature for two or three weeks after the

blooming season was notably lower than in the sections

where the Ben Davis assumes its normal shape. Since

this appeared to be the only factor constant for the

several localities, it is suggested as the explanation of

this variation. It has been shown, however, that temper-

ature is incapable of influencing form except by its action

through the functions of the plant in modifying the effect

of some other factor. It is the writer’s opinion that the

elongation is due to the relative moisture supply of the

different parts of the apple at this period of develop-

ment as influenced by the temperature ; that it is primar-

ily a modification due to the moisture relation rather

than to the direct effect of temperature, the latter being

a secondary cause. The rapidity of circulation of the sap

and therefore the supply of moisture to the organs of the

plant is greatly influenced by the temperature. It is a

well-known fact of plant physiology that much less moist-

ure passes through the plant in the cool days of spring

than during the warmer weather of midsummer. A re-

duction of the temperature at this time results in a still

more sluggish movement of the sap. In the period im-

mediately after blooming the energy of the plant, so far

as the development of the fruit is concerned, is directed

primarily to the proper nourishment of the growing seeds

and the adjacent parts. If at this time the circulation

of the sap is retarded by a temperature unwontedly low

for the variety, the moisture supply of the fruit is

lessened and a relatively larger amount goes to the seeds

and adjacent parts, while the pulpy portion of the fruit

receives a more scant supply. As a result, the axillary

development is proportionately greater than the swelling

of the fruit due to the accumulation of moisture in the

superficial tissues. After some two or three weeks the

form of the fruit becomes fixed and is not noticably in-

fluenced by the moisture supply thereafter.

No. 574] VARIATION IN THE APPLE 621

The elongation of the fruit is usually accompanied by

a constriction of the apex resulting in a conical form.

This may be due to the greater development of the basal

portion, which is adjacent to the point where the sap

enters the fruit and may therefore be better supplied,

though the physiology of fruit development is in need of

further study. In the Grimes, however, an oblong form

results. The McIntosh, as grown at Pullman, is often

decidedly obovate, a variation which the writer ascribes

to the same influences that produce the elongated conic

form of the Ben Davis and other varieties, though in this

. variety the response is somewhat different. The Rhode

Island Greening, Willow and Wagener, as a rule, fail to

assume an elongated form in localities where it is well

marked in some other varieties. Also in certain varie-

ties which are naturally conic in form and considerably

elongated, as Delicious and Yellow Bellflower, this effect

is not evident. The larger number of varieties, when

grown in this state, have a more ribbed form than the

same varieties in the east. This seems to be due to a lack

of balance in adaptation, though the particular factor

which gives rise to the variation has not been determined.

Some varieties, like the York Imperial and the Yellow

Newtown, are compressed in form, that is elliptical in

section, and have an oblique axis when grown in certain

environments. These characters seem to be in some way

related to the better development of the fruit, as they are

less evident in fruit from the elevated and unfavorable

sections of the state. Beach has noted in the ‘‘ Apples of

New York’’ a similar difference between the Newtowns

of western New York and those of the Hudson Valley, the

latter having a more oblique axis and elliptical form.

Stem.—The stem is one of the most variable structures

of the apple, and, owing to the fact that stems of different

lengths, diameters and shapes are commonly found in

any lot of apples grown under practically uniform con-

ditions, it is difficult to associate such variations with the

environment. The writer has noted, however, in the case

of some short-stemmed varieties, like the York Imperial,

622 THE AMERICAN NATURALIST [Vou. XLVIII

that those lots grown under less favorable conditions had,

on the average, longer stems than others grown under a

more favorable environment.

Cavity.—The most conspicuous variation in the cavity

is inits depth. This is of especial note in the Rome, which

has a very shallow cavity in most parts of the state. This

is doubtless due to the same cause which produces the

elongated form of the fruit in many varieties, namely the

elongation of the axis resulting from a deficient moisture

supply incident to a low temperature after the blooming

season. In this variety the elongated axis obliterates the

cavity instead of modifying the general outline of the

fruit. The same variation is also noted to a less degree

in a number of other varieties. An especially furrowed

cavity is often observed associated as a rule with the

ribbed form of fruit.

Calyx.—The writer has failed to observe any modifica-

tions of importance in the calyx lobes of the fruit. The

size of the calyx cup or ‘‘ eye ”’ of the apple is influenced

by the development of the fruit. In large fruit this open-

ing is apt to be large, so that the lobes are separated, re-

sulting in an open or partly open calyx. Small or poorly

developed apples, on the other hand, usually have the

calyx closed.

Basin.—The depth of the basin seems to depend upon

the same factors as that of the cavity and seems to be

much more readily influenced than the latter. The width

is often associated with the form of the apple, a very con- °

stricted apex resulting in a narrow basin. A much fur-

rowed basin results from a combination of factors un-

favorable to the best development of the fruit.

Skin Statements have often appeared in regard to

the effect of various climatic factors upon the thickness

and toughness of the skin. Estimates of these characters,

however, appear to be based entirely upon sense impres-

sions of the observers, although it would seem that exact

measurements would not be especially difficult. In the

absence of such accurate data, an expression of opinion

No. 574] VARIATION IN THE APPLE 623

would be premature. Dry air and sunshine are favorable

to the production of clear, smooth skin.

Color.—There seems to be no doubt that the coloration

of apples depends upon the influence of several factors of

which light is usually the most important. The impor-

tance of light is easily demonstrated by covering the fruit

during development either wholly or in part. The in-

tensity of illumination is also, evidently, quite narrow in

its limits, so that a point is soon reached at which the

color begins to pale owing to excess of illumination. It

has been frequently noted that apples grown near the

southern limit of the range of a variety are paler than

those grown farther to the north. This effect appears to

be the result of an excess of the two factors, heat and

light. It has been mentioned in the discussion of the

characters of several varieties that, contrary to the gen-

eral impression, those grown in Washington east of the

Cascades where insolation is intense were less highly

colored than those from western Washington or the east-

ern states. The most marked example of this kind which

the writer has observed is the Northern Spy. Again,

contrary to the general impression, most of the samples

from elevated locations were poorly colored, a fact which

may be attributed partly to the strong insolation and

partly to the poor development due to the low summer

temperature. It appears, therefore, that either too

strong or too weak illumination may result in poorly

colored fruit and that the best color is developed under a

condition of optimum intensity of the light.

It is suggested above that temperature may influ-

ence color. This is most commonly observed in the

case of apples grown under conditions of too

high summer temperature, though a deterioration

in color also results if the temperature is much

below the optimum for the variety. It is often

stated that apples become more highly colored the farther

north they are grown. This is only true in part. Those

varieties which are adapted to the most northerly por-

tions of the apple belt are able to develop their highest

624 THE AMERICAN NATURALIST [Vou XLVIII

color at the limit of winter hardiness of the tree. The

southern varieties, on the other hand, require for the best

development of color a higher summer temperature than

is experienced in the northern localities. The Winesap,

for example, when grown in Central New York is partly

covered with a pale red. At Pullman the majority of

varieties color poorly, due at least in part to the cool

climate. That the temperature and not the shortness of

the season is the factor involved is shown by the fact that

most of these varieties color well in central New York

which has a season of about the same length though

averaging several degrees warmer.

Cultural conditions may influence the color to a cer-

tain degree. In general those processes of orchard man-

agement which favor the early maturity of the fruit re-

sult in improved color, especially in localities having a

short growing season. Pruning and wide planting are

regarded as favoring high coloration by admitting light

into the tree, though it is possible that in regions where

the light is intense these factors may not be of so great

importance in their effect upon color as in less sunny loca-

tions. Sométhing has been said of the influence of the

soil in the discussion of the literature and it has been

noted also that studies of the effect of fertilizers upon the

color have not yielded satisfactory or uniform results.

The influence of iron compounds is worthy of brief dis-

cussion in this connection. It seems evident, from the

chemical studies which have been made, that the red pig-

ment includes iron in its composition. This has some-

times been assumed to mean. that the chief requirement

for highly colored fruit is the presence of plenty of avail-

able iron compounds in the soil. As a matter of fact,

iron is also necessary to the formation of chlorophyll and

most soils contain an abundance of that element for the

purpose. From the chemical data compiled by Stewart®

it appears that the ash of the fruit contains a much

smaller proportion of iron than that of the leaves. Itis |

logical to conclude, therefore, that soils containing suff-

5 Pa. Sta. Rept. for 1910-11,

No. 574] VARIATION IN THE APPLE ' 625

cient iron for the development of chlorophyll in the leaves

are also fully supplied for the formation of the red pig-

ment of the apple.

Internal Structure. —The form and relative develop-

ment of the core and associated structures are subject to

numerous variations, which, however, are seldom so con-

spicuous as to attract attention unless closely studied,

and appear to be of little practical importance to either

the grower or consumer of the fruit. The number of

seeds may be mentioned as an indication of the thorough-

ness of cross pollination and in most varieties the pres-

ence of one or more well developed seeds is a requisite

to the proper development of the fruit. Small or poorly

developed fruit, the result of too short a season or too

low a temperature, is apt to have the core closed and axile,

or nearly so, while in the same varieties good develop-

ment is usually associated with a more open abaxile core.

The carpels of such poorly developed fruit are usually

entire and smooth, while those of the better-grown fruit

are more or less cleft and often tufted.

Flesh Characters.—From the standpoint of the con-

sumer, these are by all odds the most important charac-

ters of the fruit, though lost sight of through the empha-

sis placed on external characters, and no grower who

has at heart the permanent prosperity, extension and

normal development of the industry can afford to look

upon quality as a secondary consideration. Neglect in

this matter is sure to result sooner or later in a bad repu-

tation for the fruit among a considerable proportion of

buyers, which appearance and advertising will not be

competent to overcome. The fact can not be denied that

the great majority of varieties fail to attain as high

quality in the northwest as when grown in the eastern or

middle states where nearly all of them originated, while

at the same time they may excel in other important char-

acters. This is especially true of most of the choice

dessert apples. Such unequal development can have no

other interpretation than that these varieties are in a

` state of unbalanced adaptation to the environment. This

626 THE AMERICAN NATURALIST [Vou XLVIII

fact being recognized, the main question is, How can this

disadvantage be overcome? Evidently the solution does

not consist in a steadfast refusal to face the situation

and vehement declaration that the fruit of any particular

district is the best that can be produced. Such tactics,

though well meant, can be permanently successful only

when the statements are justified by the facts. If apple

culture in Washington is to be maintained upon a sound

basis it will be necessary first of all that growers shall

exercise great care in planting to choose those varieties

most nearly in equilibrium with the environment in the

various sections of the state, at the same time avoiding

over-irrigation or other errors in orchard management

which may tend to an unequal development of the char-

acters of the fruit, usually at the expense of quality.

Even this, however, may be but a temporary makeshift,

since few if any of the better varieties possess the re-

quisite power of adaptation. It will be necessary first of

all to determine if the variations which appear when

apples are grown from seed in the northwest are more

favorable in character than those which are displayed by

introduced varieties. If such should prove to be the case

the writer is under the conviction that the apple culture

of the northwest should ultimately be largely made over

on a basis of new varieties of local origin. A number of

such varieties have already appeared, but unfortunately

some of them have been chosen with little regard for

quality. No work of greater value to the future horti-

culture of the region can be undertaken by the experiment

stations of the northwestern states than the development

of apple varieties of high quality and perfect adaptation

to the various sections of their respective states.

The apples of high quality which show a fair degree of

adaptation to the irrigated sections are Esopus, Yellow

Newtown, Delicious and White Pearmain. The last was

found by Lewis, of the Oregon station, to be one of the

best pollenizers on every variety tested. Jonathan,

Winesap and Stayman, though largely grown, shows in

general a poorer balance of characters. In the more

No. 574] VARIATION IN THE APPLE 627

elevated valleys Wagener, Delicious and McIntosh are

doubtless most worthy of culture. The highlands of

eastern Washington are very poorly adapted to the grow-

ing of winter apples, though some of the early apples do

fairly well, among which may be mentioned Oldenburg,

Gravenstein and Yellow Transparent. On account of

the abundance of sunshine the Oldenburg develops a high

sugar content for the variety which counteracts its natural

acidity and results in an apple of pretty good dessert

quality. Of the winter apples, Rome reaches good

marketable size but the quality is not high and the east-

ern market should not be jeopardized by shipping this

variety. The Palouse, an apple of local origin, is of

much better quality, but has little standing in the market

as yet. The Dutch Migonne, a variety from western

Europe, shows a better balance of characters in eastern

Washington than in most other sections of this country.

It is of good size, fairly well colored and excellent in

quality.

Many varieties popular in the eastern states color

better west of the Cascades than in eastern Washington,

though there is usually manifest a lack of balance in

other characters. In certain respects the environment

resembles that of western Europe and many of the va-

rieties of cherries, plums, prunes, and other fruits of

that country do very well here and, indeed, in other sec-

tions of the state as well, though in a number of instances

varieties of northwestern origin are gaining in favor

rapidly. Apple breeding, however, requires nfore time

for its accomplishment and further importations of

fruits, especially apples, adapted to the mild climate of

western Europe would no doubt prove an advantage

through the possible discovery of sorts adapted espe-

cially to the western part of the state. ?

Quality is not in itself a simple character. It depends

upon all the characters of the flesh which determine the

desirability of the fruit for eating, such as texture, juici-

ness, aroma and flavor. Fineness of texture evidently

depends upon a proper combination of favorable factors.

628 THE AMERICAN NATURALIST [Vou XLVIII

Conditions favoring rank growth result in coarse texture,

as was observed in several instances in the case of apples

grown under irrigation, especially if the fruit was over-

grown. Some of the fruit from young trees also was

overgrown and coarse. Tenderness depends upon the de-

velopment. Poorly grown, under-developed fruit grown

where the temperature is too low or the season too short

for the variety has hard flesh which becomes spongy

rather than mellow toward the end of the storage season.

Overgrown fruit of certain varieties, on the other hand,

often shows lack of coherence between the cells, often ac-

companied apparently by larger intercellular spaces, and

such fruit tends to become mealy as the season pro-

gresses. Juiciness is primarily a manifestation of the

amount of moisture in the fruit, but is also associated

with the tenderness of the cell walls and their tendency

to break rather than to separate. In general an abun-

dance of moisture results in juicy fruit though the juici-

ness is not in proportion to the moisture supply. The sub-

stances which give the apple its aroma are present in

such small amounts that their investigation is difficult.

They are volatile compounds and affect the flavor of the

apple largely by their action on the sense of smell. A

cool climate is favorable to their production and it was

often observed that they were most strongly developed in

the apples from elevated situations. Flavor depends

upon the kinds, amounts and relative proportions of the

soluble solids, especially the balance between sugars and

acids, antl will be given further consideration in the dis-

cussion of the chemical composition. Immature and

under-developed apples contain some tannic acid, which

is often sufficient in amount to give an astringent charac-

ter to the fruit.

Keeping Quality.—In its relation to the environment,

keeping quality evidently follows the same rule as other

variable characters of the apple, namely, that for any

variety the keeping quality depends upon the optimum

intensity of the various external factors. Apples grown

where the temperature is too low or the season too short

No. 574] VARIATION IN THE APPLE . 629

to develop the fruit to a proper stage to keep well, soon

wilt, lose flavor and scald, or show other evidence of de-

terioration as was frequently observed in the fruit from

high altitudes. On the other hand, too great excess of

certain factors results in overgrown or overripe fruit

having a tendency to rot, mealiness, or physiological de-

cay, as in the case of the Yellow Bellflowers from Puyal-

lup and some of the fruit from the warm valleys. - The

balance of factors favorable to good keeping quality does

not appear to differ much from that which produces the

fruit which is most desirable in other characters, though

it is possible that the required intensity of some factors

may be slightly lower. It appears, therefore, that a good

balance of the other characters of the fruit and perfect

adaptation to the environment will be accompanied, as a

rule, by good keeping quality, provided that the fruit is

properly handled and not infected with disease, while an

unbalanced adaptation of characters to environment is:

likely to result in poor keeping quality. It seems prob-

able that. irrigation in itself does not result in poor keep-

ing except when improperly applied or carried to excess:

or associated with other factors in such a way as to de-

stroy the equilibrium of the environment. The relation:

of specific gravity to the keeping quality is discussed in

a succeeding paragraph.

Specific Gravity—It has long been understood that

varieties of apples differ in their relative weights; thus

Wolf River is comparatively light and Baldwin is gener-

ally regarded as a heavy apple. The only record found

of the determination of specific gravity of apples is that

of Howard’s work in the National Bureau of Chemistry,

Bulletin 94, in which it is noted that the specifie gravity

diminished 3 per cent. to 5 per cent. during storage.

From the account it is not clear whether the determina-

tions at the different dates were made with the same

apples. The decrease of specific gravity is ascribed to the

increase of air spaces between the cells due to the soften-

ing of the middle lamella. In the specific gravity determi-

nations made by the writer a number of points was noted.

630 THE AMERICAN NATURALIST [Vou. XLVIII

The different lots of a variety may differ considerably in

specific gravity, though as a rule running somewhat close

together, thus Ben Davis and Gano are apples of low spe-

cific gravity, while Grimes, Stayman, Wagener, and Yellow

Newtown run rather high and Baldwin and Rome may be

classed as medium in this respect. Overgrown apples

were low in specific gravity, probably owing to more air

space between the cells. This is more apparent upon

examining the results for individual apples than upon

comparing the average for different lots, as in the latter

case the extremes are modified by averaging with the re-

sults for more normal specimens. On the other hand,

small and rather undeveloped apples are apt to have a

high specific gravity on account of their solid flesh and

usually closed core. Juicy apples, if not overgrown, have

a high specific gravity when the juiciness is due to a high

moisture content.

The relation of specific gravity to keeping quality is of

interest. While some late keeping varieties have nor-

mally a low specific gravity, those lots of a given variety

having a high specific gravity for the variety are usually

the best keepers. This is in line with the fact that certain

causes which give rise to fruit of poor keeping quality

also produce a low specific gravity. This is shown very

strikingly by a comparison of the specific gravities as cal-

culated month by month through the season. As the

ealeulations were made at the time the fruit was found

fit for use, the monthly averages show the steady increase

in specific gravity with the better keeping quality of the

fruit, though modified somewhat by the peculiarities of

the different varieties which happened to be in season at

different times. These averages are as follows: Novem-

ber and December, 0.787; January, 0.787; February,

0.810; March, 0.831; April, 0.852. Though these results

may seem to be at variance with Howard’s observations

it is possible that if the same specimens had been tested

at intervals a decrease in specific gravity would have

been noted.

Chemical Composition.—In order to throw some light,

No. 574] VARIATION IN THE APPLE 631

if possible, upon the relation of chemical composition to

the other characters of the apple and to determine

whether the composition is influenced by the environment,

the juice of the various samples of Grimes, Jonathan,

Winesap and Yellow Bellflower was analyzed by the de-

partment of chemistry.

The juice of the Grimes and Winesap contains, as a

rule, a decidedly higher percentage of total solids than

that of the Jonathan and Yellow Bellflower. It is also

generally higher in specific gravity and has a greater

viscosity. In Grimes and Yellow Bellflower the juice of

the eastern-grown fruit contains a large proportion of

total solids than that of the Washington grown fruit,

though this rule does not hold good in the other varieties.

The apples from the irrigated valleys and western Wash-

ington were low in total solids with the single exception

of the Winesaps from Cashmere. The analyses fail to

show any constant difference in sugar content in favor of

the fruit produced in the sunny climate with long hours

of daylight characteristic of the apple-growing sections

of the state.

In Grimes the total sugars are fairly high and the pro-

portion of sucrose is especially large. The acid content,

on the other hand, is low as a rule. The result is a rich,

mild or nearly sweet flavor. A sample from Puyallup

showed the lowest sucrose content combined with the

highest acid content, and this was the least rich as well as

the most acid in flavor.

Jonathan, on the other hand, displays a low content of

total sugars and especially sucrose, while the acid content

is slightly higher than in Grimes, indicating a subacid

apple, lacking in richness. The lots from Missouri and

Indiana were highest in sucrose but were of scarcely as

good quality as the Jonathans from West Virginia. The

latter were low in both sucrose and acid, but displayed a

good balance between these constituents, indicating an

apple with rather thin juice, not very rich, but pleasant

and refreshing. Its evident superiority resulted largely

632 THE AMERICAN NATURALIST [Vou XLVIII

from the fine texture and well-developed flavoring con-

stituents not shown by the analysis.

The Winesaps, though high in total sugars, are low in

sucrose, indicating a heavy juice rather lacking in rich-

ness. The comparatively high acid content corresponds

to the sprightly subacid character of the fruit. The high-

est acid content was found in the fruit from Cloverland,

where it is associated with a total lack of sucrose result-

ing in a comparatively poor fruit. The apples from

Cashmere and White Salmon were also devoid of sucrose

in the juice, but the acid content was low and the flavor-

ing principles well developed, as a result of which the

quality was fairly good. The poorly developed Winesaps

grown at Pullman were deficient in sucrose, acid, and

flavors and were correspondingly poor in quality.

The Yellow Bellflowers, though low in total sugars,

were rather high in sucrose and also in acid. The bal-

ance between these constituents is good and results in a

moderately rich, pleasant, subacid flavor.

SuMMARY

The opportunity for the study of apple variation was

unusually good, owing to the facilities afforded for the

examination of fruit from various localities and different

environments, and it has been possible to work out the

fundamental principle upon which variation resulting

from external factors depends and to apply it in the study

of environmental adaptations. This principle, the Law of

the Optimum, states that, for any given variety there is

for each character a certain intensity of each essential

factor of the environment at which, other conditions re-

maining the same, that character reaches its highest de-

velopment.

In the application of this law to varietal adaptations,

the essential point is the proper balance between char-

acters and environmental factors, that is, all factors

should be of such an intensity as to permit a good all-

round development of the fruit. In the absence of such

No. 574] VARIATION IN THE APPLE 633

a balance certain characters may fail to reach a proper

degree of development while others develop to excess.

The failure in quality and other respects of many of

the best dessert varieties of apples when grown in Wash-

ington is due to such a lack of balance. Practically all of

them originated under a much different environment and

were selected and came into prominence owing to their

perfect balance of adaptation in localities having a set of

external conditions similar to those under which they

originated. The hope of northwestern apple culture in

the future lies in the careful selection of varieties and the

origination locally of varieties of high quality showing

adaptation to the conditions of growth in the various sec-

tions. In the meantime plantings should be made from

those varieties of high quality which show the best

adaptation. These are Esopus, Yellow Newtown, White

Pearmain and Delicious for the irrigated valleys, and

Wagener, Delicious and McIntosh for the higher valleys

of northern and eastern Washington. Jonathan, Stay-

man and Winesap show a poorer balance and should not

be planted too recklessly. The climate of the Pacific

coast resembles that of western Europe more than that

of the eastern states, and further importations of Euro-

pean varieties is desirable especially for testing west of

the Cascades.

The moisture relation is probably the most important

factor in inducing variations, and is doubtless responsible

for certain variations which have been ascribed to other

causes which act indirectly by modifying the moisture

supply. The elongation of the fruit following a cool

period after blooming may result from a diminished cir-

culation of the sap, giving rise to an insufficient supply to

provide for the simultaneous development of the fleshy

portion and elongation of the axis. Variation in the

depth of the cavity and basin in certain varieties is prob-

ably to be explained in a similar way.

Color modifications depend to a great extent upon the

light relation and somewhat upon development as influ-

enced by temperature. The optimum intensity for the

634 THE AMERICAN NATURALIST (VoL. XLVIII

production of red pigment is quite narrow in most varie-

ties and poor color may result from either deficiency or

excess. Latitude and altitude affect the color only as

they modify the factors upon which color depends, caus-

ing them to approach or recede from the optimum. The

influence of elements in the soil is not well understood.

It is probable that soils containing sufficient iron. for the

proper development of chlorophyll contain an abundance

for the production of red pigment in apples.

Aside from such differences as depend upon the hand-

ling of the fruit, variations in keeping quality appear to

follow the law of the optimum in the same manner as the

other characters of the fruit. Conditions which favor the

best all-round development result, as a rule, in good keep-

ing quality. Apples grown under irrigation are said to

keep poorly probably because of their unbalanced adapta-

tion to the environment. Certain factors which favor de-

velopment and maturity are present in excess, resulting

in overgrown or overripe fruit.

Varieties differ in specific gravity according to the

extent of intercellular spaces in the flesh and the open-

ness of the core. Overgrown specimens are low in speci-

fic gravity. As a rule, those lots which kept best in any

variety had the highest specifie gravity.

Chemical composition is associated somewhat with

quality. High suerose content results in richness of

flavor. Fruit of high quality has the sugars and acids

well balanced and the flavoring constituents well devel-

oped. A heavy juice is usually associated with a high

content of soluble solids. Fruit grown under irrigation

is ordinarily rather low in soluble solids. There seems

to be no constant relation between the amount of sunlight

and the production of sugars, and flavors appear to de-

velop best in a relatively cool climate.

SHORTER ARTICLES AND DISCUSSION

VARIATION AND CORRELATION IN THE MEAN AGE

AT MARRIAGE OF MEN AND WOMEN

SOMEWHERE in sociological literature we have met with the

statement that whereas the mean age at marriage of men differs

from district to district because of social and economic conditions,

the mean age at marriage of women varies but little because of

these factors. In view of the high ‘‘assortative mating”

coefficient! for age of bride and groom, this statement seemed so

remarkable as to be open to question,

Its validity can be very easily tested provided the mean age at

marriage of men and women from a series of districts differing

in economic and social conditions are available. If the mean age

of women is independent of these conditions, or far less depend-

ent upon them than that of men, one should find (i) that the

variation of mean age of brides is lower than that of mean age

of grooms, and (ii) that for a series of districts the coeffiicent of

correlation between the mean age of brides and grooms is very

low.

The only suitable series of data that we have been able to find

is that given by A. Dumont? for the average age in years and

months at first marriage of the males and females of the 87

departments of France. Grouping his data in classes of five

months’ range, we find, in terms of months :*

1 See Lutz, Science, N. S., Vol. 22, pp. 249-250, 1905. For a general re-

view of the literature of assortative mating see Harris, Pop. Sci. Mo., Vol.

80, pp. 476-492, 1912.

2 Dumont, A., Rev. Ecole Anthrop. see Vol. dey 163, 1904.

3 The aese given by the ungrouped d

For Men For Women

ENE SS E A S $37.87 + 83 284.45 + 1.01

Standard deviation rssi yerr 11.49 + .59 14.00 + .72

Seo iy 492+ 25

oe the shortness of the series, the results are in as good agreement

could be expected.

For Men For Women

PEM TT PET EAE E SSE E O La ewe 337.76 + .80 284.43 + 1.03

Standard oe ey Uy bats ule Ke 11.03 + .56 14.25 + .73

3.26 & 17 6.01 + 26

4 wena Sheppard’s correction for the second moment.

635

Mean Age of Women.

636

THE AMERICAN NATURALIST [Vou. XLVIII

-260—

325 335 34

355 365

Mean Age of Men.

We note that the women marry on an average about four years

and five months younger than the men.

have been told, their mean age at marriage both absolutely, as

measured by the standard deviation, and relatively, as measured

by the coefficient of variation is more variable than that of men.

The difference in standard deviations for the ungrouped material

is 2.51 + .93 and for the grouped records 3.22 + .92. These are

2.71 and 3.49 times their probable error, and hence perhaps

significant. For the coefficient of variation, the differences by

the two methods are 1.52 + .31 and 1.74 + .31. These are 5.69

and 4.96 times their probable errors and their significance is

even more probable than those for the standard deviations.

The correlation coefficient from the grouped data by the prod-

uct moment method, using the means and standard deviations

given above, is

Contrary to what we

No.574] SHORTER ARTICLES AND DISCUSSION 637

Tmt ==.781 + .028.5

Thus on a scale of —1 to +1 the interdependence of mean

ages of men and women is very close indeed.* Expressing the

same relationship in terms of regression by the well known

formula

~ (F7—5% wm) 47H

fa (7 rm) +r m,

where the bars indicate population means and the sigmas popula-

tion standard deviations of m = males and f =— females,

f=—56.474 + 1.009 m.

Thus we see that each month’s increase in average male age is

followed by .a month’s increase in mean female age. The fit of

the straight line to the empirical means as shown in the diagram

is excellent—considering the small number of the district means

from which the equation is deduced.

Thus the available data show that the mean age at marriage of

women instead of being less variable from district to district than

that of men is actually more variable—both absolutely and rela-

tively,

In short, there is, as far as our data go, no evidence for the

assertion that while the time of marriage of men is closely de-

pendent upon the complex of social and economic conditions that

of women is. practically independent of them.

We have published this note in the hope that it may suggest

to some one with the opportunities of obtaining really adequate

data an investigation of the problem which has several rather

important points of interest.

J. ARTHUR HARRIS,

Roxana H. VIVIAN

COLD SPRING HARBOR

5 The difference method applied to the ungrouped material gives

r= .763 + .030.

The difference is of no significance.

6 Possibly, however, the relationship is in part spurious. The mean of

males and females were taken on the basis of the same N, or approximately

the same N, for the various districts. Data for investigating this question

are not available. The point should be borne in mind by a subsequent

worker.

638 - THE AMERICAN NATURALIST (Vou. XLVIII

DUPLICATE GENES

Some interesting questions are raised by a recent article by

‘Gregory: ‘‘On the Genetics of Tetraploid Plants in Primula

sinensis.’’ Reciprocal crosses of two races of P. sinensis were

made. One cross gave entirely normal results in F, as regards

chromosome number and hereditary characters. The reciprocal

cross gave an F, generation which was sterile with the parents

and produced only a giant variety in F,. This proved to have

the tetraploid chromosome number. Experiments indicated that

the genetic factors had also all been doubled, a very significant

parallelism.

Gregory uses the nomenclature AAAA, ast AAaa, Aaaa,

and aaaa to represent all the possible conditions as regards a

pair of Mendelian factors. He states that heterozygotes of the

form AAAa should give gametes AA and Aa, and should pro-

duce, on selfing, the zygotes AAAA, 2A A Aa and AAaa, and that

the last class selfed should produce recessives. On the chromo-

some theory of heredity, this assumes that the four chromosomes

concerned are equally likely to pair in synapsis in any of the

possible ways, a very interesting phenomenon if the assumption

proves correct. But it is conceivable that two independent

synaptie pairs may be formed. It may be that only chromo-

somes from the same original race pair in synapsis. It is true

that ‘the first of the original crosses shows that the chromosomes

of the two races can enter into normal mitosis and presumably

into‘synapsis with each other. But the reciprocal cross indicates,

perhaps, that in the environment of the cytoplasm of ‘this cross,

they can not enter into synapsis. If this condition continues in

later generations, we should represent the zygotes as AAA’A’,

AAA’‘a’, AaA’a’, ete. This is the way in which duplicate genes

have been represented previously as by Nilsson-Ehle, East and

Shull. With this representation, heterozygotes of the form

AAdA’a’ could never give rise to recessives after selfing for any

number of generations.

Which hypothesis is true in this case could easily be deter-

mined by experiment. The published results are not sufficiently

explicit on this point. If the original cross were of the type

AA X a'a’, producing in F, Aa’, the F,, AAa’a’, would be a

homozygote on the second hypothesis, and recessives should never

1 Proc. Roy. Soc., B 87, 1914.

No. 574] NOTES AND LITERATURE 639

appear. On Gregory’s hypothesis recessives should appear in

later generations. On the second hypothesis, homozygous races

of the types AAa’a’ and aaA’A’ would be obtainable, in appear-

ance like heterozygotes. These would breed true indefinitely

when selfed, but should give recessives in F, after crossing, as in

a case proved by Nilsson-Ehle.

SEWALL WRIGHT

BUSSEY INSTITUTION, 4

Forest HILLS, MASS., ¢

June 19, 1914 \*

NOTES AND LITERATURE

A STUDY OF DESERT VEGETATION!

Between three and four years ago Dr. W. A. Cannon, of the

Desert Botanical Laboratory at Tueson, Arizona, visited southern

Algeria in order to become acquainted with the more obvious

features of the plant physiological conditions of the desert, and

to make detailed studies of the root habits of certain desert

plants. From Algiers the journey proceeded nearly due south

about three hundred miles to Ghardaia, thence east about» one

hundred miles to Ouargla, and another hundred miles to Ee

gourt, returning through Biskra, and Batna to the northerneoast

Throughout this long and wearisome journey the vegetation was

studied in connection with the geographical and climatice environ-

ment and the results are brought together in a volume of some-

what more than eighty pages of text and thirty-seven plates, one

of which is an outline map of the region visited.

Dr. Cannon speaks of the similarity of the flora of Algeria to

that of southern Spain, France and Italy, where one is reminded

of the vegetation of portions of California. Once in the desert on

the way south low-growing shrubs on the plain become char-

acteristic, including species of Tamarix, Zizyphus and Artemisi.

Where water is available for irrigation, oases oceur with the

luxuriant vegetation of date palms, apricots, figs, mulberries!

peaches, pears, oranges, as well as artichokes, beans, carrots,

melons, peas, potatoes, squashes, ete. Further south the plain

1 Botanical Features of the Algerian Sahara. By William Austin Cannon,

640 THE AMERICAN NATURALIST [Vou. XLVIII

is covered with small stones and pebbles and ‘‘not a tree, shrub,

or herb appears to hide the bare ground. The mountains are

naked rock, while the harsh outline of desert ranges and the

distant low sand ridges give no evidence of plant life. But a

closer examination of plain, dune and mountains reveals the

presence either of living forms or of the dried remains of plants

of a preceding moist season, in numbers and in kinds not at first

suspected.’’ All of which might well describe the desert condi-

tions in our own southwest. This similarity is emphasized by the

resemblance of many of the plants to those found in our Arizona

deserts. Thus the ‘‘quidad’’ (Acanthyllis tragacanthoides) ‘‘has

a very close resemblance to small specimens of ‘ocotillo’ (Fou-

quieria splendens) of the southwestern United States.’? And

this resemblance extends to the structure of the spines and the

return of the foliage after rains. It is interesting to note that

the natives burn off its numerous spines, after which the stems

“‘are eaten with avidity by camels,’’ reminding us of the similar

treatment and use of some cactuses in Arizoria. Further to the

south the vegetation is still more sparse and xerophytic, includ-

ing Ephedra, Retama, Haloxrylon, and among grasses, Aristi

pungens. Near Ouargla, the southern point reached, there are

places where no vegetation is present, as on the dunes, and yet

on the fixed sand nearby were found Euphorbia guyoniana,

Retama retam and Genista sahare.

Much attention was given to the root habits of the plants

encountered, and in the general summary which follows the

account of the journey comparisons are made with the root habits

of Arizona plants.

With this meager introduction we must refer the reader to the

volume itself, which it is quite impossible to summarize in these

pages. One thing impresses itself forcibly upon the reader, and

that is that a desert is a hungry place in which the permanent

vegetation maintains itself against plant-eating animals by a

thorny or spiny protection. Yet Dr. Cannon points out that in

this character of spininess the American desert plants excel those

of the plants of the Sahara region.

CHARLES E. BESSEY

THE UNIVERSITY OF NEBRASKA

VOL. XLIII, NO. 575 NOVEMBER, 1914

THE

AMERICAN

NATURALIST

A MONTHLY JOURNAL

Devoted to the Advancement of the Biological Sciences with

Special Reference to the Factors of Evolution

CONTENTS

Page

I. A Comparison of the Responses of Sessile and ob Plants and Animals.

fessor VICTOR E. SHELFORD 641

II. An Apterous Drosophila and its Genetic Behavior. CHARLES W. METZ - 675

III. Shorter Articles oy amea Formulæ for the Results of Inbreeding :

Professor H. S. NINGS. ort-cut in the Computation of Certain

Probable Errors : PEER B. Frost. Galton and Discontinuity in Vari-

ation : Dr. R. giona GATES. The esua of a ae — Calf :

- 693

Professor A. M. REES

THE SCIENCE PRESS

LANCASTER, PA. GARRISON, N. Y.

NEW YORK: SUB-STATION %4

The American Naturalist

SS intended for publication and books, etc., intended for review should be

sent perei Editor of THE DOE shale NATURALIST, Garrison- -on-Hucson ee

articles contai summa

w York.

ies of research work bear on the

problems of organic alaiak are sonectaity welcome, and wiii be tues go Lo hae

in or on.

rea reprints of ear are supplied to authors free of charge.

e hu

Further 1 reprints will be supplie

ubs

s and advertisements should be sent to the Tigercat

ear.

twenty-five pal additional.

subscription price is four dollars

Canadian postage

The

Foreign postage is fifty cents and

The charge for single aer En is

page.

forty cents. The advertising rates are Four Dollars for a p

THE SCIENCE PRESS

Lancaster, Pa.

Garrison, N. Y.

NEW YORK: Sub-Station 84

Entered as second-class matter, April 2, 1908, at the Post Office at Lancaster, Pa., under the Act ot

Congress of March 3, 1879.

ARCTIC, pe deity and GREENLAND

DS’ SKINS,

Well oie Low Prices

Particulars of

G. DINESEN, Bird Collector

Husavik, North Iceland, Via Leidle, England

WANTED TO PURCHASE

aset of BIRDS OF AMERICA by J. J. Audubon,

7 or 8 volumes, please report, stating cash price, stat-

ing condition, binding and dates of volumes.

F. C. HARRIS,

Box 2244 Boston, Massachusetts

For Sale Entire

An important collection of Indian Birds’

Eggs, containing approximately 3,300.

specimens of about 620 species, and in- |

cluding among other varieties the Green-

ish Willow Warbler (Acanthopneuste

viridanus), Red-browed Finch (Calla-

canthis burtoni), Tibetan Sand-Grouse

(Syrrhaptes tibetanus), Spotted Sand-

Grouse (Pteroclurus senegallus), Ibis-

bill (Ibidorhynchus struthersi), White-

bellied Minivet (Pericrocotus erythro-

pygius), Bearded Vulture (Gypaétus

barbatus), numerous Cuckoos with their

hosts, etc., ete.

Particulars and full list may be had

from

W. F. H. ROSENBERG

57 Haverstock Hill London, N. W.

Marine Biological Laboratory

Woods Hole, Mass.

INVESTIGATION Facilities for research in Zoology,

Entire Year arg Stade Physiology and ae

of members of the staff.

for such a table is $50.00.

Courses of laboratory instructi

ures are offered in inarte-

vigie

to August 11,

$50.00.

Phil hical Aspects of Bi

and Alli camara td

SUPPLY . =. gens

DEPARTMENT

Open the Entire Year

State which

n and sli

cn tS jal

GEO. M. GRAY, Curator, Woods Hole, Mass

announcement will be sent on application ox

The

Laboratory, Woods Hole

annual

The Director, Mario

Mass-

THE

AMERICAN NATURALIST

Vout. XLVIII November, 1914 No. 575

A COMPARISON OF THE RESPONSES OF

SESSILE AND MOTILE PLANTS

AND ANIMALS

PROFESSOR VICTOR E. SHELFORD,

UNIVERSITY OF ILLINOIS

E AAA ai'e's 5 ok Soda Chg 4 Hs CRE ee Ca oes Cia te eas 642

IL. Basis of Discussion R E E hehe T A E E N A 642

aE AE PR ually ENE EN P E EE TO CV a E R E RS 643

. ` Sessile yer Motile si rie ERC EWN EA ROMP PE MICN Se wes

The er pee and it + Relation i in fe and Groups ......

(a) A and Plants made up of Single Individuals .... 644

(b) Colonial ~ Multiple Individualed Plants and Animals. 645

i; apas — éro df Individus 8 kins ei cs Ca eri 645

ii SE is ne Kgs ck be eee a cia eee e ew sa wes ee

iii. Metabeli and Reproductive Relations of Individ-

Ses Wak 6 Wine SR wt eo eee air pe bee Chae 6 647

(e) Response ‘ot atin E TAS a, PE e EE AR oa

STORE KEME SC A E E 6

Str a a aao 5004 oS £5 050 01S oa Fa cece a

(d) Response of a ile j nt A EE EN E A 651

COUR) DOMINO iiss cae oe i dev enrarir enti 651

ii, orenak _ ROR E PSUS ere non T S ete 653

(e) Behavior of Sessile-motile Organisms ..............

Response and Taxonomy of Sessile Danaa E E 653

E TS Saeg Sessile aad Motile Organisms with reference ad

Aii Cg Seren E S E E Lae EE oar TA ear ee

1. ews eee pi parmtla go crea, pear pA Ue Sine he erga

2. rakte AA of Sessile and g ws a x Asst Biota .... 655

3. Sessile Motile Organisms in Ecological Succession ..........

IV. Influence of Res nse Phenomena poe Biological Theory and Con- 3

rove

i. Teleclogical voce Sempre, eee ire oe ae eee cr

Bs NEDA othon VOW o.oo cli ees hs ce eee seiesvseeses ees

3 ee ae ng eta of Response and the Germ Plasm

4. Pasas of the ‘Study ‘of Response on Present- -day Biological

ASO a PE T A E E re O E eT ee

5. Aspects of the Untenability of the Germ Plasm Doctrine.... 662

“3 ees of Values in —— RAPD es ne ones aes cess

n Ra a EEA TA N T ook S a a

V. re bea RN ins eke S cake eeee vee ashes ehesees 672

642 THE AMERICAN NATURALIST [Vou. XLVII

I. INTRODUCTION

Durine the past few years the attention of biologists

has turned more and more from those phenomena which

were supposed to be comparatively fixed, to responses to

stimuli. Physiologists have long been concerned with

the mechanism of response; psychologists are interested

in its modification. Geographers, climatologists and

ecologists have recently turned their attention to re-

sponses in natural environments and zoologists have

become interested in response, particularly from the point

of view of its specificity. In these quite independent

investigations and compilations there has been little

attempt at analysis with a view to determine legitimate

‘lines of comparison among the exceedingly diversified

types of organisms which have been investigated, and

some confusion has resulted. For example, since the

more obvious responses of plants are structural, persons

not familiar with comparable phenomena among animals

have made erroneous comparisons of sessile plants and

motile animals. This paper is written to present in as

nearly uniform terms as practicable (a) analysis of kinds

or aspects of response, (b) justifiable kinds of compari-

son, and (c) the bearing of response phenomena on

biological theory and controversy. It aims to show that

the numerous kinds of response are reducible to a few

simple types common to both plants and animals, and that

the failure to consider all types has been responsible for

confusion and various one sided theories. It further aims

to show that study of response during the past few years

has led to an unusual broadening of our conceptions.

II. BASIS OF DISCUSSION

As a basis for discussion we must first have a clear

understanding of the character and definition of response.

Secondly, we must determine what constitutes an indi-

vidual in those plants and animals that are made up of

repetitions of parts. Thirdly, we must note whether or

not the organism is sessile or motile, capable of playing

the part of either, or colonial pelagic.

No.575] RESPONSES OF PLANTS AND ANIMALS

1. RESPONSES

The word response is used in various slightly different

senses. In general it refers to more complex and time-

requiring phenomena than ‘‘reaction.’’ In geography the

term has been used (Goode, ’04) to cover all changes in

culture supposed to be produced by climate or other

geographic conditions. It is also applied by geographers

and geologists to changes in the physical characteristics

of man (evolution) which Goode (’04) has stated are

slower than the cultural responses. In general botanists

have used the term to cover changes of plant structure

and function induced by external conditions.. Cowles

(711), however, uses the word ‘‘reaction’’ to cover these

phenomena. Coulter (’09) used the term response as

synonymous with adaptation in plants. Zoologists have

used the term to apply to changes in animals due to exter-

nal conditions, but with little agreement as to what is to

be included. We will use it here to include reactions,

changes in functions, structure, color, induced by external

conditions either directly or indirectly, without regard

to how simple or how complex the processes involved

may be.! The length of time required to bring the

changes about may arbitrarily be taken as not exceeding

the time required to breed five to ten generations of the

species concerned. All organisms respond to stimuli

because each stimulus acts upon some internal process.

Strictly speaking, the response is the change or changes

in the physical or chemical processes of the organism (or

the part or parts concerned) which results from the

disturbance.

Those things which we commonly see and term response

are often the later and less important phases of the dis-

turbance. The striking phases of responses of motile

organisms are usually movements which follow closely

upon stimulation. In sessile organisms the noticeable

responses often appear only after a considerable period.

In both sessile and motile organisms some responses are

i 1 For good representative bibliography see Adams, °13, Ch. VIII and

644 THE AMERICAN NATURALIST [Vou. XLVIII

not evident because they concern internal, chemical and

physical processes which affect neither form nor move-

ment. Changes in the enzymes secreted by digestive

glands, which accompany changes in food (Jennings, 06,

p. 347), are examples. While thus recognizing that re-

sponses are concerned primarily with internal processes,

we must of necessity refer chiefly to the external phases.

2. SESSILE AND MOTILE ORGANISMS

Sessile organisms are those which are sedentary in

habit, whether attached or possessing slight powers of

locomotion. Motile organisms are those that habitually

move about. Vagile or creeping forms as well as swim-

ming, walking, flying, burrowing types are included.

Most sessile animals are capable of moving their parts,

while only a few sessile plants possess this capacity, and

these only to a slight degree.

There is no sharp distinction between sessile (seden-

tary) and motile organisms. Every possible gradation

exists between fixed non-motile types as trees on the one

hand and the pelagic fishes on the other. It is the

extremes which we will compare.

3. Tue [INDIVIDUAL AND Irs RELATIONS IN COLONIES

AND GROUPS

The following comparison of animals and plants is an

attempt to distinguish potential or incomplete individuals

in colonial organisms and compound organisms which,

while not commonly recognized as colonial, are made up of

incomplete individuals.

(a) Animals and Plants made up of Single Individuals

The vast majority of animals belong here. Most pro-

tozoa, solitary sponges, solitary hydroids, sea anemones,

worms not preparing for asexual division, echinoderms,

mollusks, arthropods and vertebrates. Only single-celled

plants, young seedlings and possibly a few adults of multi-

cellular plants which possess but one growing point

No.575] RESPONSES OF PLANTS AND ANIMALS 645

(exclusive of roots) belong in this group. Single indi-

viduals as described here are the basis for determining

what shall be called individuals in colonial and compound

types.

(b) Colonial or Multiple Individualed Plants and Animals

A number of animals and the vast majority of the

plants belong here. The group can be roughly divided

into two types, (a) those having a chain or plate arrange-

ment of incomplete individuals and (b) those having a

branching or tree-like arrangement. The groups of in-

complete individuals of type a occur among the Protozoa,

worms undergoing asexual reproduction, many of the

Bryozoa and some of the Tunicates; both sessile and

pelagic (plankton) forms occur. On the plant side type

a includes plate-like colonies of alge, filamentous alge,

some thallose plants and probably some of the fungi,

though the great multiplicity of forms makes the separa-

tion of this group from the branching tree-like types,

difficult.

Type b includes some of the colonial Protozoa, the

majority of the sponges, hydroids, corals and the branch-

ing Bryozoa. The alge, fungi, mosses, ferns and flower-

ing plants are all represented. The colonies are usually

attached to the substratum (sessile).

i. Numbers of Individuals——Among the animals the

number of so-called zooids is the number of incomplete

individuals. In the sponges there are as many zooids as

there are excurrent openings (oscule) (Minchins, ’00,

p. 91). Zooids usually possess a mouth opening and

organs for securing food, though in some cases they may

be specialized for reproduction, defence or locomotion as

in some of the Celenterates. Among the colonial plants

there are as many incomplete individuals as there are

buds or growing points (vegetative regions). There are

no regularly occurring organs in animals, strictly com-

parable to leaves. However, any organs such as tentacles,

gills, ete., which secure or absorb nutriment may be re-

646 THE AMERICAN NATURALIST (VoL. XLVIII

garded as analogous to leaves. Each potential bud with

its leaf may be compared to a zooid. In comparing plants

and animals, roots can perhaps be compared with the

holdfast organs of hydroids. In both groups, roots and

root-like organs are individuals of a very low order of

individuahization and of a type not well represented

among animals. The holdfast organs of animals are not

important absorbers of food and water.

ii. Stems and Other Connecting Organs (Conducting

Tissues).—The most striking difference between the in-

complete individualed or colonial plants and colonial ani-

mals is the presence in the former of specialized stems and

highly complex conducting tissues (Cowles, 711; Piitter,

"11, pp. 861-66). The conduction of food materials from _

the root to other parts of the plant and from the leaves to

the root is a functional necessity not paralleled even in

those colonial animals showing the greatest division of

labor. In animals stems are relatively undifferentiated

and are often made up of living, relatively unspecialized

zooids, as, for example, in many Bryozoa such as Crisis.

The tendency to cauliflory in some plants and the ability

of cambium to produce shoots and of the stems of most

hydroids to produce individuals indicates that such a con-

dition may be potentially present in all. In stalked

Protozoa the stems are solid, while in most Coelenterates

they are tubes, usually simple though sometimes complex,

made up by mere elongation and branching of the stock.

of the simple single forms such as the Hydra. The lumen

is usually ciliated and makes possible a transfer of mate-

rial which renders practicable such division of labor as

occurs in this group (Piitter, 711). In the Bryozoa the

different zooids have their body cavities joined in the

simpler forms merely as a branching lumen of the main

wall of the colony; in others by small openings the more

specialized of which are sieve-like plates (Harmer, ’01,

pp. 471 and 496; Delage and Herouard, ’97, Vol. 5, p. 62).

The connection between the individuals of the tunicate

colonies is often very complex, due to the fact that in the

No.575] RESPONSES OF PLANTS AND ANIMALS 647

most complex types the stolon (stem) gives rise to new

individuals and possesses all the layers of cells which

take part in forming them. The connection between

different individuals differs in different groups and is

determined by the particular mode of asexual reproduc-

tion. As the individuals are quite independent of one

another in function, these connections do not have the

Same significance as in plants. Even where there is a

common blood cireulation, as for example in the Clavel-

linide (Harmer, ’04, p. 71), there is no noteworthy divi-

sion of labor. —

iii. Metabolic and Reproductive Relations of Individ-

uals.—The flat worms at certain times consist of chains

of zooids at various stages of development and with

various degrees of independence. Child (713) has found

that these chains of zooids present a series of gradients

in rate of metabolic reaction. The rate is highest at the

anterior end of the whole chain and decreases toward the

posterior end, not uniformly, however, for the rate is

lower immediately in front of each head region than it

is in the head region itself. A gradient is present in the

axis of each zooid. The most anterior head dominates so

long as the chain remains intact. In the corals certain

zooids dominate (Wood-Jones, ’11) over the others.

Some types have a single dominant zooid and some more, `

while in other cases all are equal.

Among plants whose form is that of a chain or a plate

the individuals are less closely bound together and domi-

nant vegetative regions are probably less well developed.

In the branching types, dominant vegetative regions occur

(Cowles, ’’11, p. 747; Goebel, ’00, Vol. I, p. 206). In the

conifers, for example, there is a leader, a dominant grow-

ing region at the tip of the main stem just as in certain

madrepore corals (Wood-Jones, p. 83). Other plants like

the elm have several vegetative regions which dominate

over others, as they do in the branching madrepores.

Growth form or colony form varies according to cer-

tain laws dependent, in part at least, upon the metabolic

648 THE AMERICAN NATURALIST [Vou. XLVIII

relations of individuals. Thus Wood-Jones says of the

corals—

a colony may grow according to five different types of vegetative growth

. it may grow as (1) a spherical mass, (2) an encrusting layer, (3) a

free plate, (4) a branching tree-like growth, or (5) a mere amorphous

lump.

He further notes the division of all the corals into two

groups of normal growth-forms; for all the zooids may

take an equal share in the asexual reproduction or, again,

some may be of greater importance than others, and the

asexual reproductive functions may be lodged in a very

few individuals only. Considering the first division

(all zooids taking equal share, the principal types of bud-

ding vary from each other in the actual site of origin of

the daughter zooid from the parent, in the degree of final

separation of the two zooids, and in the thickess of the

intervening partition between the two zooids. The

amount of rising above the general surface by each indi-

vidual zooid is likewise subject to variation.

Turning now to the corals that constitute the second

class (some zooids of greater importance than others)

which in the words of Wood-Jones have some of their

units specialized as active agents of growth,

it is at once seen that the possibilities of variation of normal vegetative

habit are greatly increased. 1 the elaborate branching forms, plates

and leaf-like growths belong to this class; and all are evolved by special

peculiarities of the growing point. The zooids that constitute the grow-

ing point may take various forms; they may be arranged as a cluster, as

a creeping edge, or as many varieties of terminal shoots of branches.

In the first instance, it is necessary to draw very sharp distinctions

between two subdivisions of this group. In Group 1 come all those

forms like Montipora, whose distal zooids are the newest formed mem-

bers of the colony; and in Group 2 are included the PEAT whose

distal zooid is the most ancient individual in the whole p.

In dealing with Group 1 many forms have to be pate Hy for when

the youngest are the active zooids their growth cluster may be variously

disposed, and on its disposition the resulting vegetative form entirely

depends.

In Group 2, however, this state of things is entirely altered, for there

one zooid, which is situated at the extremity of the stem, and which I

No.575] RESPONSES OF PLANTS AND ANIMALS 649

shall call throughout the “ dominant apical zooid,” constitutes the grow-

ing point; and this zooid is the parent of the entire colony.

Various writers make comparable statements or show

comparable principles among hydroids (Motz-Kossowska,

08) and Bryozoa (Davenport, ’91, et al.) and among

plants (Goebel, 700). Of the colony form of the tunicates

Herdman (’04, p. 82) says:

The marked differences in the appearance of the colonies of compound

Ascidians is largely due to the methods of budding; even in those of

stolon type where the budding is practically the same in essential nature,

the results may be different in superficial appearance, according as the

buds are formed on a short stolon close to the parent body, or from the

extremity of the post abdomen or from the long epicardiae tube which

may extend for some inches from the ascidiozooid.

Thus we conclude that the innate causes of different

growth-forms (colony forms) of colonial organisms are

(a) the mode of division of the zooids or vegetative

regions, (b) the ratio of stem elongation to number of

zooids or buds produced or uniformity or lack of uni-

formity of stem elongation (Wood-Jones, p. 76) closely

related to (c) the presence or absence, number, position

and region of influence of the dominant growing regions

or dominant zooids, and (d), in some cases, the grand

period of growth and the length period of the internodes

(Johnson, 711). The innate tendencies are thus reducible

to a few principles applicable to both plants and animals.

i. Movements.—In motile organisms the most striking

responses are changes in position brought about by

movements usually more or less-random, and which bring

the organism into various conditions one of which usually

relieves the disturbance. The organism resumes normal

activity in conditions which brought the relief (Jennings,

06). These conditions are not necessarily advantageous,

but are usually so when the stimuli are those encountered

in nature (Mast, 711). Behavior of motile organisms is

also modified by repetition of action even in animals as

low in the animal series as the Protozoa (Holmes, 711).

650 THE AMERICAN NATURALIST [Vou. XLVIII

Jennings (‘06) has quoted various botanical workers’

observations on motile plants the behavior of which prob-

ably follows the general laws governing the behavior of

motile animals. As a result of the quick behavior re-

sponses of motile organisms, their distribution at any

given time is a better index of the conditions at that time

than the distribution of sessile organisms, because when

the conditions at a given point become unfavorable the

motile organisms usually move to another situation,

while the sessile forms remain and perhaps die.

ii. Structural Responses—Among motile animals,

structural and color changes occurring as a response to

environmental conditions (stimuli) are usually not of

importance to the organism concerned. The color differ-

ences induced in Lepidoptera by heat and cold (Stanfuss ;

Fischer) and the structural differences in Crustacea such

as were brought about in Cladocera by Woltereck, and

other modifications brought forward recently, are usually

of no known advantage or disadvantage to the animals

concerned (Bateson, 713, Ch. IX and X). Such re-

sponses in color and general form do not ordinarily take

place in adults subjected to such conditions. The strik-

ing structural responses of motile animals are often

responses to the organism’s activity. The use and disuse

phenomena of the Lamarckians, the increase in size and

form of muscles, thickening of skin in man and mammals,

are well-known examples of a type of responses which

have influenced zoological speculation. Child (’04) con-

trolled the form of Leptoplana by controlling activity.

Holmes (’07) found that the movements of pieces of

Loxophyllum have an important part in shaping the

general outline of the bodies of the resulting forms. The

general forms of motile animals are correlated with their

activities but whether form or structure correlated with

it appeared first in the course of evolution has been the

subject of considerable fruitless speculation.

No. 575] RESPONSES OF PLANTS AND ANIMALS 651

(d) Responses of Sessile Organisms

i. Structural Responses——The striking phases of re-

sponses among colonial sessile organisms are often

changes in form and structure, or the relative position

of the parts. The changes in structure or position of

parts are not necessarily advantageous or useful, but are

usually so when the stimuli are those commonly encoun-

tered in nature (Cowles, ’11; Loeb, ’06, p. 124; Wood-

Jones, 711; Ch. VIII). Indifferent and detrimental re-

sponses are often given under experimental conditions and

no doubt the absence of such variants among sessile ani-

mals collected in a wild state is due in part to the failure

of such organisms to survive. A few sessile colonial

organisms such as cacti (Cowles, 711) show little or no

plasticity.

Among sessile animals, the observations of Wood-

Jones form the belt examples of response. He found

that the branching type of corals dominated in barrier

pools, tall slender non-branching types in deep water,

and massive boulder types on surf beaten shores. Thus

he figures similar colonies of each of three genera which,

while possessing certain peculiarities of their own, are in

general agreement as to growth form just as sessile

plants usually are; and this in part for comparable rea-

sons. Thus various conifers occur as Krummholz in the

high mountains, due to severe conditions (Cowles, 711,

p. 732), wind, snow, and in part to the injury of terminal

growth regions of the main stem which gives rise to

lateral branches. The boulder-like corals with the zooid

at the same level occurring on the surf-beaten shores of

coral islands are due, in the case of Madrepora, for ex-

ample, to repeated injury of the terminal dominant zooids.

Conifers in protected situations often grow into tall

slender trees comparable with the (deep) stil-water

corals. The barrier pools afford conditions where the |

terminal buds are less often injured than in the surf and

the tree-like branching corals result from minor injuries

to dominant zooids.

652 THE AMERICAN NATURALIST [Vou. XLVIII

Wood-Jones finds further that still-water corals are

less strongly calcified than those in rough water, the

strains producing increased secretion analogous to in-

creased tissue production as a result of mechanical

strains in plants (Cowles, p. 669). Corals show different

kinds of growth under different environments partic-

ularly when injured. The new part may be different

from the rest and adjusted to the environment thus

making it appear as though two ‘‘species’’ occurred in

the same colony. The mode of division of the zooid is

also different under different conditions. Plants show

similar variation with changes of conditions, particularly

in the leaves which are divided in submerged portions of

amphibious plants and entire in the emerging portions

(Cowles, 711, p. 595).

As has been noted, there is nothing in sessile animals

that is more than roughly analogous to leaves. Leaves

show marked structural differences on different parts of

the same tree where the environmental conditions are

different, as, for example, in the differences which occur

between the upper and lower portions of a forest tree.

While there are, no doubt, differences in similar details

(histology) in the organs of display in different parts

of the same colony of sessile animals, little or nothing

has been done upon them. As a further indication of the

prevalence of structural response in sessile organisms

of the hydroids Hickson states that there is probably but

one species of Millepora which occurs in a large number

of growth forms. The commercial sponges (Moore, ’08)

and common freshwater sponges and polyzoa show many

different forms under different environmental conditions.

The major differences in growth form induced by ex-

ternal stimuli in colonial organisms result from modifica-

tions of the rate and character of growth with respect to

= the four innate tendencies toward various growth or

colony forms discussed above, and which may be briefly

enumerated as follows: (a) mode of division, (b) amount

_ of stem elongation, (c) influence of dominant regions and

No.575] RESPONSES OF PLANTS AND ANIMALS 653

(d) grand period of growth and the length of period of

internodes.

The principles are concerned with asexual reproduction

and apply to motile organisms only exceptionally as for

example in the case of colonial pelagic forms. The laws

are applicable to both plants and animals.

ii. Movements.—Movements of sessile animals are

usually contractions or extensions of parts or of the

entire body. Tentacles and comparable organs are capa-

ble of movements for securing prey. Such organs often

tend to wrap about objects which are in motion. Many

Sessile animals are capable of opening and closing a

mouth opening and of bending or twisting the entire body.

Plants possess a comparable capacity only occasionally.

(e) Behavior of Sessile Motile Organisms

Most sessile animals are capable of some movement

and react by contraction of parts. The reactions may be

modified by repeated stimulation (Jennings, ’06) and

usually by physical factors. Some animals, as Hydra,

Stentor and many others are both sessile and vagile or

free-swimming, and show different types of behavior

when attached and when free. Jennings states that such

protozoa have a more complex behavior than motile

forms. This is due to their combining the types of

behavior of sessile and motile animals,

(f) Response and Taxonomy of Sessile Organisms

Hickson (’98) has stated that there is but one species

of Millepore and believes that sex organs will be found

to be the best taxonomic characters. Wood-Jones states

that there are far fewer species of corals than has

formerly been supposed, and states further that growth

form can not be used to distinguish species. Among

fresh-water sponges and Bryozoa reproductive bodies

(gemmules and statoblasts) have been found to possess

Satisfactory taxonomic characters. This is a situation

quite parallel with that in plants where reproductive

654 THE AMERICAN NATURALIST [Vov. XLVIII

organs are used as classification characters. The ideas

of the reproductive organs of plants are now at the

“fixity”? stage which on the animal side is paralleled by

the idea of fixed tropisms and fixed instincts, of a few

years since. Variability of tropisms is now well recog-

nized and reproductive organs in plants are being found

plastic, as those of animals will probably be found also.

III. PARALLELISM BETWEEN SESSILE AND MOTILE ORGAN-

ISMS WITH REFERENCE TO ECOLOGY

From a summary of the considerations above it will

be seen that for practical comparison the division of

organisms into plants and animals may be abandoned and

only reference to sessile and motile organisms made. We

may now turn to a discussion of a few general principles

making the division into sessile and motile organisms only.

The behavior of motile organisms is plastic. There

are innumerable cases of modification of reaction by

variations of physical factors (Jennings, ’06; Loeb, ’06;

Mast, ’11). If for purposes of discussion we put the

usual ‘‘normal’’ reactions of motile animals over against

‘‘normal’’ structure of sessile animals, we note that the

behavior response of the former parallels the structural

response of the latter.

1. BREEDING

Motile Organism Fixed. (Sessile) Organisms

(a) The breeding activities take

place within narrower limits ae

-= other activities. Merri

790; Herrick, 02; Reighard, "08;

Shelford, "11a, b, c, 12a, b.

(b) The selection of breeding

place and breeding activities, in-

eluding first activities the

young, are governed by the same

_ general laws as other activities.

(a) Breeding and other activi-

ties within same limits, except that

dispersal may take place over wide

areas through detachability of

seeds and other reproductive bod-

les.

(b) Less marked because a se-

lection of abode by sessile organ-

isms takes place through the be-

havior of motile young stages or

through wide dissemination of non-

motile bodies by wind (ete.) with

growth under favorable conditions

and failure elsewhere.

No.575] RESPONSES OF PLANTS AND ANIMALS 655

probably least modifiable and least and early embryonic stages are

regulatory. less modifiable than the vegetative

parts.

(a, b,c) The maple tree, a sessile organism, is entirely

stationary in its adult stages. The seeds are blown by

the wind. One would not accomplish much in the study of

ecology by studying the distribution of the seeds of the

maple, or, on the other hand, by the study of the distribu-

tion of adult birds, without some further discrimination.

Sessile organisms are not difficult to associate with

their proper environmental conditions in their adult

stages. As we proceed in our study to forms which can

move readily and rapidly, the difficulty of associating

them with their definite environmental conditions in-

creases. Sessile organisms have stages which are small

and capable of easy dispersal, as in the case of the maple.

Sessile marine animals and some sessile plants frequently

have motile forms in young stages. In these motile

stages they are governed by the same laws as other motile

organisms. The conditions under which the motile stages

develop into the sessile forms are crucial.

Most fresh-water forms and some marine forms of

sessile organisms are without the free-swimming stage,

and they produce non-motile stages physiologically

comparable to the seeds of higher plants. The winter

bodies (statoblasts) of the Bryozoan (Pectinatella) com-

mon near Chicago, and which is a strictly sessile organ-

ism, are comparable to seeds and probably require

‘‘ripening’’ by cold, just as do many seeds and the repro-

ductive bodies of some other species of the same group.

Organisms which are highly motile in the adult stages

are not motile in the egg and young stages. The eggs

and young of birds, for example, do not move about, yet

birds are the most motile of all animals.

2. COMPARISON OF THE SESSILE AND MOTILE ELEMENTS OF

THE ‘Brora

(a) The motile organisms of a (a) The sessile organisms of a

given habitat usually react simi- given habitat (particularly plants)

656

larly to two or more stimuli not

differing greatly in intensity from

their optimum, i. e., the percent-

age of positive or negative trials is

essentially the same for standard

intensities. There is also probably

similarity in the rates of metabol-

ism, ete.

(b) The specificities of behavior

such as the mode of moving the

organs, e. g., of locomotion, and

in some eases the combined results

THE AMERICAN NATURALIST

[Vou. XLVII

usually show coe functional

rates, sue s similar rates of

transpiration among sand dune

plants.

(b) The various structural de-

vices which meet the conditions of

the environment are ecologically

equivalent.

of different behavior reactions are

similar and hence are ecologically

equivalent. The size and efficiency

of the organs are also involved.

A testing, for example, of the rheotaxis of a large num-

ber of brook-rapids animals has shown them to be

strongly positive, and when active individuals only are

considered the percentage of positive trials is very

similar for the entire rapids community. Likewise they

are in accord in their avoidance of sand bottom. Many

of the animals have special means of attachment which

may be brought into play with speed.

As has already been pointed out elsewhere, ecological

equivalence is illustrated here. The darters (fish) are

strong swimmers and are able to live in rapids by virtue

of their swimming powers and positive reaction, while

snails meet the same general conditions through positive

rheotaxis and the strong foot which enables them to hold

to rocks.

3. SESSILE AND MOTILE ORGANISMS IN ECOLOGICAL

SUCCESSION

(a) Ecological succession is succession of ecological

(physiological) types over a given area, due to changes of

conditions which both cause migration of physiological

types and transformation of such types as remain (Shel-

ford, ’1la, ’11b, ’11d, ’12a, ’12b and citations). Changes

of conditions are geographic, i. e., physiographic, climatic,

No.575] RESPONSES OF PLANTS AND ANIMALS 657

etc., and biological (due to organisms). Sessile plants

are the chief biological cause of succession on land and

in fresh water, while sessile animals are the chief biolog-

ical cause in the shallow portions of the sea, especially in

coral reef regions (Wood-Jones, ’11). Sessile organisms

are more important causes of succession than motile ones

because they (a) build up the substratum with detritus

and skeletons, (b) interfere with the movement of the

surrounding medium, (c) cut off light from the sub-

stratum where other organisms must reside and their

own young secure foothold, and (d) they usually affect

their own environments with excretory products more

than do motile organisms. In general we recognize

ecological succession of motile animals through the differ-

ences of behavior which accompany changes in conditions.

The differences are physiological; differences in behavior

are the easiest index of the physiological condition. The

character of nests, burrows, etc., are often good indi-

cators also.

IV. INFLUENCE OF RESPONSE PHENOMENA UPON BIOLOGICAL

THEORY AND CONTROVERSY

A glance at some aspects of biological speculation

since before the publication of Darwin’s ‘‘Origin of

Species’’ is essential to our understanding of the atti-

tude of biologists until recently, toward responses.

1. TELEOLOGICAL View

In the matter of animal behavior response, the earlier

workers interpreted the reactions as intelligent and pur-

poseful, ascribing human sensations, ete., to animals as

low in the seale as protozoa. This teleological tendency

was paralleled on the plant side by the idea of purposeful

adaptive responses. Many common plants respond

(structurally) readily to environmental conditions. As

has been noted, the commonest of the surviving responses

of the wild state are apparently advantageous. This led

some botanists to a Lamarckian teleological conception of

response, perhaps best represented by Kerner and

658 THE AMERICAN NATURALIST [Vow.XLVIII `

Oliver’s work on the natural history of plants. Accord-

ing to this view, responses are advantageous and for the

purpose of preserving the plant. Thus response and

adaptation become synonymous (Coulter, ’08), a usage

quite inapplicable to animal structure. At the beginning

of the recognition of the response phenomena of corals

Wood-Jones takes essentially the view of adaptation

which botanists have tried and rejected.

Lamarck, who was for many years engaged in botanical

work, must have noted many cases of advantageous

structural response in plants. Later he undertook the

study of invertebrates which show great plasticity, and

was naturally much influenced in the development of his

theory of transmutation of species by the response phe-

nomena in the plastic organisms which he studied. Thus

the responses of motile (as well as sessile) organisms

which result from their own activities or the action of

their environments formed an important feature of

Lamarck’s (Packard, ’01; Cope, ’96) theory of transmu-

tation of animal species. His theory is clearly in accord

_ with the material he studied most. The nature of his

contention and various well-known circumstances caused

his ideas not to be accepted.

2, NATURAL SELECTION VIEW

Characters used in classification of motile animals

before and since the time of Darwin are quite frequently

adaptation characters. Thus the large pectoral fins and

absence of an air bladder are characteristics of an entire

group of fishes, the darters. The divided eyes of the

Gyrinide, which swim at the surface of the water, are so

adjusted that one half looks downward into the water,

and the other outward into the air. This character com-

bined with the paddle-like hind legs would have served to

distinguish the family. Again larve with a head and

thorax modified to fit a circular burrow and with hooks on

the dorsal surface of the fifth abdominal segment, which

is supposed to be an adaptation to prevent the animals

No. 575] RESPONSES OF PLANTS AND ANIMALS 659

from being drawn from their cylindrical burrows by

their prey, could serve to distinguish the entire family of

Cicindelide (tiger beetles). Such cases might be multi-

plied indefinitely.

Following Lamarck came Darwin, who, being more par-

ticularly a zoologist, was probably (proportionately, at

least) less familiar with structural response phenomena.

He was apparently impressed with the ‘‘fixity’’ of the

so-called adaptation characters in motile animals, and

with the fact that they are often family, generic or specific

characters. With the assumption that they originated in

the environment in which they are now found, Darwin

and his followers on the zoological side credited ‘‘natural

selection’’ of structural characters with the origin of

species. Though broader than Lamarck, this important

feature of Darwin’s theory was quite clearly drawn from

data on motile animals. After the acceptance of Darwin’s

theory, biologists were for many years engaged in elabo-

rating the ideas of phylogeny and natural selection by

working out recapitulations and homologies and by point-

ing out cases of adaptation. The investigation was

largely confined to the highly individuated animals. The

morphological method of this period, which indeed has

still continued in use among a minority of zoologists and

which finds a parallel in the recent morphological study

of the sex organs of plants, belongs to descriptive rather

than to analytical science. Since its conclusions are often

based upon the arrangement of species or of stages in

development into series chosen by the investigator, it is a

method which often allows free play of subjective fancy.

Thus unconsciously experimental study of modification by

environment became more and more neglected, and the

dominant type of investigation being such as to show

only the usual course of events in development, the ideas

of fixity grew more and more. Thus the fact that the

external form, structure and color of animals are not

easily modified without careful experimental methods,

and that the structural responses of sessile animals were

660 THE AMERICAN NATURALIST [Vou. XLVIII

so little known, resulted in structure in animals being fre-

quently regarded as fixed and every resemblance and

peculiarity being too often regarded as significant. The

explanations of supposed adaptations among animals fell

largely to the theory of natural selection which was

strained by some (see, for example, in Romanes, ’92, p.

269) to explain origins in great detail, largely on the basis

of the competition of species for food, ete. Explanations

along this line were carried to a reductio ad absurdum as

indicated by Livingston (’13) and have by no means dis-

appeared from the scientific calendar. This tendency

was less important on the plant side. More attention was

given to speculation concerning adaptive response.

From a consideration of the facts just presented, we

note that the characters of the two leading early view

points in evolution were no doubt influenced if not actually

caused to crystallize into their peculiar form by the failure

of workers to recognize the entire series of phenomena

which we have presented above. Thus a review of the

responses of sessile and motile organisms throws much

light on the influences leading to the first conceptions and

later modification of these two leading doctrines. Botan-

ists for many years dwelt mainly on the response of sessile

organisms and crystallized a Lamarckian conception of the

origin of adaptations through the fixing of advantageous

responses as hereditary characters. During the same

period zoologists essentially ignored sessile and other

multiple individualed animals and their great plasticity

and crystallized the Darwinian idea into Weismannian

germplasm doctrine based on highly specialized single

individualed animals.

3. SUPPOSED Non-INHERITANCE OF RESPONSE AND THE

ERM Puasm DOCTRINE

The theory of the independence of the germ-plasm from

the soma, and its continuity from generation to genera-

tion, was brought strongly to the attention of zoologists

in 1885 by Weismann. It was the natural outgrowth of

the methods and theories of the preceding period and

No.575] RESPONSES OF PLANTS AND ANIMALS 661

was largely based upon the non-inheritance of mutila-

tions and the fact that the germ cells of a few organisms

are, morphologically, early differentiated from the soma.

Timing to its influence upon ideas concerning response,

we note that from this viewpoint details of structure

were not of fundamental importance unless traceable to

the germ plasm. Still, structural details were more im-

portant than response, because, with the exception of

instincts, responses were believed to occur independently

of the germ plasm and hence were of interest only on

their own account. Thus the methods used in applying

Darwin’s theory led to neglect of experimental study of

response and culminated in the extreme views of Weis-

mann. The germ-plasm theory or the ideas of heredity

which are associated with it has dominated zoological

thought almost if not quite down to the present day.”

4. THE INFLUENCE OF THE Stupy or RESPONSE ON

Present-Day BrotocicaL THEORY

One of the most striking developments of recent years

has been the discovery that behavior responses are modi-

fiable to a high degree. Small traces of reagents reverse

2 Unconsciously suggestions of the supernatural which come up in connec-

tion with heredity and evolution have stimulated investigators to study and

speculation, though they have often approached the question of heredity with

an unscientific attitude. This is indicated by ‘such statements as ‘‘I could

not, however, resist the temptation to endeavor to penera Fed mystery of

this most marvelous and com a chapter of life’’ and ‘ momentous

issues involved’’ and ‘‘no more crepes problem oe well be

stated’’ bear out this sta HH ardeney which appears here and

elsewhere in the discussion of A questions, ne to the writer to

be associated with the discussion of pL which can not be referred

to existing facts for solution. Few the ial caveat of scientific

men acquired a working kaowledge of the methods of science before the

age of twenty-five years, and the early habits of mind were formed in the

atmosphere of the supernatural and dogmatic, which has characterized

human thought for centuries. It is doubtful if the majority o

stantly come back to our tests and principles. This may account for many

of the contradictions regarding scientific safer te which one finds in the

conversation of scientific men. When the methods of science have become

the methods of society we may expect a ‘ile of scientific men far more

effective than we ourselves can hope to be. $

662 THE AMERICAN NATURALIST [Vou. XLVIII

reactions. Intelligent behavior occurs in the lower

Arthropods. Even Paramecium shortens the time re-

quired to turn around in a tube, by repetition. Actions

formerly regarded as instinctive now appear to be mere

innate tendencies perfected by repetition. Thus the ideas

of fixity have essentially disappeared from this field.

The response of organisms to injuries and the general

control of form in the lower groups has done much to

break down the ideas of fixity developed by Weismann

and embryological schools. Thus Child, the leading

American worker in this line, is able to control

size, form, number of eyes in the case of Planarians.

Various writers have found modifications inherited after

several generations of repeated stimulation (see Bateson,

13). The development of anti-bodies (immunity) has

been shown to be a response occurring in connection with

many normal processes. The discovery of responses of

so many types has led to abandoning ideas of fixity even

among students of embryology and genetics. Thus we

note the recent decline of the doctrine of continuity and

independence of the germ plasm and kindred doctrines

and points of view, which constitute the central ideas of

fixity. It will accordingly be profitable to consider some

further facts which make the germ-plasm doctrine un-

necessary.

5. ASPECTS OF THE UNTENABILITY OF THE GERM

Puasm DOCTRINE

The presence of primordial germ plasm is assumed

even in sessile colonial organisms such as plants, cœlen-

terates, and in flatworms, etc., where under certain con-

ditions any small part of the body may give rise to a

complete organism. Here the theory is not needed to

explain the facts.

Child (711) said:

The theory of the continuity of the germ plasm as a system, inde-

pendent of the soma, except as regards nutrition, has played an im-

portant part in biological thought during the last two decades, but I

am convineed that it has led in the wrong direction and that it is re-

No.575] RESPONSES OF PLANTS AND ANIMALS 663

sponsible for many pseudo-problems of heredity and development,

which on the basis of a different theory could never have oceupied the

attention and wasted the energy of biologists. Briefly my position is,

that the gonad primordium is, at least up to a certain stage of develop-

ment, physiologically a part of the individuality as are other organs,

and that its further history of differentiation into male and female

gametes indicates that it becomes specified in a particular direction, at

least partly in consequence of its correlative environment in the or-

ganism.

The independence of the germ plasm is not well sup-

ported physiologically. Thus Wilson (712, p. 163) says

of the effect of prolonged ingestion of alkaline salts by

‘mice:

No obvious changes were evident in the liver, kidneys, lungs, spleen

and intestines but in the testes some extraordinary alterations were

found. These results are of especial interest because as the cells of the

testes except the basal cells are regarded by many eytologists as out of

“coordination with the somatie cells. As a result of these experiments

it would seem that they are more susceptible to changes in reactivity

than the surrounding plasma.

Dungay (713) and authors cited have thrown compara-

ble light on this question.

The facts of embryology themselves are but a pseudo

argument in its favor. The organisms in which continu-

ity is supposedly demonstrable are highly individuated

and their organs highly specialized and many different

organs are early separated from the common mass of

cells. The germ cells thus follow the general law of

development in such animals. The germ plasm is prob-

ably no more independent of other parts of the organism

than is the liver or any other special tissue. ‘‘Germ

plasm” and “‘ germinal continuity,” if such exist, may

thus be merely incidental to the particular type of organi-

zation of the specialized individuals in which they occur.

It should further be noted that on the botanical side

this doctrine of the independence and continuity of the

germ plasm has received little attention and has n

given little credence because ‘‘germ plasm’’ arises from

different tissues and is neither set aside early from the

soma nor is it in any other sense clearly continuous.

664 - THE AMERICAN NATURALIST [Vou XLVIII

Furthermore, the plasticity of plant structures made the

application of the doctrine of natural selection to sup-

posed adaptations untenable, and this type of explanation

has received little more attention with botanists than have

Lamarckian speculations with zoologists. The adaptation

characters of plants can not ordinarily be used as taxo-

nomic criteria (Coulter, ’08).

6. Tur Measure or VALUES IN BIOLOGICAL SCIENCE

One hears reference to pure science as something quite

apart from applied science. It is indeed true that inves-

tigators in pure science are to some degree prompted to

push forward in research by interest in the problems for

their own sakes. But the human mind does not work long

isolated from practical affairs or the main channels of

human interest, and it is doubtful if the pure-science

investigator continues long in this way. Observations are

soon connected up in some way, actual or possible, with

some human interest, be it as remote as the improving of

human stock in remotely future generations. Thus ‘‘ pure

science” defined as investigation for investigation’s sake

hardly exists so far as the pure-science workers are

concerned, but may be best defined as an indirect method

of attacking problems of general importance. It differs

from applied science in that application to practical

problems is not its aim, though the estimated value of

theories and results in ‘‘pure’’ science are often greatly

modified by applicability to practical questions.

Certain problems and groups of facts in biology are

sometimes referred to as fundamental. Some one has

said that a fundamental problem is one the solution of

which biologists have decided will give greatest progress.

It is doubtless true that a few leaders reach such decisions

with regard to particular questions, but the real causes

of their general acceptance as fundamental are social

and imitative. Thus when one investigator or a small

group of investigators arrives at such a decision many

others usually become active along the same lines largely

because it is a popular topic. Thus under the influence

No. 575] RESPONSES OF PLANTS AND ANIMALS 665

of a group of investigators among whom Weismann was

a conspicuous leader, problems of the germ cells, the

ege’s early development, and heredity, became ‘‘funda-

mental problems.’’ They evidently argued that since all

comes from the egg and germ cell, all must be discover-

able in the egg. If germ plasm were as independent from

soma, as completely insulated from environment as con-

tinuous from generation to generation as has been

assumed, the study of germ plasm would be the only way

to the solution of the problems of heredity and evolution.

This follows no matter whether the chromosomes or

almost the entire egg are credited with carrying heredi-

tary qualities; only the postulation of continuity and

independence from soma and insulation from environ-

ment are necessary. If the independence of germ plasm

from soma be accepted even in a weakened and modified

form it follows that studies of somatic characters can at

most be of secondary importance from the point of view

of heredity and evolution. Thus in some quarters the

value of various lines of zoological work has been esti-

mated largely, unconsciously, no doubt, in proportion to

the nearness or remoteness of their relation to the “germ

plasm’’ question.

Thus it is true that in biology as in all other fields

values are measured consciously or unconsciously by

criteria. In recent years another better criterion of value

has made its appearance among zoologists. The germ

plasm criterion already discussed was primarily morpho-

logical; the second is physiological, borrowed no doubt

from physiologists. It measures values on the basis of

the analysis. of the organisms into terms of physics and

chemistry or is concerned with a mechanistic conception

of life in all its manifestations. From this viewpoint the

study of each and every part of the organism is important

because the discovery of laws governing one part is

usually or at least often of general importance. Investi-

gations from this viewpoint have shown that the germ

plasm criterion is clearly illogical in its application to the

study of somatic characters because it is based upon the

666 THE AMERICAN NATURALIST [Vou. XLVIII

tacit assumption that the soma is governed by different

laws from the living matter which makes up the germ

plasm from which it arose. In other words it is assumed

that the germ plasm is so different from the soma that

the discovery of laws governing the soma is a type of

investigation of relatively little significance.

Some criterion of values is of course necessary in sci-

ence as well as elsewhere, and for the sake of argument we

would be willing to accept the second when broadly stated

and the first when broadened and modified so as to accord

with the second as appears to be the case among certain

students of genetics. In other words, problems of the

germ cells, the egg, and heredity, are of much importance

when the germ cells themselves are regarded as dynamic

and in their relations to the dynamics of the organism

as a whole.

Granting that these are true and tenable criteria of

values in present-day biological science, what. is to be the

method of application? Should biology demand that

results be of direct application to these ‘‘central’’ prob-

lems? One has but to look at the history of almost any

branch of science to find that great, if not the greatest,

advances have come through following up results at

points where relations to the central problems of the

period were quite unsuspected, or by the transference

of methods, principles and results from one field to an-

other where relations between the two were not suspected.

Take, for example, immunity and immunization, the his-

tory of which is ably sketched by Adami (’08, pp. 451-

528). It has been known for ages that one attack of many

infectious diseases yields more or less complete immunity

from subsequent attacks. Thus for centuries in India

and the East individuals, chiefly children, have been pur-

posely inoculated with matter or by contact. The prac-

tise grew out of experience showing that diseases thus —

communicated to healthy individuals from weaker ones

are less severe. In 1796 the results of Jenner on vaccina-

tion with cowpox were published. This may have influ-

No.575] RESPONSES OF PLANTS AND ANIMALS 667

enced Pasteur, who over eighty years later laid the

foundation for the modern epoch of development, by

combating a plague of diarrhea in poultry (1880).

During the twenty years following, various investigators

added noteworthy contributions, and about 1900 Ehrlich

and Morgenroth evolved the ‘‘side-chain theory’’ by

which a large number of possible conditions can be pre-

dicted and all the observed facts of immunity explained.

While not expressed in strictly chemical terms, the theory

and the experiments which support it are very important

both practically and theoretically. In recent years the

knowledge of immunity and comparable phenomena have

been greatly extended. Various workers (Pfeffer, Vol.

II, p. 262) have shown similar phenomena in the increased

resistance of plants to poisons, thus making the responses

of plants and animals still more generally comparable.

Most recently workers on problems such as fertilization

(Lillie, 713), standing in close relation to the older germ-

plasm doctrine, have discovered facts belonging to this

field and made use of Ehrlich’s theory to explain the ob-

servations. This development has helped to confirm the

conclusion of some investigators that immunity phe-

nomena represent important features of the chemical

mechanism of life. Adami has remarked,

That a plague of diarrhea in a poultry yard, studied by a professor

of chemistry, should be the seed from which has grown the vast de-

velopment of later years is a strange fact, but a fact nevertheless.

What was the attitude of pure science so called, of

germ-plasm doctrinairies, and biologists generally during

the long period which elapsed before they could make

use of his results? Clearly it was one of indifference, if

not disgust, toward the subject. The probable result of

such attitudes on the progress of the investigation of

immunity phenomena, had it not been for their immense

practical significance, is clear. They could not have

received their proper share of attention. Thus in the

pursuit of the analysis of the chemical mechanism of life

men who sought it directly have failed in this one impor-

668 THE AMERICAN NATURALIST [Vou. XLVIII

tant step, and the chief contribution has come from very

remote indirect methods. Generally speaking the inves-

tigators who choose a direct method of attack often put

themselves somewhat in the position of the chemist who

would make chemical analysis of living matter when his

first step defeats its own purpose by killing the substance

to be analyzed. The failure of exclusively direct methods

is often evident. Still the ability to obtain results by the

method of direct attack, combined with a far too rare

ability to tie with them indirectly obtained data, some-

times gives noteworthy contributions.

It accordingly remains to be seriously considered

whether or not biology can afford to apply criteria to the

measure of the values of investigation. Their application

is of course largely unconscious, but the effects are not

thereby modified. Noteworthy results of their applica-

tion are (a) concentration of work in certain lines indi-

cated by a given criterion, and (b) an actual abandoning

to a large degree of remote and indirect methods of

attacking the problems which the criterion involves. This

means the partial abandoning of the methods for which

pure science stands.

Criteria can be safely used only in a very broad gen-

eral way, and in application more often to past progress

than to current investigation. They are perhaps most

valuable as a guide to individual investigators working

on problems remote from these more or less central

‘‘pure science’’ questions. That some guide should be

in the hands of such workers is beyond question. In the

hands of those attacking the problems directly they often

appear detrimental because they soon take on an extreme

form and become regarded as fundamental. At this

stage they are usually in need of extensive revision. If

the investigator is contributing observations and details

only, he is doing a great service, for, such information is

needed everywhere. If he is able to combine his own and

others results, he almost invariably draws data from all

sources, direct and indirect, far and near. Granted the

No.575] RESPONSES OF PLANTS AND ANIMALS 669

ability to synthesize, the opportunity to use the ability

sometimes comes to those who attack the so-called cen-

tral problems directly. It comes equally often (we believe

more often) to those who have led up to the central prob-

lem from some remote viewpoint, frequently condemned

by the followers of direct method of attack. Granting

the importance of synthesis, if the biologist seeks the

solution of such a problem as the germ-plasm problem,

he should encourage workers to start at points as remote

from the subject as possible, that they may approach it

with new light and from new angles.

In judging the work of another, its value should be

determined more by the (a) strictness of scientific method

used, (b) the thoroughness and completeness of the in-

vestigation, and (c) (and perhaps most important of all)

evidence of ability to synthesize and combine other re-

sults with his own with a view to broader generalization.

It must, however, also be recognized that there are many

biological problems of much human importance, which

must be solved quite independently of the ideal central

problems of pure science.

6. Summary AND CONCLUSIONS

From the data presented above, we note that the doc-

trine of purposeful, advantageous response (including

anthropomorphic ideas) arose from the uncritical non-

experimental study of the responses (structural) of ses-

sile and (behavior) motile animals. The idea of the all-

sufficiency of natural selection is largely the outcome of

observational study of apparently fixed and yet appar-

ently adaptive characters of motile highly individuated

animals. The doctrine of the continuity of the germ

plasm is likewise the outgrowth of the study of highly

individuated animals in which the various organs are

early differentiated in the dividing egg. No one of the

doctrines is wholly tenable; no one is more than a partial

truth. Each appears to have arisen from a recognition

of certain more or less unconsciously selected and un-

critically interpreted phenomena by each of several men

who secured different facts and attempted explanations.

In a few animals the ‘‘germ plasm’’ may be morpho-

logically early differentiated and reasonably continuous,

though governed by the same laws as other tissues. In:

others, any part of the general tissues may give rise to a

complete organism. The behavior of some organisms is

intelligent and purposeful, while that of others is largely

mechanical. Some structural responses of sessile organ-

isms are advantageous, some indifferent and some harm-

ful. Some of the more fixed structures of the highly indi-

viduated animals are advantageous, some indifferent,

and some disadvantageous (Metcalf, 713). No other type

of general statement appears to be tenable, yet each

extreme of each proposition has at some time or other

been the subject of some all-inclusive doctrine.

Such are the limitations of an individual’s knowledge

and the psychic limitations of our race and generation.

In considering the psychology of religion, Ames (’10,

p. 594) points out similar well-recognizable tendencies in

development as follows:

Much energy has been wasted or nearly wasted, in the exclusive and

intolerant advocacy of special schemes—single tax, prohibition, state

socialism and the like, each of which was imagined by its adherents to

be the key of millennial conditions. Every year makes converts to the

truth that no isolated scheme can be a good scheme, and that real prog-

ress must be advanced all along the line.

Advance all along the line is what biological science

must achieve. This I believe means the encouraging of

all lines of indirect attack, whether they at first throw

light on the ideal central question of pure science or

important practical problems or not. It means the exer-

cising of extreme caution in the application of criteria of

values to scientific results. Such measures tend not only

to stifle the best initiative in good investigators, but also

tend to check the building up of fruitful hypotheses.

The latter danger is greatest in connection with the

= mechanistic criterion referred to above: As has already

_ been stated, criteria of values ean be safely applied only

No.575] RESPONSES OF PLANTS AND ANIMALS 671

as broad general guides, and investigation should be

measured on the basis of its thoroughness, the originality

shown, ete.

In science special schemes of course do a exist recog-

nized as such, but intolerant application of criteria of

values results in essentially the same condition. One

often hears the statement made by so-called scientific

men, that this or that line of investigation has been pur-

sued for several years, but has failed to yield important

advances or generalizations, but they add, we will be very

glad to recognize it as soon as its value is proven. This

seems to us to be a distinctly unscientific attitude, and

but a polite modern statement of a spirit which in former

generations often sent men to the stake or dungeon. This

is true because to these oe objectors its value is

rarely or never proven. It is ‘‘schemes’’ (preconceived

theories) thus presented that have in the recent past

stifled the study of responses by discouraging efforts in

that direction and thus contributed materially toward

making zoology the unorganized science which it is

to-day. We must recognize that the various aspects of

zoology pure and applied have never been well corre-

lated, less so we believe than in any other branch of

natural science, clearly less than in botany. In general.

animal physiology has been isolated in medical schools

and genetics, faunistics and morphology have not been

properly influenced by it, while morphologists for many

years held themselves aloof from other workers.

In a discussion dealing mainly with the doctrine of

natural selection in the origination of adaptations,

Mathews (713) has sounded the keynote of a growing

attitude toward all response questions. Out of the infi-

nite different combinations which may enter into the

proteid molecule and the varying rates at which metabolic

action may go forward, innumerable types of irritability

and correlated structure have been and still are arising

under the influence of environment external and internal.

Of these some are incompatible with life, others indiffer-

672 THE AMERICAN NATURALIST [Vou. XLVIII

ent, and others advantageous. Upon these physiological

characters natural selection has operated to eliminate,

and with time has perhaps rendered of less frequent

occurrence, those characters that are incompatible with

their conditions of existence. External form, color orna-

mentation, ete., while no doubt often of importance them-

selves are more often the advantageous or indifferent

correlatives of physiological or irritability types which

are compatible with their conditions of existence. The

study of irritability and response may be pursued in

many ways—by experiment, by observation in nature

alone or combined with experiment. The mapping of

stimulating conditions in nature, of the distribution of

types of irritability and response, which is one function

of field ecology and modern geography, can hardly fail

to contribute materially to the advance of knowledge in

-many lines, including that of the physico-chemical

mechanism of life. The student of experimental ecology

has an infinite field of problems and methods thrown

open to him by the organization of such information

relative to responses. Still in our attempt to make ad-

vances along the line of the study of responses, we must

not forget that it is but one of several lines of advance,

all of which must sooner or later be correlated with a

view to broader generalization.

HULL ZOOLOGICAL LABORATORY,

UNIVERSITY OF CHICAGO

April 1, 1914

LITERATURE CITED

Adami, J. G. (’08). Principles of Pathology; Vol. I, General Pathology.

Philadelphia. i

Adams, Chas. C. (’13). Guide to the Study of Animal Ecology. Chaps.

II, IX. New York.

Ames, E. S. (710). The Se of Religious ees, New York.

Bateson, W. (°13). Problems of Geneties. New Hav

Brooks, W. K. (’93). Salpa in its Relation to its Nvotitibi of Life.

Johns Hopkins Univ. Studies. Bicl. Lab., Vol. 5, pp. 129-211.

Child, C. M. (’04). Studies on Regulation, II. Jour. Expt. Zool., Vol. II,

pp. 95-1

(711). A Study of Senescence and Rejuvenescence Based on Experi-

ments with Planaria dorotocephala. Roux. Arch., XXXI,

pp. 537-616

No.575] RESPONSES OF PLANTS AND ANIMALS 673

(713). — Dynamic Factors in Experimental Reproduction.

x. Arch., Bd. XXXV, pp. 598-641.

Coulter, J. oy (*00). The Theory of Natural Selection from the Stand-

point of Botany. Fifty Years of Darwinism, pp 75.

Cowles, H. -z (711). A Text-book of Botany. Vol. II, Ecology. New

ork.

Cope, E. D. eae Primary Factors of Organic Evolution, Chicago.

Davenport, C. B. (’91). Cristatella: The Origin and Development of the

SpE in the Colony. Bull. Mus. Comp. Zool, Vol. 20,

-152.

PP.

Delage, Y., and Heron, E. (’96). Les Vermudiens, Zoologie Concrete

BY:

Dungay, N. S. (13). Pe Study of the Effect of Injury on Fertilizing

Powers of Sperm. Biol. Bull., Vol. XXV, 213-260,

Forbes, S. A. (’80). Some Interactions of Organisms. Bull. I. St. Lab.

Nat. Hist., III, pp. 3-1

Goode, J. P. (’04). Human Responses to the Physical Environment.

Jour, of Geography, III, pp. 333-343,

Goebel, x (?°00). Organography of Plants. Part I, English ed., Oxford.

Harmer, S. F. (’01). Polyzoa. Cambridge Natural History, Vol. TI,

pp. 465-534.

Herdman, W. A. (’04). Ascidians and Amphioxus, l. c., Vol. VII, pp.

35-138.

Herrick, F. H. (’05). The Home Life of Wild Birds. New York.

Hidtson, S. J. (’98). On the Species of the Genus Millepora. Proc. Zool.

Soc. London, 1898, pp. 241-257.

Holmes, S. J. (’07). The Behavior of Loxcphyllum and its Relation to

Regeneration. Jour. Expt. Zool., Vol. IV, pp. 399-418.

(711). Evolution of Pol Intelligence. New York.

Henderson, L. J. (’13). The Fitness of the Environment. Am. NAT.,

eae ia 105-115.

Jennings S. (706). Behavior of the Lower Organisms. New York.

Jo dhnson, My E. irons A Qualitative Study of the Development of

e Salpa Chain in Salpa fusiformis runcinata. Univ. of Cal.

ac in Zool., Vol. 8, No. 7, pp. 145-176.

Johnstone, J. (’08). Conditions of Life in the Sea. Cambridg

Klebs, Œ. (’03). Willkiirliche Entwicklungsiinderungen bei Pflanzen. Jena.

Lillie, Frank R. (’13). The Mechanism of Fertilization. Science, N. S.,

Vol. XX

, pp. 524-528.

Livingstone, B. E. (13). PEE in the Living and Non-living. Am.

Nar., XLVII, pp. 7

gl J. (706). ‘Dynamics of Pia Matter. New Yor

ugal, D. T. (’11). Organic Response. AM. NAT., Tok XLV, pp. 5-40.

‘ied S. O. (’11). Light and the Behavior of Organisms. New York.

Mathews, A. P. (’13). Adaptation from the Point of View of the Physi-

ologist. Am. Nart., XLVII, pp. 90-105.

McGee, W J (’96). The Relations of Institutions to Environment. Smith.

Rep., 1895, pp. 701-711.

Merriam, C. H. (90). Results of a Biological Survey of the San Fran-

674 THE AMERICAN NATURALIST [Vou. XLVIII

cisco Mountain re ate and the Desert of the Little Colorado.

North American Fauna 3.

Metcalf, M. M. (713). kantita through Natural Selection and Ortho-

genesis. AM. NAT., Vol —

Minchin, E. a tts Sponge. Takata: s Treatise on Foalog’: Part II.

Moore, H. F. rey “The Commercial Sponges and Sponge Fisheries. Bull.

S. F. C., 1908, pp. 403-511

Motz-Kossowska, 8. (708). Quelques asiaa à propos de Plumularia

lichtensterni Mark eim et les via affines. Arch. Zool. Exp.

et Gen. Notes et Revue, 4, T. IX, pp. 55-59.

Packard, A. (’01). Lamarck the AFE of Evolution. New York.

Piitter,-A, (711). on ae Physiologie. Jena.

Reighard, Jacob (’1 ethods of Studying the Habits of Fishes with

an ae ae the Breeding Habits of the Horned Dace.

Bull. Bur. of Fish., XXVIII, 1908, pp. 1111-1136.

Ritter, W. E. (711). The onkaan between Materialism and Vitalism.

Can it be ended? Science, N. S., Vol. XXXIII, No. 847, pp.

437—441.

Romanes, G. J. (’92-’97). Darwin and After Darwin. 3 vols. London.

Shelford, V. E. (’lla). Ecological Succession, I. Stream Fishes and the

Method of Physiographic Analysis. Biol. Bull., Vol. XXI,

9-35.

PP.

(11b), ee Succession, II. Pond Fishes. Biol. Bull., Vol.

I, pp. 127-151,

(’1l1c). GEE SG eS Animal Geography. Jour. of Morph. (Whit-

an Vol.), Vol. XXII, pp. 551-618.

(711d). Taigaa Succession, IIT. A Reconnaissance of its Causes in

Ponds with Particular Reference to Fish. Biol. Bull., Vol.

XXII, pp. 1-38.

(’12a). Ecological Succession, IV. Vegetation and the Control of

Land Communities. Biol. Bull., Vol. XXIII, pp. 59-99.

(7126). Ecological Bora. V. Aspects of Physiological Classifi-

cation. Biol. Bull., Vol. XXIII, pp. 331-370.

Strassberger, E., Schenk, Noll F., and Schimper, A. W. F. (’03). A Text-

oak of Botany. Translation by Porter and Lang. London.

Stockberger, W, W. (°13). A Literary Note on the Law of Germinal Con-

tinuity. AM. Nat., XLVII, pp. 123-128.

Washburn, M. F. (09). The prng ESR Chap. I. Yor

Weismann, A. (’91). 92, 2v. Essays upon Heredity and Faa Biolog-

ical Problems. ATE Translation. E. B. Poulton & A. E.

Shi pley. Oxford.

(°93). The Germ Plasm a Theory of Heredity. Translation by N. W.

Parker

& H. Rönnfeldt.

Wilson, F. P. (712). Cell Stimulation by Means of Prolonged Ingestion of

Alkaline Salts. Biochemical Journal, Vol. IV, p. 162.

Wood, Jones He 4D: On Growth Forms and spe apee Specter of Corals.

ool. Soc. London, i pp. 518-556

(733), nie e Atolls. Londo:

AN APTEROUS DROSOPHILA AND ITS

GENETIC BEHAVIOR

CHARLES W. METZ

DEPARTMENT OF ZOOLOGY, CoLUMBIA UNIVERSITY

Among the various mutants of the fruit-fly, Drosophila

ampelophila, which have arisen from cultures in this

laboratory, is one entirely destitute of wings, and hence

called apterous.! The study of the heredity of this form

has been difficult because of its almost complete (appar-

ent) sterility. In order, therefore, to determine to which

of the three groups of linked characters of Drosophila it

belonged I was obliged, in most crosses, to make use of

heterozygous flies that carried the factor for apterous.

As this process is unique in certain regards, it will be

described i in some detail.

METHODS or STUDY

At first it was thought that the apterous mutant was

completely sterile, since none of the first flies, as they

appeared occasionally in certain cultures, could be crossed

even with normal individuals. At last, however, offspring

were obtained from an apterous female by a wild male,

and a permanent line started. But this line could not be

perpetuated by means of apterous individuals, for these

were unable to breed.? It had, therefore, to be kept up

by means of heterozygous, winged flies. The method was

as follows: The original cross of winged by apterous gave

in F, approximately 3 winged to 1 apterous. Of the

winged class approximately two thirds were heterozygous

for apterous, and when mated together gave the same

1 This apterous fly is quite distinct from that called wingless in earlier

papers by Morgan, and now known as vestigial.

2 Only twice, aside from the original mating, were apterous individuals

successfully crossed, and then only to winged specimens, never to their own

kind. These two cases are given in experiments II and ITI.

675

676 THE AMERICAN NATURALIST [Vou. XLVIII

3:lratio. Selecting again from the winged flies, the proc-

ess could be repeated indefinitely. The only difficulty lay

in the fact that no visible character differentiated the

heterozygous from homozygous winged flies, and conse-

quently all matings had to be made in pairs taken at

random, with the result that about 56 per cent. of the

cultures were rendered worthless. In actual practise large

numbers were mated in pairs, and then all discarded save

those producing apterous.* This was the method used in

keeping up stock.

To obtain the necessary combination of apterous with

other mutant factors, winged offspring from apterous-

throwing parents were mated in pairs to flies of the

desired stock. One third of the normals from apterous

stock were pure for the normal allelomorph of apterous

and rendered worthless all matings in which they were in-

volved; but the other two thirds were heterozygous for

apterous, and when crossed with the desired stock gave

in F, some apterous offspring. If the F, flies were bred

en masse, approximately 15 winged to 1 apterous were ob-

tained, but if bred in pairs, certain pairs (those in which

both members were heterozygous for apterous) gave 3

winged to 1 apterous. The latter method was the one

actually used in most cases. In this manner the same end

result was attained as would have been secured by using

apterous individuals in crosses with other stocks, the

only difference being in the amount of labor involved in

making up a larger number of cultures. Both kinds of

crosses were, in fact, used, as will be seen below.

The use of symbols in this paper follows the system

recently adopted by Morgan and other students of Droso-

phila (Morgan, 1913, a and b). That is, for any pair of

allelomorphic characters a capital letter is used to indicate

the dominant, and a small letter the recessive factor—the

symbol being taken from the name of the mutant. Since

the apterous character is recessive, the symbols for the

3 In the fourth experiment a character (black) was introduced which dif-

ferentiated homozygous from heterozygous and thus made it possible to pick

out the heterozygous individuals.

No. 575] AN APTEROUS DROSOPHILA 677

apterous fly become arap, and those for the winged fly

v-Ay. In other words, Ap is a factor in the wild fly

necessary for wing production, while a is its modified

homologue responsible for lack of wings in the mutant.

The apparent contradiction in using Ap, not for the factor

responsible for apterous, but for its normal allelomorph,

may be confusing at first sight, but a little familiarity with

the system obviates this difficulty.

EXPERIMENTS

Experiment I.—Long-winged, red-eyed & by apterous,

white-eyed ? (from miniature wing stock).

F, All winged. Long-winged, red-eyed females.

Miniature winged, white-eyed males.

F, Winged and apterous as follows:

Long-winged, red-eyed males and females.

Long-winged, white-eyed males and females.

Miniature-winged, red-eyed males and

females.

Miniature-winged, white-eyed males and

females.

Winged

Ast Apterous, red-eyed males and females.

i as Apterous, white-eyed males and females.

This experiment shows the inheritance of the apterous

character to be Mendelian, giving in F, all winged, and in

F, approximately 3 winged to 1 apterous. Table I con-

tains a summary of the offspring from 21 pairs of the F,

and F, individuals, giving a total of 1,405 winged to 450

apterous,—a ratio of 3.12 to 1.

The absence of apterous flies in F, indicates at once

that the apterous character is not sex-linked. The pres-

ence of miniature-winged flies in F, and F, indicates that

the apterous factor is independent of the miniature-wing

factor, which latter must have been carried by the apterous

female (coming from miniature wing stock), and trans-

mitted to her offspring unaffected by the apterous factor.

678 THE AMERICAN NATURALIST [Vou. XLVIII

Analysis of the cross:

Ap, factor necessary for wing production. ap, its alle-

lomorph, in the apterous fly.

M, factor necessary for the production of long wings

(sex-linked).

, allelomorph of M responsible for miniature wings.

factor necessary for the production of eye color

(sex-linked).

w, allelomorph of W responsible for white eyes.

P, Long, red male A,MW X—Ap,

Apterous, white female a»mwX—apmwX.

F, Long red females AMW X-amwX,

Miniature white males AÁprapmwX.

F., leaving out of account the sex-linked factors and

considering only winged vs. apterous :

Gametes of F, Ar, ap

Ap, Ap.

F. Ag-Co.

Winged 4 Ap—Ap>.

pp.

Apterous dp—dp.

Experiment II.—Long vermilion Ẹ by apterous, white 3.*

This cross is practically the reciprocal of Exp. I, except -

that vermilion replaces red eye color in the winged parent.

Like Experiment I it involves two pairs of sex-linked char-

acters, aside from the apterous character. The results are

essentially like those of Experiment I and may be passed

over briefly.

P, Long, vermilion? A»MWX-—A,MWX,

Apterous, white J apmwxX—dp.*

4 The white-eyed, apterous ¢ in this cross is white-vermilion, i. e., the

double recessive, and therefore when crossed with vermilion it gives vermilion

instead of red in F,.

No. 575] AN APTEROUS DROSOPHILA 679

F, Long, vermilion? A»MWX-amwX,

Long, vermilion ¢ A»MWX-«.

F, Long vermilion Ẹ and g

Long, white g :

Miniature, vermilion ¢ Winged.

Miniature, white ¢

Apterous, vermilion 2 and ¢

Apterous, white | Apterous.

TABLE I

OFFSPRING FROM PAIRS HETEROZYGOUS FOR APTEROUS IN EXPERIMENT I.

PARENTS TAKEN FROM F,, F, and F,

Mating No. | Winged Apterous Ratio

41 | 122 32 38:1

53 | 4 14 2.9 :1

56 | 29 12 25:1

64 | 46 30 15:1

65 | 29 16 1.8 :1

80 85 28 3.04: 1

83 71 20 < i

84 13 3 43:1

85 | 40 9 4k: i

111 | 183 64 98:1

112 20 10 S34

117 42 24:1

118 | 28 10 28:1

130 32 7 45:1

131 | 76 26 3.1:1

132 71 19 37 +1

134 92 36 2.6 :1

151 | 58 12 48:1

170 | 78 10 78-1

171 63 15 | 4.5 :1

77. 129 35 | 3.6 :1

| 1,405 450 |

Average ratio of winged to apterous, 3.12:1.

An analysis of the F, is not essential here and is omitted

for the sake of brevity. It may be derived from the F,

formulæ. Table II indicates the expected classes and

ratios in F, and gives the actual numbers obtained in cul-

ture No. 59, in which each class was recorded separately.

In subsequent cultures of this experiment no attempt was

made to separate any but the winged and apterous classes.

Counts of the latter are given in Table III.

680 THE AMERICAN NATURALIST [Vou. XLVIII

TABLE II

MATING 59

Fz — Expected Classes Expected Ratio Aotus Ralio Actual No.

TOUR Va Vy aA RS | 12 138:1 185

Lone VOUS Ooo Ses Oa eee ee i 3 4.251 57

HW WHILE OW is. 3s ere. 3 24:1 31

Minintiire vor: ol ee oo ve ens 3 eee 26

Miniature white oi cee eee | 3 Ari 53

Aptirons vor. 9.60. ee as sk | 4 Saek 44

Apterous verm. Goose veecccceessves | 2 1221 15

ADCOPOUR WHITE: OS oe ask sa ee es 2 a ey : ane

TABLE III

F, FROM MASS CULTURES

Calture Wo, | Winged’ | -Apearouk Ratio

58 367 78 4:7 onl

59 352 3 By Beso |

96 556 114 48:1

7 306 70 Ao 21

120 597 103 5.08: 1

135 554 104 ES 203

137 405 49 6 Si

298 6:6 $1

157 405 53 16.2

3,840 699 Average ratio, 5.5 : 1.

F, FROM PAIRS

Culture No. Winged Apterous Ratio

155 141 35 AL Sk

156 38 11 G4 2:3

179 46 Average ratio, 3.9 : 1.

F, FROM PAIRS

Culture No. Winged Apterous Ratio

160 119 25 4.7531

162 123 27 46:1

172 266 62 CS: eau |

173 87 26 SA st

174 165 37 45:1

175 167 42 4: 1

176 171 46 SF BL

77 129. 8.0: 1

178 92 23 A tl

1,319 323 Average ratio, 4.08: 1

179 46

1,319 323

Total from pairs, 1,498 369 Ratio, 4.6 : 1.

No. 575] AN APTEROUS DROSOPHILA 68 |

It will be noted that the apterous classes fall a little

below the expected numbers in most cases. This is char-

acteristic of all weak races of Drosophila, and is doubtless

due to the inability of some flies to mature. Of the winged

classes the first two and the fifth exceed the Mendelian

ratios, while the other two fall short, due to linkage be-

tween white, vermilion and miniature.’ The distribution

of apterous, however, is entirely independent of the

others, showing that the factor responsible for it is not a

member of the group containing those responsible for

vermilion eyes, white eyes, or miniature wings.

The ratio of winged to apterous in this particular cul-

ture is 4.2:1. Table III includes a summary of this and

nineteen similar cultures in which the parents were all

descendants of the long-winged, vermilion-eyed female by

the apterous male mentioned above. The first nine are

mass cultures, the next two are pairs, from F, flies. Below

these are offspring from nine pairs of F, flies.

It is noticeable that the ratio of apterous to winged is

greater in cultures where pairs are used than in mass cul-

tures, though all parents in the latter are heterozygous.

This, I believe, is unquestionably due to the low viability

of the apterous flies, which prevents some of them from

maturing in cultures where the competition is severe. For

this reason the averages are given separately for pairs

and for mass cultures. The average from pairs is 4.06: 1,

while that from mass cultures is 5.5:1. This low viability

is also shown by pairs, if the food conditions are not good,

or if the culture becomes very dry.

Experiment III.—To determine the relation between

apterous and characters in Group ITI.

It is obvious from Experiments I and II that apterous

is not a sex-linked character (Group I). The present ex-

periment is for the purpose of determining its relation to

characters of Group III. As a representative of the latter

group pink eye color was chosen. The results of the cross

between this and apterous may be passed over briefly

5 For discussion of linkage between these characters see Morgan, 1911.

682 THE AMERICAN NATURALIST [Vou. XLVIII

since they are similar to previous results in showing no

linkage. A winged, pink-eyed male bred to an apterous,

vermilion-eyed female (from Experiment III) gave, as

expected, winged, vermilion-eyed males and winged, red-

eyed females in F,. These inbred gave four classes of

winged and four classes of apterous, i. e., red, vermilion,

pink and orange.® The ratios are such as to show inde-

pendent segregation of apterous and pink. Below is a

summary of the expected and actual results.

P, Apterous, vermilion 2 apvPX-avPX,

Winged, pink ¢ ApV pX—App.

F, All winged. Red 9 apvPX—ApV >X,

Verm. 5 adpvPX—App.

Eight kinds of eggs and four kinds of spermatozoa are

formed by these F, flies, giving, through random fertiliza-

tion, 32 classes of oe divided into eight groups, as

shown i in oe

TABLE IV

F Expected Results Actual Results in Experiments te

Classes | Ratio | 627 | 6287 | 629 | 630 | 681 | Total

baer COS oi cee ic: 9 85 | 402

inged, vermilion.. >... gie ieg] S | sa n ne = |

waar DINE Ee 4 3 25 | 4 | 111

Winged, orange......... gf" 110s" | , pO : |

Apterous, rod. io. 004 2.3 3 22 | | 114

pterous, vermilion...... 3 . 13 a | a pa D i bic.

Apterous, pink.......... 1 T 7 34

Apterous, orange........ 1 1 8 15 | 9 i9 7

In the table red and vermilion have been considered

together as one class, because they both contain P; and

similarly pink and orange have been considered together

because they both contain p. The total numbers for the

four classes give the ratios 402:111:114:34, or 10.8:3.6:

3.35:1,—a sufficiently close approximation to the expected

6 Orange eye is the double recessive pv-pv. :

7 In this culture the ratios are seen to diverge widely from the expected,

due, I believe, to the poor cultural conditions in this case which prevented

_ some of the weaker pink and orange flies from maturing

No. 575] AN APTEROUS DROSOPHILA 683

9:3:3:1. These results clearly show the independence

(i. e., lack of linkage) of apterous and pink.

Table V includes all matings (giving apterous) in Ex-

periment III, for the purpose of showing the ratio of

winged to apterous.

TABLE V

Mating No. Winged | Apterous | Ratio

627 178 50 | 3.56 : 1

628 37 15 2.47:1

629 158 42 | 8.7 :1

630 46 13 3.63 : 1

631 94 28 3.326: 1

646 48 10 | 4.5.42

649 81 27 Soko S

650 44 14 3,15: 1

651 124 37 3.35 : 1

652 92 34 ay ae |

GA 50 | 19 2.63: 1

692 75 | 27 3.15. 1

1,027 | w eee

Average ratio winged to apterous, 3.25:1.

Experiment IV.—To determine the relation between

apterous and characters in Group II.

The mutant called ‘‘black’’? (having black body and

wings) was used in this experiment as a representative of

Group II. No direct matings with apterous individuals,

such as obtained in the three preceding cases, could be

effected here, and consequently the winged brothers and

sisters of apterous had to be used for crossing with black,

according to the method described in the introduction.

Matings of this kind (in pairs) gave, in F,, winged flies,

some of which were heterozygous for apterous and black.

These inbred (also in pairs) gave winged and apterous,

and gray and black, as shown below:

P, Black ¢ homozygous for wings Apb—Apb,

Gray 9 heterozygous for apterous A,B-apB.

F, Winged, heterozygous for black A yb—ApB,

Winged, heterozygous for black and apterous A:b-%B.

Only pairs in which both members were of the second

type (4,b—a»B,—heterozygous for apterous) could pro-

684 THE AMERICAN NATURALIST [Vou. XLVII

duce apterous. The others, therefore, are ignored. Con-

sidering the second type alone, the analysis becomes:

F, gametes (expected) A,b—A,B-apb—apB,

A po—A p B —p b-a B š

F, Expected classes.

apB-apB

apterous, gray.

a>B-apb apterous (heterozygous for black).

B-B winged, gray.

aB-Ayb winged, heterozygous for black.

apb-arpB apterous, heterozygous for black.

apb-@b - apterous, black.

ayb—ApB winged, heterozygous for black.

apb—Apb winged, black.

A,B-aB winged, gray.

A,B-ayb winged, heterozygous for black.

A,B-A,B winged, gray. -

A,B-A,b winged, heterozygous for black.

Arb-apB do.

Apb-apb winged, black.

A,b-A,B winged, heterozygous for black.

Ayb-Apb winged, black.

Expected ratios: 3 winged black; 6 winged heterozygous

for black; 3 winged gray; 1 apterous black; 2 apterous

heterozygous for black; 1 apterous gray.

Actual results: winged black, winged heterozygous for

black, and apterous gray, as shown in Table VI.

In the last two matings black and heterozygous off-

spring were counted as one class.

Total: winged 701; apterous 174 or 4.02:1,

The expectation for the F, if Ap and B segregate inde-

pendently is equal numbers of black and gray among the

winged and among the apterous offspring. Actually, how-

ever, the apterous flies are all gray, and the winged flies

are all black or heterozygous for black. Furthermore, the

No. 575] AN APTEROUS DROSOPHILA 685

ratio of heterozygotes to pure blacks in the winged class

shows that the flies which should have been gray accord-

ing to expectation have been added to the heterozygotes.

Likewise the gray flies in the apterous class are about four

times as numerous as anticipated, showing that the ex-

pected heterozygotes and blacks are here gray. From this

it is evident that the factors a and B, on the one hand,

and A» and b, on the other, have remained associated in

the combination which they formed in the parents, instead

of independently segregating. Such an explanation ac-

counts for the absence of A,B and apb gametes in the F,

generation, and consequently for the absence of gray,

winged flies, and of black or heterozygous apterous flies

in F,. The evidence accords with that obtained for many

other mutant characters in Drosophila, and the explana-

tion is the same as that given for the previous cases (e. g.,

Morgan, 1911, 1912; Morgan and Lynch; Sturtevant, 1913

a and b; Dexter).

TABLE VI

ACTUAL RESULTS

ete pm tee Witt ‘Bee Apterous SS

Mating | ER |

w i Bi A TBD 0

Logie s 22 | 71 s i 0 0

78a ee Â 20. | 54 15 G4 0

i o W 29 | 59 15 a 0

741 0 19 | 50 19 ae 0

745 0 35 86 30 oo 0

746 0 15 42 13 OY 0

| ae ak l oN |

—____#.__*

725 53 14

734 79 24

The presence of a definite linkage or association be-

tween apterous and black (i. e., between either ap or Ap

and b or B depending upon the nature of the cross) as

shown by this experiment, together with the absence of

any such linkage with characters in Groups I and ITI, as

shown by the preceding experiments, indicate that apter-

686 _ THE AMERICAN NATURALIST [Vou. XLVIII

ous is a member of Group II and is, presumably, asso-

ciated or linked with all other characters belonging to

that group.

Supposedly this association of the characters results

from an association of the factors responsible for them

in the germ cells. And this latter association has been

explained upon the assumption that factors responsible

for linked characters are located in the same chromosome.

The hypothesis has even been carried so far as to postulate

a linear arrangement of the factors within the chromo-

some—the relative position of the factors being deter-

mined by the degree or amount of linkage existing between

them. This conception and the data upon which it is

based have been amplified by Morgan and by Sturtevant,

and need not be dwelt on here. In the present case no

attempt has been made to ascertain the exact degree of

linkage between apterous and other characters in the

group, except black, because of the difficulty of breeding

the apterous flies. Apy tly the linkage between apter-

ous and black is very close, if not complete, since no case

of ‘‘crossing over” was observed among the 875 F, off-

spring in this experiment (Table VI). There is a possi-

bility that the classification of the F, apterous flies as all

gray is not absolutely correct, because, owing to the diffi-

culty of distinguishing gray from heterozygous black in

apterous specimens, an occasional heterozygous fly might

have passed for pure gray. However, if there had been

any appreciable number of cross-overs in this direction,

there would also have been some in the opposite direction,

which fact would have been indicated by the presence of

winged, gray flies. And since none of these were observed,

it is safe to conclude that few or no cross-overs occurred,

and hence that apterous is very closely, if not completely,

linked to black.

Experiment V.—To determine the relation between the

apterous mutant and the ‘‘vestigial’’ mutant.

Among the mutant characters of Group TI is one called

‘‘vestigial wing.” Flies having this character are more

No. 575] AN APTEROUS DROSOPHILA 687

like the apterous individuals than are any of the other

mutants, and since the two characters belong to the same

group the question arose as to whether or not the factor

responsible for one might be simply a modification of that

responsible for the other. Experiment V was performed

to determine this point.

Long-winged flies heterozygous for apterous were

crossed with vestigial winged individuals, and F, and F,

generations raised. The F, flies were all long winged,

which fact in itself indicates the independence of the two

characters, for if they were allelomorphs either apterous

or vestigial should have appeared. In F, both vestigial

and apterous, as well as long-winged, flies appeared,

showing conclusively the independence of the two

characters.

Summary OF EXPERIMENTS

Experiments I and II show that the apterous character

is a simple Mendelian recessive, which independently

mendelizes with miniature wings, white eyes and ver-

milion eyes, and hence is not sex- linked (i. e., not a mem-

ber of Group I).

Experiment III shows that the apterous factor is trans-

mitted independently of the factor for pink eye, thus indi-

cating that apterous is not a member of Group III.

Experiment IV shows a linkage ratio to result from

crosses involving apterous and black, the ratio being such

as to indicate a very close linkage between apterous and

black, and to identify apterous as a member of Group II.

Experiment V shows apterous to be distinct from ves-

tigial wing, to which it bears a considerable degree of

resemblance.

ORIGIN AND CHARACTERISTICS OF THE APTEROUS MUTANT

The description of the apterous fly has been deferred

up to this point in order that it might be combined with a

discussion of the experimental results.

The mutant has appeared upon several occasions, but

688 - THE AMERICAN NATURALIST (VoL. XLVIII

always in the same stock (miniature white), and always

with the same evidence of weakness and low viability.

Fig. 1 is a camera drawing of a typical specimen, made

by Miss E. M. Wallace. In morphological characters the

fly differs from the normal in

being entirely destitute of wings

and in possessing greatly re-

duced balancers. Likewise in

Ra panos pe or physiological characters it devi-

ates strikingly from the normal.

This is best shown by means of a comparison between

apterous flies and normal flies from which the wings

have been removed. The latter are not appreciably

inconvenienced by their loss of wings; they show char-

acteristic vigor in their active running and jumping

movements, they easily right themselves if overturned,

or extricate themselves if entangled in food or cotton, and

they are long lived and breed as prolifically as do winged

individuals. In fact they show no ill effects except the in-

ability to fly. The true apterous individuals, on the other

hand, show marked abnormalities in all these respects. In-

stead of being vigorous and active they are weak and usually

sluggish; if overturned they have great difficulty in right-

ing themselves; or, if entangled in food or cotton, they are

usually unable to extricate themselves and consequently

perish. Moreover, they are always short lived, even when

kept under the best possible conditions and prevented from

becoming entangled in food or cotton. And lastly they

exhibit a most marked inability to breed, as noted in the

experiments. This characteristic, as has been mentioned

above, is so marked that the apterous flies were at first

thought to be sterile. I am convinced now, however, that

the difficulty is not one of sterility at all, but is due to a

physical weakness which makes it extremely difficult for |

the flies to copulate, and for the females, even when fertil-

ized, to produce and lay eggs. Cytological examination

has shown that the males produce spermatozoa in an ap-

parently normal manner, yet prolonged observation of

No. 575] AN APTEROUS DROSOPHILA 689

the flies has not revealed a single copulation or attempt at

copulation on the part of an apterous male.’ Similarly

the females have been shown to produce rudimentary eggs

in an apparently normal manner, and in two cases females

have produced offspring when fertilized by winged males,

thus indicating their fertility. But many other cases

have been observed in which apterous females were fertil-

ized by winged males (or at least in which copulation took

place), and yet in these observed cases the females invari-

ably died without producing offspring,® because, I believe,

of their physical weakness.

From these facts it appears practically certain that the

apparent sterility is not due to infertility of either sperm

or eggs, but results from a weakness which makes it very

difficult for the apterous flies to perform the reproductive

processes.

This explains why no crosses have been secured between

apterous and apterous, although each sex has been suc-

cessfully crossed to winged. It is also supported by the

fact that from the cross between apterous male and

winged female a large number of offspring were secured,

since the winged female could produce many eggs,—

whereas in the two crosses between apterous females and

winged males only a very few offspring were secured,

because the apterous females could only produce a few

eggs.

When the experiments were first begun it was hoped

that sooner or later one or more inherently vigorous

apterous flies would appear which might give rise to a

vigorous race. But nothing of the sort took place,

although numbers of the apterous flies were given oppor-

tunity to breed all through the course of the experiments.

Obviously, then, the physiological characteristics, as

8 Copulation in normal flies can be observed with very little difficulty. It

is evident that at least one case of copulation by an apterous male occurred,

namely in Experiment II.

9 Judging from the cases observed a large number (probably one spara

_ Oor more) of apterous females must have been fertilized by win winged

during the course of these experiments, yet only three of these gave ue.

690 THE AMERICAN NATURALIST [Vou. XLVIII

shown by vigor and viability, are directly associated with

morphological characters and are not to be separated from

them by selection. In other words, the ‘‘factor’’ respon-

sible for lack of wings is also responsible for physiological

disturbances.

The only suggestion of an inherent difference between

different races, or strains of apterous, is the slight differ-

ence in the percentage of apterous offspring in Experi-

ments I and III as compared with II and IV. In I and IIT

the ratios of winged to apterous are 3.12:1 and 3.24:1,

respectively, while in Experiments II and IV they are

4.06:1 and 4.02:1. This deviation is not great, but it is

fairly constant, and is sufficient, I believe, to indicate a

real difference. But whether it is to be explained upon

the assumption that in I and III the apterous parents were

inherently stronger than in II and IV is not so clear. It

might equally well be explained upon the basis of differ-

ences in the winged races to which apterous was crossed.

Unfortunately, an experimental analysis of the question is

prohibited by the difficulty of breeding the apterous flies,

and it must, therefore, be left open. One fact, however,

is clear, namely that there is no progressive increase in

viability of the apterous flies, for the apterous parent in

Experiment II, where the viability appears to be low, was

descended directly from that in I where it appears to be

high, and likewise the parent from the apterous side in

IV was obtained directly from III.

In conclusion it may be profitable to call to mind

briefly the bearing of certain of the above data on the

question of the nature and behavior of Mendelian

‘*factors.’’

The present case of a definite correlation between lack

of wings, reduction in size of balancers, and weak physical

constitution in the apterous race of Drosophila, shows

clearly that one factor may have far reaching effects, and

not be limited to any particular part or organ,—a fact

No. 575] AN APTEROUS DROSOPHILA 691

which has been long known, and often mentioned,!° but

by no means universally recognized.

Correlated with, or resulting from this principle is the

conception that the final result of ontogenetic develop-

ment is not due to the independent action of various

factors and their products, but is due to the combined

action, or the interaction of these products during devel-

opment. To illustrate by the wing of a fly,—it is probable

that the normal development of such an organ is not

dependent solely upon one factor, but that it is influenced

by many factors. This is strongly suggested by data

derived from the various wing mutations in Drosophila.

These have dealt with a large number of factors, each of

which is responsible for a definite wing modification. For

instance, one factor is responsible for miniature wings,

another for vestigial, another for rudimentary, another

for curved, ete.!!_ From the fact that these mutant factors

(which may be considered as modifications of factors in

the normal fly) influence the wings, it seems highly prob-

able that their normal allelomorphs also influence wing

production in the wild fly.

Finally I wish to thank Dr. T. H. Morgan for kindly

assistance and advice in connection with this work.

BIBLIOGRAPHY

Dexter, John S.

1912 a Coupling of Certain Sex-linked Characters in Drosophila.

iol. Bull., Vol. 23, p. 183.

Morgan, T. ra

1911. An Attempt to Analyze the Constitution of the Chromosomes on

the Basis of Sex-limited Inheritance in Drosophila, Jour. Exp.

ool., Vol. 2, p. 365.

1912a, Bight Factors that Show Sex-linked Inheritance in Drosophila.

ence, N. S., Vol. 35, p. 472.

10 Most recently, perhaps, by Morgan (1913a, page 9): ‘‘A change in a

factor may have far-reaching consequences. Every part of the organism

capable of reacting to the new change is affected. Though we seize upon th

most conspicuous difference between the old type and its mutant, and make

use of this alone, every student of heredity is familiar with cases where more

than the part taken as the index is affected. Weismann’s theory, on the

other hand, seems to identify each character with a special determinant ...’’

11 The same is true for various eye colors, and body colors

692 .

1912b.

1912c.

1912d.

1913a.

1913b,

Morgan, T.

1912.

Morgan, T.

1913

Sturtevant,

19134

19136.

THE AMERICAN NATURALIST [VoL. XLVIII

A Modification of the Sex Ratio, and of Other Ratios, in

Drosophila through Linkage. Zeit. f. ind. Abst. u. Vererb.,

Bd. 7, p. 323.

Heredity of Body Color in Drosophila. Jour. Exp. Zool., Vol.

» p. 27.

The Explanation of a New Sex-ratio in Drosophila and Com-

plete Linkage in the Second Chromosome of the Male. Science,

N. B., Vol. 36, p: 718

Factors and Unit Characters in Mendelian Heredity. AMER.

Nart., Vol. 47, p

Simplicity versus Adequacy in Mendelian Formule. AMER.

72.

H., and C. J. Lynch.

The Linkage of Two Factors in oe that Are Not Sex-

linked. Biot. Bull., Vol. 23, p. 1

H., and E. Catt ell.

Additional Data for the Study of FE oat Inheritance in

Drosophila. Jour. Exp. Zool., Vol.

A. H.

The Linear Arrangement of Six Sex-linked Factors in

Drosophila, as Shown by their Mode of Association. Jour.

ap. Zool., Vol. 14, p. 43.

A Third Group of Linked Genes in Drosophila ampelophila.

Science, N. S., Vol. 37, p. 990.

SHORTER ARTICLES AND DISCUSSION

FORMULA FOR THE RESULTS OF INBREEDING

IN connection with Pearl’s recent valuable analyses of the

results of inbreeding (1, 2, 3), a comparison of these results with

those from self-fertilization is of interest. In my note on the

latter (4), I gave a formula for the rate at which organisms

become homozygotie through continued self-fertilization. This

occurs more slowly in the various types of inbreeding, but Pearl

gives no general formula for it. For purposes of comparison I

have worked out from Pearl’s data the general formula for the

rate at which organisms become homozygotic through continued

brother by sister mating; as such formule appear to be of perma-

nent value, it is here given.* What the formula gives is, pre-

cisely, (1) the proportion of individuals that will be homozygotie

or any given character after any number of unbroken genera-

tions of such inbreeding, (2) the average proportion of the char-

acters of a given individual that will be homozygotie after any

number of unbroken generations of such inbreeding. The nu-

merical value so obtained may conveniently be called the co-

efficient of homozygosis.

The formula turns out to be a combination of the successive

powers of 2, with the successive terms of the Fibonacci series,

which appears in so curious a way in various natural phenomena.

In this series every term is the sum of the two preceding terms,

the series beginning: 0, 1, 1, 2, 3, 5, 8, 13, ete.

Let «=the coefficient of homozygosis.

n= the number of inbred generations (the number of

times successive brother by sister mating has

occurred).

fis fo» fa, ete., = the successive terms of the Fibonacci series

üs f0, f= 1, ate).

Then the formula for the coefficient of homozygosis is:

w Eh aha . ete,

o 2

y —

L =

(The terms in the numerator are continued until the exponent

of 2 becomes

1 In conversation, Dr. Pearl urged the publication of the present note,

otherwise I should not at this time have dealt with a matter which he has

under analysis.

693

694 THE AMERICAN NATURALIST [VoL. XLVIII

Thus, if the number of inbreedings (n) is 1.

t= z= 1/2, or 50 per cent.

If n= 4

3 2 1 0

pest ce i TAPE At Ab én 68.15 par eat

If n=9

28-4027 + 1,29 4 1.25 4 2.244 3.93 4 5.9? +. 8.914 13.29

ia 53 ).2? + 8.2" + 13.

== 457/512, or 89.26 per cent.

It n= 16

x or 97.38 per cent.

_ 63819

65536

As these examples show, the formula gives the results that

were obtained by Pearl in the detailed working out (so far as

this was carried), as given in Pearl’s table I (2, p. 62). (It will

be noted that Pearl counts as generation 1 the one before inbreed-

ing has occurred, so that his generation 10, for example, is that in

which there have been 9 inbreedings (n=9).

If one is working out the values of the coefficient x for a series

of generations, the above formula may be expressed as a simple

rule, applicable after the value for »—1 is obtained. This

e is:

The value of the coefficient of homozygosis x for any term (as

the nth) is obtained by doubling the numerator and denominator

of the fraction expressing the value for the previous term, and

adding to the numerator the corresponding (n— 1th) term of

the Fibonacci series.

Or, in view of the peculiar nature of the Fibonacci series, the

rule may be expressed as follows:

Double the numerator and denominator. and add to the nu-

merator the sum of the last two numbers so added.

Thus, since

x for 1 inbreeding = 1/2

2x1+0

?? 2? en aetna nee were nti

2” 2 =F = 3/4

2x2+1

79) 2? (ec ee ee,

x 3 eve a 5/8

» ar 2K5+1._

z 4 O Ga == 11/16, ete.

No. 575] SHORTER ARTICLES AND DISCUSSION 695

After obtaining x, or the proportion of homozygotes for any

one pair of characters, the proportion y for any number m of

pairs is obtained simply by raising x to the mth power, that is:

ya,

Thus, after two generations of brother X sister mating, the

proportion of homozygotes for three pairs of characters is

(1/2)*==1/8, or 12.5 per cent. After 8 generations of such

inbreeding the proportion homozygotie for 10 pairs of char-

acters is:

(555) = 24.05 per cent.

The corresponding value in the case of continued self-fertili-

zation is 99.61 per cent. (4, p. 491).

Whether it may be possible to obtain a similar formula for

the coefficient of homozygosis in the cases of mating of cousin X

cousin or of parent X offspring, remains to be discovered.

Pearl’s ‘‘coefficient of inbreeding’’ gives the percentage of

lacking ancestors in a given pedigree, as compared with the

number that would be present if all the parents were unrelated.

In order to compare self-fertilization with inbreeding in this

respect, Pearl’s formulæ for the coefficient of inbreeding may be

expressed in terms of the number of successive inbreedings (7) ;

for many purposes the formule appear more convenient so ex:

pressed. The following gives these formule for self-fertilization

and the three types of inbreeding, together with those, so far as

worked out, for the proportion of individuals homozygotie with

respect to a given character. In all these, n is the number of

successive self-fertilizaticus or inbreedings.

Coefficient of Inbreeding. Coefficient of Homozygosis.

2" —1

Self-fertilization 7 =

Qn-1 + f,-2"-2-+ fa 27. o -etc.

Brother x Sister aR

eire

Omala x Coada oa ?

a E E ?

Parent x Offspring ree

~ It will be observed that in self-fertilization the value of the

coefficient of inbreeding is, curiously, the same as that of the

coefficient of homozygosis, while in the other cases there is no

evident simple relation between the two. Further, the coefficient

696 ` THE AMERICAN NATURALIST (Vou. XLVII

of inbreeding in brother X sister mating is the same as for self-

fertilization, save that it lags one generation behind the latter;

thus the coefficient for the fourth generation of self-fertilization

is the same as that for the fifth of brother X sister mating. Pearl

(1, p. 592) has already pointed out that in cousin mating the

coefficient is one-half that for brother X sister, with a lag of one

generation; as compared with self-fertilization the lag is two

generations. No such simple relation is apparent between the

proportions of homozygotes resulting from the diverse methods

of breeding, though possibly such may yet be discovered.

H. S. JENNINGS

PAPERS CITED

1. Pearl, R. A contribution toward an analysis of the problem of inbreed-

ing. This JOURNAL, XLVII, October, 1913, pp. 577-614.

——. On the results of inbreeding a Mendelian population; a correction

and extension 7 previous conclusions, This JOURNAL, XLVIII, Jan-

uary, 1914, pp. 57-62

On a pace formula for the constitution of the nth generation vf

a Mendelian population in which all matings are of brother X sister.

This JOURNAL, XLVIII, August, 1914, pp. 491-494

4 Jennings, H. S. Production of pure homozygotie organisms from hetero-

zygotes by self-fertilization. -This JOURNAL, XLVI, August, 1912, pp.

487—491. .

A SHORT-CUT IN THE COMPUTATION OF CERTAIN

PROBABLE ERRORS

In his handbook of statistical methods, on p. 38, Dr. C. B.

Davenport! gives a short method for the calculation of the prob-

able errors of some of the commonest statistical constants, in a

table of logarithmic formule. It would seem that the simple

and obvious short-cut involved has not been given the attention

it deserves in connection with non-logarithmie calculation. The

logarithmic formule are as follows :?

(1) log E, = log .6745 + log ø — $ log n [since E, = 6745 =|.

log E, = log E, — 4 log 2 [since E, = .6745 T

or, B, =~E,~+ |,

1 Davenport, C. B., ‘‘Statistical Methods with Special ae to Bio-

logical Variation,’’ 2d ed., 1904, New York, John Wiley & s.

2 A indicates the weighted arithmetic mean, o the standard ‘dative. and

-Oth coefficient of variability.

(2)

No. 575] SHORTER ARTICLES AND DISCUSSION 697

(3)8 log Ee = log E, — log A [since E, = E, + A].

Now, if one is working with a calculating machine, he can

simply carry the value of E a to two or three more decimal places

than are to be retained, and then divide by the square root of

2 to get E,; similarly, the latter’ value, divided. by the mean,

gives Eo.

The writer prefers, however, to caleulate the values in the

ordinary w pe on the machine, using Miss Gibson’st table for

6745

, and then to use the short method in checking.

guns

The original computations can be indicated and performed with

great confidence and rapidity, since it is hardly possible to make

an error that will not be discovered in the checking.’ It is

obviously safer, as well as much quicker, to check in this way

than to repeat the original processes. Howarp B. Frost

CITRUS EXPERIMENT STATION,

RIVERSIDE, CAL,

GALTON AND DISCONTINUITY IN VARIATION

Ir seems not to be generally realized that Galton recognized

both continuity and discontinuity, both in variation and inherit-

ance. Of course, all biologists are familiar with ‘‘Galton’s poly-

gon,’’ in which slight oscillations of the polygon on one of its

faces, but without a change of face, are compared with ‘‘small

unstable deviations’’ (fluctuations), while a larger oscillation, in

which the polygon moves over to a new face, is compared to a

sport ... of such marked peculiarity and stability as to rank as a new

type, capable of becoming the origin of a new race with very little as-

sistance on the part of natural selection.*

Galton’s polygon illustrated for him how the following uaii

tions may co-exist:

(1) Variability within narrow limits without prejudice to the purity

of the breed. (2) Partly stable sub-types. (3) Tendency, when much

disturbed, to revert from a sub-type to an earlier form. (4) Oceasjonal—

sports which may give rise to new types.

These four types would seem to correspond rather well to what

3 Formula (3) gives, of course, the approximate or uncorrected value

Ee,

4Gibson, Winifred, ‘¢Tables for Facilitating the Computation of Prob-

able Errors,’’ Biometrika, 4: 385-393. 3 tables.

5 Unless, of course, one misreads the figures from the machine in checking.

1 Natural Inheritance, ’’ London, 1889, p. 28.

698 THE AMERICAN NATURALIST [Vou. XLVIII

are now called (1) fluctuations or ‘‘non-inherited’’ (in reality, I

think, partially inherited) continuous variations; (2) instability

resulting from a heterozygous or partially heterozygous condi-

tion; (3) reversions, now believed to result chiefly from cross-

ing; and (4) mutations.

Galton is equally explicit i in other statements on this subject.

Like Darwin, he admitted the facts both of continuity and dis-

continuity in variation; but, unlike Darwin, he also recognized

discontinuity or E A as well as continuity or blending, in

inheritance. Thus he says, in a paragraph headed ‘‘stability of

sports’’ :?

ae a does not show that those wide varieties which are called

“sports ” are unstable. On the contrary, they are often transmitted to

successive generations with curious persistence. Neither is there any

reason for expecting otherwise. While we can well understand that a

strained modification of a type would not be so stable as one that ap-

proximates more nearly to the typical center, the variety may be so wide

that it falls into different conditions of stability, and ceases to be a

strained modification of the original type.

In another paragraph,* headed ‘‘Evolution not by minute

steps only,’’ he says:

The theory of evolution might dispense with a restriction, for which

it is diffieult to see either the need or the justification, namely, that the

course of evolution always proceeds by steps that are severally minute,

and that become effective only through accumulation. That the steps

may be small and that they must be small are very different views; it is

only to the latter that I object... . An apparent ground for the com-

mon belief is founded on the fact that wherever search is made for in-

termediate forms between widely divergent varieties, whether they be of

plants or of animals, of weapons or utensils, of customs, religion of

language, or of any other product of evolution, a long and orderly series

can usually be made out, each member of which differs in an almost im-

perceptible degree from the adjacent specimens. But it does not at all

follow because these intermediate forms have been found to exist, that

they are the very stages that were passed through in the course of evo-

lution. Counter evidence exists in abundance, not-only of the appear-

ance of considerable sports, but of their remarkable stability in hered-

itary transmission,

Again, Galton not only believed in the existence of both

blended and alternative inheritance, but he recognized the im-

2 L. c., p. 30.

sL ¢., p. 32.

No. 575] SHORTER ARTICLES AND DISCUSSION 699

‘portance of the latter in connection with the survival of new

races. Thus he writes

The quadroon child of the mulatto and the white has a quarter tint;

some of the children may be altogether darker or lighter than the rest,

but they are not piebald. Skin-color is therefore a good example of

what I call blended inheritance. .

Next as regards heritages that come io aia from one progenitor to

the exclusion of the rest. Eye-color is a fairly good illustration of

There are probably no heritages that perfectly blend or that abso-

lutely exclude one another, but all heritages have a tendency in one or

the other direction, and the tendency is often a very strong one.

On the following page Galton remarks that

A peculiar interest attaches itself to mutually exclusive heritages,

owing to the aid they must afford to the establishment of incipient races.

He thus recognizes the invalidity of Darwin’s objection to

‘‘single variations’’ as a factor in evolution, namely, that they

would certainly be swamped by crossing with the general popu-

lation.

It would, therefore, appear that in his recognition of continu-

ity as well as discontinuity both in variation and heredity, Galton

was in advance of his time, and more in accord with some of the

current views. R. RUGGLES GATES

UNIVERSITY OF LONDON `

REPULSION IN MICE

IN the February number of the AmERICAN NaruRaList Dr. C.

Little criticizes the results of my mouse-breeding experiments

which I published in the Zeitschrift für Induktive Abstam-

mungs- und Vererbungs-lehre Bd. VI, Heft 3. The chief point,

‘on which he disagrees with me, is the interpretation of the results

I obtained in breeding black and albino mice together.

The fact is, that in my paper on mice, I overlooked a serious

error. In three sentences on page 126, relating to test ATIR

of albinos, the words ‘‘black’’ and ‘‘agouti’’ changed places.

printed in the paper these sentences run:

Without exception they have given black or equal numbers of black

and albino young, depending upon the purity of the black used. But

never has one of these albinos produced a single agouti young in a mat-

ing with black. Counting together the colored young of such families I

get 89 black ~~

iE api

5 Cates of kde in such crosses are of course now well- known.

700 THE AMERICAN NATURALIST ([Vou.XLVIII .

These errors were corrected in an “‘errata’’ in Band VI, heft 5,

which Dr. Little unhappily did not find. The sentences should

read:

Without exception they have given agouti, or equal numbers of agouti

and albino young, depending upon the purity of the black used. But

never has one of these albinos produced a single black young in a mating

with black. Counting together the colored young of such families I

get 89 agouti young.

Professor Punnett was so kind as to draw my attention to

these mistakes. They were corrected in the reprints sent out.

The facts were simply these: Albinos were bred of two sorts,

with and without @ (the gene which agoutis have more than

blacks). These albinos can only be distinguished by test-mating

them to blacks. The albinos with G (aG) give agouti young, if

mated to black (Ag), the ag albinos give black young from such

a-mating. In one series, some agoutis were produced, which

were heterozygous for A as well as for G@(AaGq). Ordinarily,

such agoutis, when mated inter se, produce 9 agouti (1 AAGG,

2 AAGg, 2 AaGG, 4 AaGg), 3 black (1 aaGG, 2 aaGg) and 4

albinos ( 1 aaGG, 2 aaGg, 1 aagg) in every sixteen. Mated to

albinos without @(ag) the ordinary AaGg animals give four.

kinds of young, agoutis (AaGg), blacks (Aagg) and two kinds

of albinos (aaGg) and (aagg) in equal numbers.

Now these particular AaGg animals did not produce four

kinds of gametes, as expected, namely, AG, Ag, aG and ag, but

only two kinds, Ag and aG. Thirty one agoutis were test-mated

to aagg albinos. -These test matings gave 181 young, of which 94

were black (Aagg) and 87 albino (aaG@qg). No agoutis were

produced.

As a further proof, the result of breeding these agoutis inter se,

can be adduced. These matings gave 73 agouti (AaGg), 37 black

(AAgg) and 32 albinos (aaGG@). Of these 32 albinos, thirteen

were tested by mating them to blacks. If one of them should

have lacked G, it would have given black young. But no black

young were produced. Some young were albino (when the black

parent was heterozygous for A), but all the colored young were

agouti (89 in all). —

This, I hope, will make it perfectly clear, that in this series

we have been dealing with a case of repulsion between the genes

A and G. A. L. HAGEDOORN

_ Bussum, HOLLAND

No. 575] SHORTER ARTICLES AND DISCUSSION 701

THE OSTEOLOGY OF A DOUBLE-HEADED CALF

THROUGH the kindness of Mr. Charles O. Reed, taxidermist, of

Fairmont, W. Va., the writer received the skulls and anterior

cervical vertebrae of a double-headed calf which seemed of suffi-

cient interest to warrant a brief description.

According to Mr. Reed the calf’s mother was a four-year-old,

thoroughbred Herford, living at Grafton, W. Va., owner not

‘mentioned.

At her first labor this cow gave birth to twins, supposedly

normal, though it was not so stated. The second calf was

“‘slightly deformed,’’ but in what way Reed did not know. The

third labor produced the double-headed calf in question, which

was of unusual size, and was killed in parturition. According

to Reed ‘‘This calf would have lived if it could have been brought

through O. K.’’ He dissected it and found the ‘‘alimentary

canal, blood vessels and trachea normal.’’

The bones in the occipital region are slightly broken, probably

done in disarticulating the skulls from the neck; and in the left

skull the left premaxilla was lost and was replaced by a roughly

carved piece of wood for the sake of symmetry.

In macerating the skulls, for the purpose of removing all the

flesh, many of the loose sutures separated, and in gluing the bones

together again it was not always possible to completely close the

sutures.

As may be seen in the figures there is a considerable though not

very great difference in the size of the skulls, the right being the

larger. They were detached from the cervical vertebre when

received, but the photographs show their approximate position in

relation to the neck and to each other.

Each skull is twisted and bent away from the other, the bend

being most marked just cephalad to the orbits. The left skull is

the more distorted. .

The articulation of the skulls with the fused atlas was so crude —

that Reed, who had seen the skulls before disarticulation, had to

be appealed to to decide which skull was right and which was left.

Fig. 1 is a photograph of the dorsal aspect of the skulls and

the first three cervical vertebre. The distortion of the two skulls

is of about the same character but is, as noted above, more marked

in the left skull.

The parietal (p) is normal. The posterior regions of the fron-

tals (f) are normal, but their anterior ends are bent laterally,

702 THE AMERICAN NATURALIST [Vou. XLVIII

which causes a slight curvature in the sagittal suture. It is in the

region of the lachrymals (1) that the distortion is most marked,

so that the lachrymal on the convex side of the bend, especially

Fig. 1. DORSAL VIEWS OF THE Two SKULLS AND OF THE First THREE CERVICAL

VERTEBR®. MANDIBLES IN POSITION.

a, atlas; aw, axis; e, extra bone between maxilla and premaxilla; f, frontai ;

l, lachrymal; m, malar; mz, maxilla; n, sal; o, occipital; p, parietal; pm

premaxilla.

in the left skull, is much longer than that on the opposite side;

the same is true of the malars (m), of the maxillaries (ma), and,

to a less degree, of the premaxille (pm). The nasals (n) are also

unsymmetrical, but do not differ much in size; they are simply,

as a pair, pushed to the side.

Fig. 2. The ventral aspect of the skull shows even greater ab-

normalities than the dorsal. The occipital (0), as noted above,

was somewhat injured by the person who disarticulated the skulls

from the neck, but it is quite unsymmetrical, especially in its

exoccipital region. In the left skull (right in this figure) all the

other bones seen in this aspect are bent, but in the other skull

most of the bones are comparatively straight.

In the right skull a suture in front of the teeth separates off an

extra bone (e) on each side, between the maxilla and the pre-

maxilla, that of the right side being much the larger. In the left

skull these extra bones are not present though a partial suture,

No. 575] SHORTER ARTICLES AND DISCUSSION 703

extending about half way through the left maxilla, is visible in

this view of t

~—

1e skull.

FIG. 2. VENTRAL VIEWS OF THE SKULLS AND THE neat Two CERVICAL VER

TEBRÆ. iaoa REMO

Fig. 3 shows the curious distortion of the mandibles, which seem

to be bent in more or less the same direction. In the right man-

dible the left half has four incisor teeth, the right half has three.

In the left mandible the right half has four teeth, the left half has

three, though one tooth is missing from each half.

As noted above, the skulls, when received, were disconnected

from the vertebre: but the latter, three in number, were strung

together on a small piece of rope and presumably were the first

three cervicals; they are shown in a dorsal view in Fig. 1, ven-

tral view in Fig. 2. and anterior view in Fig. 3. The first of

these is presumably a compound atlas (a) since it articulates

with each of the skulls. though in a very crude way. It consists

of eight loosely united elements which became completely sepa-

glued together again. In the

rated in cleaning and had to be

bone from which

dorsal view, Fig. 1, is seen a small, irregular

radiate three somewhat symmetrical bones, the largest lying it

the median plane between the bases of the skulls. This larger

704 THE AMERICAN NATURALIST (VoL. XLVIII

bone is pierced by two large foramina; each of the other two

bones shows in this view a foramen which branches and opens

both on the antero-median and the postero-lateral surfaces. The

Fig. 3. DORSAL VIEWS OF THE MANDIBLES AND ANTERIOR VIEWS OF THE FIRST

THREE CERVICAL VERTEBRÆ.

ventral view, Fig. 2, shows a very irregular group of bones, the

smallest of which is for articulation with the following vertebra.

The second vertebra (ax), supposedly the axis, exhibits no

indication of an odontoid process and articulates in a very crude

way with the preceding bone. Its dorsal spine is rather elon-

gated in an antero-posterior direction, but otherwise it bears no

closer resemblance to an axis than to any other cervical vertebra.

Its centrum was so loosely fused with the arch on either side

it became detached in cleaning and had to be glued in place.

The third vertebra exhibits no peculiarities that warrant de-

scription.

A. M. REESE.

WEST VIRGINIA UNIVERSITY,

MORGANTOWN

THE

AMERICAN

NATURALIST

A MONTHLY JOURNAL

Devoted to the Advancement of the Biological Sciences with

Special Reference to the Factors of Evolution

CONTENTS

Page

. The Failure of Ether to Produce Mutations in Drosophila.

Professor T. H. MORGAN 705

bot

The Analysis of a Case of Continuous Variation in — ad a Spats

of its Linkage Relations. Professor JOHN S. DEXTER 71

Ill. Shorter Articles and Correspondence : On the Progressive Increase of Homo-

zygous Brother-Sister Matings. H. prim- - - ~ = oo D

soal TOS

. NotesandLiterature: Mendelian Fluctuations: G.U.Y.

Index to Volume XLIII -

THE SCIENCE PRESS

LANCASTER, PA. GARRISON, N. T.

NEW YORK: SUB-STATION 84

The American

Naturalist

intended for pae t and books, etc., intended for review should be

MSS

sent to the Sec! of THE AMERICAN NATURAL

Short articles containing summaries of res

problems rod organic evolution are especially rlon.

in ops: oe

RALIST, Garrison- -on- agrbees = York.

earch work bea on the

and will be Sey afne an

ndrea reprints of si f are supplied to authors free of charge.

Further ı Pariak will be supplied at c

Subscript

ons and p saaa aeina ae be sent to the Ponas. The

subscription price is four dollars a year. oreign postage fty cents and

Canadian postage twenty-five ak additional. The ppi for single poiri is

forty cents. The advertising rates are Four Dollars for a

THE SCIENCE PRESS

NEW YORK: Sub-Station 84

Entered as second-class matter, April 2, 1908, at the Post Office at Lancaster, Pa., under the Act ot

Congress of March 3, 1879.

Lancaster, Pa.

Garrison, N. Y.

FOR SALE

ARCTIC, ICELAND and GREENLAND

BIRDS’ SKINS,

Low Prices

rticulars of

G. DINESEN. ra Collecto

Husavik, North Iceland, Via Leidie, enatand

Well Skr

JAPAN NATURAL HISTORY SPECIMENS

Perfect Condition and Lowest Prices.

Specialty: Bird Skins, Oology, Entomology, Marine

Animals and others. Catalogue free. Correspond-

ence solicited.

T. FUKAI, Naturalist,

Konosu, Saitama, Japan

For Sale Entire

An important collection of Indian Birds’

Eggs, containing approximately 3,300

specimens of about 620 species, and in-

cluding among other varieties the Green-

ish Willow Warbler (Acanthopneuste

viridanus), Red-browed Finch (Calla-

canthis burtoni), Tibetan Sand-Grouse

(Syrrhaptes tibetanus), Spotted Sand-

Grouse (Pteroclurus senegallus), Ibis-

bill (Ibidorhynchus struthersi), White-

bellied Minivet (Pericrocotus erythro-

pygius), Bearded Vulture (Gypaëtus

barbatus), numerous Cuckoos with their

hosts, ete., ete.

Particulars and full list may be had

from

W. F. H. ROSENBERG

57 Haverstock Hill London, N. W.

| Marine Biological Laboratory

| Woods Hole, Mass.

INVESTIGATION Busilities for research in Zoology,

ryology, Physiology and Bot-

Entire Year a ty-six private labora-

rar, $100 each for not over three

y tables are avail-

able for bataar in — who

desire to work under the direction

of aaa the. of the sale The fee

for such a table is $50.0

Courses of labora ok n

INSTRUCTION with E tenbared are fered i in Inverte-

July—August

plants n i

¢ ia embe ma mic stages:

all s

raae of Algae, Dr oor badi

SUPPLY

DEPARTMENT

Open the Entire Year

z for. the m

Sasori oF furnished in season

as ordered. Price li of we

logical an $ oieniees TEk

po on Ppr, estion À and all

information ma

address

GEO. M. GRAY, Curator, Woods Hole, Mass

The annual announcement will be sent on application nt

The Director, Marine Biological Laboratory, Woods Hole

Mass-

THE

AMERICAN NATURALIST

Vout. XLVIII December 1914 No. 576

THE FAILURE OF ETHER TO PRODUCE MUTA-

TIONS IN DROSOPHILA

PROFESSOR T. H. MORGAN

COLUMBIA UNIVERSITY

THE many mutants of Drosophila ampelophila that have

appeared ‘‘under domestication’’ have raised the question

as to the cause or causes that have brought about the re-

sult. Since every fly that has passed through our hands

has been etherized once in its life, usually before it begins

to lay its eggs if a female or before mating if a male, it

might appear that this recurring condition was respon-

sible for the mutations. At any rate it seemed worth

while to put this view to a test, if for no other reason

than to remove from one’s mind the suspicion that ether

UGIL.

Preliminary trials showed that two drops of ether (on a

piece of cotton). in a quart milk bottle, tightly stoppered

with a cotton plug, would not noticeably affect the flies in

half an hour, three drops made them slightly ‘‘stupid,’’

four drops more so, and five drops quieted them. It was

found that they would for the most part recover even

after 6, 7 and 8 drops of ether. If etherized twice daily

the flies were so far weakened that they generally died

without laying any eggs. Therefore in the later experi-

ments the flies were etherized only once a day or once in

two days.

The larve (beginning two days after the eggs were laid,

705

706 THE AMERICAN NATURALIST [Vou. XLVIII

at which time the eggs have hatched) can stand more

ether. Twice daily throughout their larval and pupal

lives (approximately 11 days) 6, or 7, or 8, or 9 or even 10

drops of ether were added to the quart bottles. The ether

excited the larve at first, then quieted them; later they re-

covered. In such tests the larve were kept almost con-

tinuously in an atmosphere of ether from birth to emer-

gence of the fly and in a few cases the etherization was con-

tinued with the flies also. By covering a wide range of

stages and conditions I hoped to find the critical point, if

any such existed, when ether would act. Since, as the

sequel will show, no specific results were obtained it seems

unnecessary to give the details of all these trials.

Double, and in one ease triple recessive, stocks were

used for the work, because experience had already shown

that even with great care contamination may occur. One

or two flies that came from escaped mutants would ruin

the value of the data, but the operator can protect himself by

using stocks that have already two or more recessive char-

acters. If such flies mutate in one of the characters in-

volved the presence of the other one will make it certain

that the mutant belonged to this culture, and had not come

in from outside; if a change appeared in some other part,

the double recessive character would still identify the

stock. Two of the stocks used had sex-linked characters,

i. e., eosin miniature and cherry club vermilion. If a

mutation should appear that involved these characters it

would become evident at once in the male offspring; for,

the male gets his single sex chromosome from his mother

and exhibits her sex-linked factors. Of course this

would be equally true for any other sex-linked char-

acter that appeared, but in practise it is impossible to

thoroughly examine each fly in every possible part, so that

I had to confine my attention to certain organs, and in

these cases I concentrated on the mutant characters. Con-

spicuous mutations in other parts would, I think, have

been picked up, but minor ones would probably have been

missed. On the other hand, if changes taking place in

No. 576] MUTATIONS IN DROSOPHILA Pita!

the chromosomal material are the basis for mutation it

would seem perhaps a priori unlikely that the same

changes should occur at the same time in both members of

a pair, and if not the effect would not appear in the next

generation, and not until two flies of the later progeny

each carrying one mutant factor met. Whatever weight

may be attached to this argument—we know really nothing

as to the origin of mutations—it seemed necessary to

carry some stocks to another generation; and this was

done.

The following are the totals of offspring produced by

flies from larve that had been etherized twice daily from

the time of hatching to the winged state:

Bhk væk r oesG voc es eae E GEE T io

Puk Dak oes. E RE e ES T AE A a bees 1,390

Ronm DOMINARE soo a oot aa aes cK ee 871

Cherry olub Vermilion 05.0000. is peace es sis orria 364

Pink ebòùy SORIG s 6205 5a Wis ee eee oes 1,311

ital o ie ee kes 4,802

In the next case fewer drops of ether were used—four

or five throughout larval and pupal life. The parent flies

were changed to new bottles quite often to prevent crowd-

ing and abundant food was supplied:

Black vestigial .........0cscececeer cesses ceeeeeees 2,122

Pink HACK ay oss es ae ws EVER Bens Fomine eos Vee oe be 6,762

Hoin miniature... vs cs 6 ce erases ees hese toe pes oes 2,603

Pink ebony sepia ....... esses cee reereetenereecess 953

12,440

T reo Ss es ewe ive we be OE CS ee oe

The following data are from the offspring of the flies

that had 8 and 10 drops of ether twice daily:

Pink blak oc Fo ee a wk ee areen 3,440

i i 2,775

Fosin miniature 2. ..cccc cece ccs ret ss eects es eera 2,

a rs ae cet hee ae re 6,215

The next data are the records of the offspring of

708 THE AMERICAN NATURALIST [Vou XLVIII

adult flies that had been etherized several times (usually

twice a day) just after they had hatched:

Dinek Voutimia) (6! CGN) ies is esi ae Ves e es 3g koe 870

Black vestigial (V times} Soares ie ses wie sates 143

Black vestigial (8 times) ................ sey he aa Ea 694

Toró Tame (S See ee a ie ce kee ee 81

Eons mimiature: (3 UMA) s. asair os 5 es a 206

Mosin miniature (O°: Wee os re e eh 428

Cherry club vermillion (8 times) > roii. cereri dirr 713

Cherry club vermillion (8 times) .............-2500%5 476

FONE ae yao OW a ikia Secu ee roles ro 3,611

Finally some of the flies that had appeared, in the ex-

periments in which 8 and 10 drops of ether had been used

throughout the larval and pupal life, were bred and gave

in the next generation the following records:

PUR NOR oo E eens a Sha ae ase ee hone sui 2,186

MOOT) DIOR ONG ks bk sy wicca bh oe ON a a

Carry CBD VADANO icy ee cake hron KASA eh ees 709

PER MOUS BOER os i ins ks as ECA eS RS ee 539

ae EN a aera Oe pe car grater Grain ie ata gine 4,100

In a grand total of 31,168 flies subjected to ether, there

was not a single mutation observed. It seems highly

probable therefore that ether has no specific effect in pro-

ducing mutations in Drosophila ampelophila. It might,

of course, still be said that mutations are so rare, that,

although caused by ether, they still are not frequently

enough produced to make 31,000 flies a sufficient guaran-

tee. Granting this, it still remains that since no mutants

appeared under this excessive treatment, ether does not

play the rôle of a specific agent causing the. mutations of

Drosophila, and one is inclined to look elsewhere for a

solution of the problem.

One of the first mutants that I observed in ampelophila

appeared in the offspring of flies that had been treated

with radium and although there was no proof that the

radium had had a specific effect I felt obliged to state the

actual case, refraining carefully from any statement of

No. 576] MUTATIONS IN DROSOPHILA 709

causal connection.! Nevertheless, I have been quoted as

having produced the first mutants by the use of radium.

I may add that repetition of the experiment on a large

scale both with the emanations of an X-ray machine and

from radium salts has failed to produce any mutations,

although the flies were made sterile for a time. Loeb and

Bancroft also tried the effect of radium.? They found a

black mutant type after treatment with radium but since

the same type appeared in the control they do not believe

that its appearance had any connection with the radium.

They also state that after treatment a white-eyed female

appeared in the first generation, and suggest that a white

eyed male may have existed in a previous generation that

escaped notice, but if it had been found in a previous gen-

eration, the mutation or the contamination must have

been earlier than the one that produced the white-eyed

female; for, a white-eyed male takes two generations to

reappear again. Pink-eyed flies were also found both in

the control and in the treated flies. In regard to another

mutant type, they state:

We succeeded in producing short winged specimens in two different

cultures by treating them with radium, while thus far we have not

observed this mutation in cultures not treated with radium.

But although ‘‘two hundred different cultures’’ were

subsequently treated with radium and no short-winged

(miniature) flies appeared, I get the impression that

Bancroft and Loeb must have had stock that was already

contaminated by some recessive mutant factors. All of

these mutants had been obtained and described by us, and

the stock used by Bancroft and Loeb was obtained in part

at least from my friend Dr. Frank E. Lutz, who had at

that time in his possession, as a letter I have from him

states, certainly two of these mutants, black and minia-

ture, that he had received from me. It seems to me not

improbable that the collector, who got the stock from Dr.

1 Science, XXXIII, 1911.

2 Loe. cit.

710 THE AMERICAN NATURALIST [Vow XLVII

Lutz for Professor Loeb, included by mistake some flies

heterozygous for these two characters; for in our very

extensive experience with wild stock from Cold Spring

Harbor (the origin of most of Dr. Lutz’s stock) and else-

where these mutants have never arisen again. .

At various times experiments have been made in this

laboratory involving wide ranges of temperature,® salts,

sugars, acids, alkalis without any resulting mutation.

In fact, our experience with Drosophila has given us the

impression that mutations are rare events, although the

actual number of our mutants is now quite large.

Guyénot* also has treated ampelophila to high tempera-

tures, to radium and to X-rays without result. When the

adult flies were treated with ultra-violet light, however, a

definite type of ‘‘black’’ fly was obtained. The first eggs

that such females lay are normal and give rise to normal

flies. The eggs laid later fail to hatch, although they ap-

pear to begin their development. On the third day

amongst the abnormal eggs some were found that gave

rise to flies that were apparently normal. It happened

that they were not examined again until after the flies of

the next generation had appeared (many of them had

died). Both among the living and the dead flies there

was a considerable percentage of black flies. The black

females laid eggs which did not develop, even although

normal males were added. Why the black males were not

also tested by outcrossing is not apparent. The descrip-

tion of the black flies given by Guyénot tallies in some

points with our stock of ebony in which the females were

at first usually infertile but the males fertile. At first,

indeed, we kept the stock by breeding the ebony males to

the heterozygous females. These are intermediate in color.

In fact, Guyénot seems to have had heterozygous flies but

did not, according to his account, obtain any black flies

from them. However, if the ultra-violet light is a specific

agent for these mutations the experiment can easily be

repeated.

3 Science, XXII, 1910.

4 Bull. Scientifique, XLVIII, 1914.

No. 576] MUTATIONS IN DROSOPHILA TII

It should be added that only one of Guyénot’s two

lines gave dark flies after treatment with ultra-violet light.

This might seem to indicate that the first result was acci-

dental, or due to the presence of a recessive mutation in

the stock prior to treatment were it not that a careful

control is recorded. Guyénot himself speaks with much

caution concerning the interpretation of his results. De-

cision as to their value may be reserved until repetition of

the experiment gives confirmation. Our own experience

with Drosophila shows that mutations appear under con-

ditions where all the other flies in the same culture are

normal and we have become unduly sceptical perhaps

towards evidence which refers a particular mutant to some

unusual treatment to which the flies have been subjected.

Until we can get definite information as to how mutants

arise, whether through external influences, through acci-

dents of mitosis, through hybridizing, or through changes

in the chromosomes with its consequent dislocations of

the machinery of crossing over, or in some other way, it

seems futile to discuss the question.

THE ANALYSIS OF A CASE OF CONTINUOUS

VARIATION IN DROSOPHILA BY A STUDY

F ITS LINKAGE RELATIONS

PROFESSOR JOHN S. DEXTER,

OLIVET COLLEGE

I. Introduction.

II. The Germinal Constitution of Beaded Flies.

A. Crosses between Beaded and Normal Wild Flies.

1. Behavior in First Generation.

2. Behavior in Second Generation.

3. Behavior in Third and Fourth Generations.

B. paa between Beaded Flies and Other Mutants.

e F, Generation.

$ Linkage Relations.

(a) Sex Linkage.

_ (b) Linkage to Sex-linked Genes

(d) Linkage to Third Chromosome Genes.

III. The Effect of Environmental Conditions on the Development of Beaded

Wings.

A. General Statement.

B. The Effects of Relative Moisture

C. The Effects of Covering with Paraffine the Mouth of the Bottle

in which the Flies are Developing.

D. The Effects of Acidity and Alkalinity.

E. The Effect of cape ger aie ald

F. The Effect of Dar

IV. The Effect of Selection on me Production of Different Types of Bead-

edness,

V. Mutation in Beaded Stock.

A, General Statement.

B. Perfect Notched Wings.

C. Spread Wings

D. Stumpy Wings.

VI. ee

I. INTRODUCTION

Hise had the principles of Mendelism been worked

out in one species of plant than apparent exceptions to

these principles were discovered. Mendel’s own case of

the breeding true of species hybrids in Hieracium was the

first of these, and since 1900 others have been reported.

te 712

No. 576] VARIATION IN DROSOPHILA 713

Further analysis has shown that many of these early

cases are readily interpreted on Mendelian principles,

while for other exceptions, like that of Hieracium, for

instance, the true explanation has been found without in

any way coming into conflict with Mendelism.

The masking of a Mendelian ratio may be effected in

many ways, and some of the most important of the recent

work in genetics has dealt with this problem. Among

the conditions so far brought to light may be mentioned

the following:

(a) Multiple Factors.—Recent papers by MacDowell

(1914) and Shull (1914) have discussed at length the

literature and history of this subject. In brief, the work

that has been done shows that in both animals and plants

the production of certain characters is brought about

through the action of two or more independently Men-

delizing pairs of genes that have similar effects on the

developing organism. If the effect of these genes is

cumulative, so that the character is more or less produced

according to the number of dominant genes present, the `

type of inheritance known as blended inheritance is

produced. If the effect is not cumulative, the recessive

character does not appear with the frequency of 1:3, but

with the frequency of 1:15, 1:63, etc., according to the

number of pairs of genes concerned.

(b) The Effect of the Environment.—A typical case

of this sort is reported by Baur (1912). In crossing a

dark red to a red strain of Antirrhinum, a complete series

between the red and the dark red appeared in the F, gen-

eration; the effect of light on the plants was such that

plants that had developed in a bright light had a darker

color than those that had developed in a less intense light.

The analysis of the F, generation, however, proved con-

clusively that one fourth of the F, plants had been homo-

zygous dark reds, one fourth had been homozygous red,

and two fourths had been heterozygotes. Morgan (1912a)

has described a case in Drosophila in which moisture

conditions in the bottle in which the flies are developing

determine to a certain extent whether or not certain

714 THE AMERICAN NATURALIST [Vou. XLVII

characters shall appear; and Hoge (1914) has shown

that certain temperatures are necessary for the develop-

ment of reduplicated legs in Drosophila. Other examples

may be found in the literature of genetics.

eral instances in recent years of animals and plants which

are unable to live if homozygous for certain genes. The

No. 576] VARIATION IN DROSOPHILA 715

case of yellow mice, Baur’s Aurea-strain of Antirrhinum

(Baur, 1912) and the modified sex-ratios in Drosophila

reported by Morgan (1912d) are examples of this phe-

nomenon.

The object of the present paper is to describe a case of

inheritance in Drosophila that for some years seemed to

716 THE AMERICAN NATURALIST [Vou. XLVIII

defy Mendelian analysis. Though all the details of the

case have not been worked out, enough has been done to

show that it is brought about by factors which segregate

in the ordinary Mendelian fashion, and that the diffi-

culties which it still presents are not opposed to that

hypothesis.

The case under consideration is that of Beaded wings,

which, according to Morgan (1911a), first appeared in

May, 1910, among flies that had been exposed during part

of their early life to radium rays. :

The appearance of these wings can best be understood

from the figures (Figs. 1-12), which represent a few of

the forms that may appear in a stock culture. All grada-

tions may be found between wings perfectly normal and

mere strips, such as shown by Figure 11.

In the early days of its history, according to Morgan,

the Beaded-winged flies did not breed true, but for many

generations produced many normal-winged offspring.

At the time when I took up the experiment, however, the

stock bred almost 100 per cent. pure; that is, almost every

fly hatched had wings more or less Beaded. I have at

present a strain which breeds true, throwing only Beaded-

winged offspring, and most of the offspring have the

Beading in an extreme form. Most of my work has been

done with this stock.

II. THE GERMINAL CONSTITUTION OF BEADED FLIES

A. CROSSES BETWEEN BEADED AND WILD FLIES

1. Behavior in First Generation

When a Beaded fly is mated to a normal fly of a normal

Wild stock, a considerable number of flies with Beaded

wings usually appears in the first generation (F',). The

percentage is not constant, but varies between zero and

about fifty per cent. (See Table I.) From Chart 1, it

appears possible that the average percentage of Beaded-

= winged offspring per pair is near 10-15 per cent. or else

near 30-35 per cent. of the total offspring. The exact

No. 576] VARIATION IN DROSOPHILA Til

average is 25.5 per cent. That there is a bimodal curve

produced may perhaps not be significant, as will appear

TABLE I

CROSSES OF BEADED TO WILD FLIES, SHOWING PERCENTAGES OF BEADED-

WINGED OFFSPRING

TEA EBT 2 ee rea > n a

Per Cent. of Flies with Beaded Wings

| | Le olaia i a] o a sen

$| Si 212] 8] 8] 4/8) sie) 8

(3 tatala sla sida g

Number of broods giving this per-| | pe |

a E E P EEAS | H 5 10| 6 7 7 14 3a 2 2) 2

Average size of brood......... +--+ ++, 203 172 148 120 180 113 101 95 130 50 73

from the following facts, although later evidence will

show that it very possibly is significant.

The per cent. of Beaded-winged offspring given by one

pair (Beaded X Wild) may vary at different times and

ee SER SS eS Aa

Se a &£ SE 3s

3 4 aa ae 2

wu 2 asrazse ts R

CHART I

Numbers of broods giving certain percentages of Beaded-winged off-

spring in F, generation of Beaded X Wild. (See Table I).

under different conditions. For instance, if a pair are

put into a bottle with food and are left there for ten days,

and are then put into another bottle with fresh food and

left another ten days, the percentage of Beaded-winged

offspring will be different in the two broods. Table II

gives the records of such tests. The first two were made

718 THE AMERICAN NATURALIST [Vou XLVIII

with single pairs. In the third case, a Beaded male was

given four virgin females, so that although all the off-

TABLE II

DIFFERENT PERCENTAGES OF BEADED-WINGED OFFSPRING BY THE SAME

PARENTS DURING Two SEPARATE TEN-DAY PERIODS

IN DIFFERENT BOTTLES

First Ten Days | Second Ten Days

| No. of Off- Per Cent. No. of Off- Per Cent.

| spring Beaded | spring Beaded

First wate 65 00S C196 te a

Second pairs. 2.. ks Ss 117 rå | 146 : 22

One father X4 mothers. . 389 20 301 28

spring have the same father, they come from four

mothers. Inspection of this table shows that it is quite

impossible to assign the parents of any one brood to any

definite class based on the percentage of Beaded-winged

offspring that they give.

Table II shows also that the parents gave a larger per-

centage of Beaded-winged offspring during the second

ten days than during the first period. That this is a

coincidence appears from Table III. Here it is shown

from the records of fifty broods chosen at random, that

TABLE III

PERCENTAGES OF BEADED-WINGED FLIES IN THE First CounT or A BROOD

COMPARED WITH THOSE OF THE LAST COUNT (INTERVAL OF

FROM EIGHT TO TEN Days). BASED ON COUNTS FROM

Firty Broops, CHOSEN AT RANDOM

iret | rase | Fit | Læt | rie | raot f Ph | bai | rat | bai

Count | Count | Count | Count Count Cvunt Count Count Count | Count

36 zi 10 0 10 33 10 13 3 7

-+ 0 42 24 1 0 25 24 51 30

10 0 71 25 8 1 24 0 23 11

20 0 0 1 19 7 11 3 15 12

1 0 5 5 3 0 32 0 43 8

48 20 64 15 36 10 40 0 42 0

45 4 37 3 25 3 46 9 32 6

37 40 17 6 18 0 60 0 50 40

33 15 10 0 47 10 16 0 29 4

21 0 10 0 28 20 52 0 56 29

Larger percentage of Beaded-winged offspring the first count, 44 broods.

Larger percentage of Beaded-winged offspring the last count, 5 broods.

ood.

: Ee percentage of Beaded-winged offspring both counts, 1 br

No. 576] VARIATION IN DROSOPHILA 719

the counts made in the first few days after the flies of

any brood begin to hatch show almost invariably a very

much larger percentage of Beaded-winged offspring than

do the last counts made. This fact will be considered at

some length in the section on environmental effects.

Enough has been said, at least, to show that, whether

the results here described are genetic or environmental

effects, the F, generation is remarkably inconstant with

reference to the percentage of Beaded-winged offspring

that appear. It is evident that this percentage can be

readily altered by (1) changing the length of the period

CHART 2.

Bd of X Wild?

daughters, 33% Bd sons, ji bots RAR

Bd o&' X Wild 9

daughters, 16% Bd sons, 3% Bd

Bd o& X Wild 2

daughters, 23% Bd sons, 9% Bd

during which the brood is allowed to run; (2) by chang-

ing the parents from one bottle to another. Extensive

studies of environmental effects have shown other ways

in which the percentages can be altered, but of this we

will treat later.

2. Behavior in the Second Generation

The question at once arises whether the Beaded and

normal F, flies are alike genetically. To the solution of

this problem two different breeding tests were applied:

viz., matings of F, normal by normal, normal by Beaded,

and Beaded by Beaded; and back crosses of both normal

and Beaded to Wild stock. The results of these tests are

given in Tables IV and V. These tables show that when —

720 THE AMERICAN NATURALIST [Vou. XLVIII

Beaded-winged flies of the F, generation are used as par-

ents, more Beaded-winged young are produced than when

normal-winged F, flies are used. This holds true for each

TABLE IV

MATINGS BETWEEN F, FLIES OF THE CROSS BEADED BY WILD, SHOWING

PERCENTAGES OF BEADED OFFSPRING IN INDIVIDUAL BROODS

Per Cent. of Flies Beaded

SPT alalalala ajaja

ETE TIT ARTIR TTS

lel" leleialals|siei/s|sieis|s

Normal XNormal........ }8|2)2]...)... aR es ep iy a Be ty ese Heres i Ut

Orme X Denes. rn F T heh Sid S E e a

Beaded XBeaded........ Rhee a as ee ee oe i 2 ila

TABLE V

BACK-CROSSES TO WILD OF F, FLIES OF THE Cross BEADED X WILD, SHOW-

ING PERCENTAGES OF BEADED OFFSPRING IN INDIVIDUAL BROODS

Percentage of Offspring Beaded

alelaleisieal3iel3|

erais g s shel Ss]

3I r ii

Normal XWild.... o.oo... ZETTE T EA ee le

Headed x Wid 03. a a Ba E a a ee ge Q a

of the five crosses shown in the two tables. Normal-

winged F, flies do, however, have some Beaded-winged off-

spring, both whi mated among themselves, and also,

though more rarely, when back crossed to Wild.

These F, and back-cross results give little satisfaction

at first sight to the student of Mendelism. If we suppose

that there is one gene on which the Beaded condition

depends, and that it is partially dominant, then Beaded

TABLE VI

BEADED AND NORMAL OFFSPRING BY SEXES WHEN ONE PARENT IS BEADED

ND THE OTHER WILD

29 9:9 Pg TF 9 9 Bd.

1,246 s 4,488 948 — 4,481 21.7 | 175

894

| Beaded | Normal Beaded Normal Per Cent. ey se

Mother Beaded. 2,959 1,139 2,684 23.2 29.8

No. 576] VARIATION IN DROSOPHILA 721

and normal F, flies should give the same results when

used as parents. Or if we were dealing here with a case

like the ‘‘yellow mouse’’ case, in which homozygous

yellows do not exist: that is, if homozygous ‘‘Beadeds’’

do not exist, then one quarter of the flies produced by two

Beaded parents from the stock should be normal. But as

was said before, the stock breeds true, every fly produced

having Beaded wings.

It may be noted that a pair of F, normal flies usually

produce less than 10 per cent. of Beaded offspring. If

these normal flies carried a recessive gene for Beaded-

ness, they should produce twenty-five per cent. Beaded

offspring. The.Beaded F, offspring, on the other hand,

though they produced in all cases more than twenty-five

per cent., did not produce 75 per cent. Beaded offspring,

as they should have done if a single. dominant gene for

Beaded wings were heterozygous in them.

3. Behavior in Third and Fourth Generations

Beaded offspring, that appeared in the F, generation

of the cross Beaded X Wild, were back crossed to Wild.

The process was again repeated with the Beaded off-

spring that appeared, till four generations had been pro-

duced. The results of this test are given in Tables VIT

and VIII and in Chart 4.

A striking result is that an F, Beaded fly or even a fly

of later generations heterozygous for Beaded wings some-

TABLE VII

REPEATED BACK-CROSSES OF BEADED-WINGED FLIES FROM THE Cross BEADED

By WILD TO WILD STOCK TO SHOW PERCENTAGES OF BEADED-

WINGED OFFSPRING. See Chart IV

ao o i ao o a èo

E 1 SE) e R a

SE pa | SE |ua | S8 |å. |58 gals E| na. | se | va

es S S 5 S oa

eration 1......... 86 25.6] 460 28.9] 690) 23.2] 48| 4.2] 82| 15.9]1,266

Generation 2. 226 25.7|1,711| 19.3] 646| 15.9]137 1.5|314| 7.6/3,034) 17

Generation 3. 515 20.8|2,512) 24.6|2,241) 19. 1) 1.8319 16.3]6,038) 21.9

Generation 4......... 135! 8.9] 196| 24.0}..... Te 4.0|132| 25.0| 760| 13.7

722 THE AMERICAN NATURALIST [Vou. XLVIII

TABLE VIII

NoRMAL FEMALES FROM FAMILY 2, GENERATION 2, BACK-CROSSED TO WILD

MALES, SHOWING PERCENTAGES OF BEADED-WINGED OFFSPRING

¢ Ba.

| Beaded | Normal

2 | 1,040 | .02

Sas Doo:

times has as large a percentage of Beaded-winged off-

spring when mated to Wild, as does a fly direct from pure

Beaded stock when mated to Wild, though a comparison

of Tables I and V shows that this is not the usual occur-

rence. This suggests at once the action of a lethal gene

(Morgan, 1912b). Morgan has shown that in a certain

stock of Drosophila there are twice as many females as

males in the offspring of one half the females. No matter -

to what male such a female be mated, her daughters are

twice as numerous as her sons, and one half of her

daughters also repeat this phenomenon, and one half of

the daughters of these again. This fact finds its explana-

tion in the assumption that there is in one of the sex-

chromosomes of such females a gene which prevents the

development of any male which gets it.

Now if such a gene had the power of expressing itself

as a dominant in those flies that carried it in the hetero-

zygous condition, if, for example, it caused the wings to be

Beaded, it would be possible to select such flies at sight,

and these flies could then be depended upon to repeat the

phenomenon. (Morgan accomplishes the same end by

mating such flies to mutants carrying a gene with which

the lethal gene shows close linkage, such as that for white-

eyes. He then finds that the red-eyed females carry the

lethal gene, unless, as rarely happens, a ‘‘eross-over’’ has

occurred.)

Such a sex-linked lethal gene producing a dominant

wing character has actually been found to occur in the

case of a mutant which arose in the Beaded stock, and

which will be discussed later. For the present we must

note that if the lethal gene were not associated with sex,

- its presence could be detected by the absence of certain

No. 576] VARIATION IN DROSOPHILA 3 723

expected ratios, or classes, or in some other peculiarity of

genetic behavior. In the case before us, we found that

the F, generation consisted of at least two types; viz.,

Beaded and not-Beaded flies. These were shown to differ

genetically. To obtain such a result must mean that at

least one of the parents was heterozygous in at least one

gene. This result is however a fairly constant one; and

by virtue of the long-continued inbreeding of the Beaded

stock this heterozygosity must surely have been weeded

out before now if there were no serious hindrance to

homozygosity. The classic example of this sort of effect

is that of the yellow mice.

But the development of Beaded wings can not be

brought about by the action of a single lethal gene, for if

this were true it would be impossible to obtain a stock

of Beaded flies that would breed true, and yet such a

stock, as has already been said, is the one from which

these very crosses derive their Beaded ancestors. There

must therefore be at least one pair of allelomorphs of

which one member is effective in producing Beaded wings,

and can exist in the homozygous condition and possibly

also another pair of allelomorphs of which one member is

a recessive lethal gene. We can explain many of the

facts so far obtained on the supposition, that there are

these two independently Mendelizing pairs of allelo-

morphs concerned in the production of Beaded wings.

The pair containing the lethal gene we will call L (nor-

mal) and | (lethal); and the other pair B’ (Beaded) and

b’ (normal). The occurrence of the two genes B’ and 1

in one individual usually causes such an individual to

have Beaded wings, though Beaded-winged flies also

occur which do not carry the lethal gene, but are homo-

zygous for B’. |

It should be possible then to isolate a stock of Beaded-

winged flies not carrying this lethal factor, 1. Such flies

should give a much smaller percentage of Beaded-winged

offspring in the F, generation of a cross with Wild stock

(or perhaps none at all, if B’ were recessive), than would

those flies carrying IL. Such a stock has not yet been ob-

724 THE AMERICAN NATURALIST (Vor. XLVIII

tained, but occasionally a strain of Beaded flies is met

with that gives only low percentages of Beaded-winged

offspring. See, for instance, Family 4, Table VII. Pos-

sibly such a stock would not be recognized at once, espe-

cially if it were so affected by environmental conditions

that even flies homozygous for the factor B’B’ sometimes

had normal wings. Normal-winged flies, as will be

pointed out in a later section of this paper, do very fre-

quently appear in Beaded stock, but these flies when

mated to each other appear to throw as many Beaded-

winged offspring as do the Beaded-winged flies of the

stock, and often 100 per cent. of their offspring have

Beaded wings.

In this connection it will be of interest to recall that

Chart 1, and Table I gave results that might be inter-

preted as evidence of the bimodal curve that should be

expected if the above hypothesis is correct.

Normal females from the second generation of Family

2 were also back-crossed to Wild males. The results are

given in Table VIII. Most of these normal females gave

very few or no Beaded offspring (Type X) while two of

them gave a considerable number of Beaded offspring

(Type Y). The explanation here is perhaps that the type

Y females were genetically like most of the Beaded

females of an F, generation (on our hypothesis, B’ L b’ 1)

while the females of Type X were genetically lacking in

the factors that are usually present in Beaded F, flies

(i. e., they were B’ L b L). That such an occurrence is

not infrequent in Drosophila is seen in Table IV in which

three broods out of fifteen raised from normal F, flies

gave 25 per cent. or more of Beaded offspring though the

other twelve broods gave less than fifteen per cent., and

eight broods less than five per cent. of Beaded offspring.

It seems certain therefore that there are two types of

normal-winged offspring in the F, generation of the cross,

Beaded by Wild; one of these is genetically like the

Beaded flies of the same generation and the other is

genetically different from its Beaded brothers and sisters.

Types X and Y have been found to occur in all of the

No. 576] VARIATION IN DROSOPHILA 725

tests made of F, flies whether of matings to Wild stock

or of matings to other mutants such as Black, Pink, Are,

Ebony, ete. Table XXVI shows these two types as they

appeared in back crosses to normal Pink males of normal

and Beaded females of the cross Pink Beaded by Wild.

Here it was found that more of the normal than of the

‘Beaded F, flies were of Type X, and conversely that more

of the Beaded than of the normals were of Type Y.

It has not been possible to distinguish with certainty

between these two types even by their offspring because

of the large amount of fluctuation that occurs in the per-

centages of Beaded offspring. For example it would be

difficult to say whether a fly giving five per cent. of its

offspring Beaded would belong to Type X or Type Y.

It would be expected that Type Y would be given by

those flies that carried both factors for Beaded, and

Type X by those that lack the lethal factor. and it wil)

be seen later that on the whole the evidence supports

this view.

B. Crosses BETWEEN BEADED FLIES AND OTHER MUTANTS

1. The F, Generation

If we examine the F, generation when Beaded flies are

crossed to other mutants, i. e., to flies of a stock that is

perfectly normal so far as Boadedness is concerned, but

which is unlike the normal Wild flies in some other wing

character, or in eye color or body color, etc., we find an

even greater amount of variability in the percentage of

Beaded-winged offspring than in the F, generation of

Beaded by Wild. (See Tables I and IX; also Charts

1 and 3.)

The details may be gathered from Table IX, where it

can be seen that there is a certain specificity in the per-

centage of Beaded offspring that appear in any specific

mating.

For instance, it appears that more of the offspring

have Beaded wings if a cross is made with Vermilion-

eyed flies than when Beadeds are mated to Pink-eyed

726 THE AMERICAN NATURALIST [Vou. XLVIII

TABLE IX

THE PERCENTAGES OF BEADED-WINGED FLIES IN THE F, GENERATION OF

ROSSES BETWEEN BEADED FLIES AND OTHER MUTANTS

Percentages

EEPE aidiaie gdigiy

Ls Jeon Rabari Se o eer a ae

Vermillion: 21). ie BET Bard Ae | EAE ee eel. Pe ie

Miniature Eosin... 1 EEN

E noe e aaa.

Pink Be aded X White i124 JL] Apak a]

Pink Beaded XWild.....)4/4/4/1/1]1)..)....)..)..

ME Sea Behe tec Lege E

Pink Ebony.. 2: 222: ape Lj fee| Dl. feels BY

rod Sue ape cg 1 UP a a ok

Da ma Des ee fee AE leche baa

eerie ns iil Pea

ack E a e a k IER RA AN ears

PMN. ia et eee pape |* chased e Ble aime Ey cep be aero

Voie 05 1 ee pibes toate ts als | a Liste e el ate eas

nai eer Spt ego Se Bet ad ok PoP) Bae Sua A Pee an Gry Gr om Figo Os aed Bg GPE es Eh a

Bap RRO a i ee a a a a ees Res ard cee

oo SENDA Gap 8/6/5)1/5/6)8/5/7/2)-./2/1]..[ hfe ladat 1 fas

Beaded X Wild (Table I).| 1 |5 [10/6 |7 |7 |14|3/2)2/2]..]..)..]..[..[..]..[.h.

Grand Total... 9 |11)15) 7 1213221810 l4 21 1 Eh, fi i

S ek

CHART III

* oa <8 EE YR 2k SB ce wa Sf eS

eh PR CRSA Ee 8 eed

=x WSS Se 2 2 AS a EE a

WRMARBRARRRWIVSRRRa Zr aa

Distribution of broods giving certain percentages of Beaded-winged off-

sprin Beaded X Normal (other Mutants or Wild).

(See Table

ı generation of

No. 576] VARIATION IN DROSOPHILA 121

flies, or more in the crosses with White-eyed flies than

in those with Black body color. (In every case, where

the contrary is not stated the flies are normal in other

respects than the one named, e. g., White-eyed flies in

these crosses have Gray bodies and Long normal wings.)

No explanation of this specificity by the assumption

of a segregation of factors in the germ cells appears to

be available here, though such a possibility has not yet

been ruled out, or can be ruled out till certain other phe-

nomena are understood. The easiest way of ‘‘explain-

ing’’ it is that the dominance of the genes for Beadedness

varies in accordance with many other circumstances,

among which are differences in the other genes present,

such as those for Vermilion, White or Pink. Such an

assumption as this, as will appear later, would seem to be

fully in accord with the behavior of the genes for Beaded

wings when in still different relationships.

It is assumed, then, for example, that the percentages

of Beaded-winged flies in the F, generation of a cross

between Beaded and White are higher than those in the

F, generation of a cross between Beaded and Black, be-

cause the gene for Black is relatively to the gene for

White eyes an inhibitor of Beadedness. It would appear

as though a series might be made of the mutants of -

Drosophila beginning with those in which the genes for

Beaded wings are most dominant and ending with those

in which the Beaded genes are recessive. In order to

construct such a series a large number of pairs would

have to be made for each cross in order to determine the

limits of variability of Beadedness for the cross econ-

cerned. The work would probably be greater than the

value of the results obtained, and therefore the attempt

has not been made to earry out this test. From what has

been done incidentally in carrying out other experiments,

it will be seen that in general the darker eye colors and

body colors are associated with a low percentage of

Beadedness in the F, generation, and the brighter colors

with a higher percentage. This may, however, only be a

coincidence.

728 THE AMERICAN NATURALIST [Vou. XLVIII

2. Linkage Relations

(a) Sex Linkage

If in the crosses thus far described the sex of parents

and offspring that show Beaded wings be considered, it

may appear at first as though we may be dealing with a

partially sex-linked gene. For it very frequently happens

that when the mother is Beaded, and the father is normal

(either of Wild stock or of some mutant stock not carry-

ing Beadedness), more of the sons than of the daughters

are Beaded. For example, in one such brood, there

were 17 Beaded to 128 normal females, and 5 Beaded

to 130 normal males, or 12 per cent. of the females

and 3.5 per cent. of the males. Both of these examples

were deliberately chosen because they were good ex-

amples of the phenomenon described. It would be possi-

ble to select from my records several examples of the

reverse phenomenon, where Beaded females had more

Beaded daughters than Beaded sons, and where Beaded

males had more Beaded sons than Beaded daughters.

Nevertheless, the records of all broods available have

given the numbers shown in Table VI, where it appears

that more sons are Beaded when only the mother is

Beaded and more daughters when only the father is

Beaded.

It may perhaps be significant, on the other hand, that

when the mother is Beaded a slightly larger percentage

of her daughters is Beaded than of the daughters of a

Beaded male, while a very much larger percentage of her

sons is Beaded than the sons of a Beaded male. In other

words, it seems that the daughters are affected to ap-

proximately the same extent, whether they get their

Beadedness from father or mother, while the sons are

affected also by the mother, whether or not she carries

Beadedness. This might mean that there is some gene in

the sex chromosome that does not show except when other

Beaded factors are present. That this is not the case

will appear from Chart 2, which records three generations

of flies in each of which the mother was normal (Wild)

No. 576] VARIATION IN DROSOPHILA 729

and the father Beaded. This shows that although the

father transmitted his Beadedness more to his daughters

than to his sons, yet his Beaded sons also had the capac-

ity to affect their daughters more than their sons, and

these sons again repeated the phenomenon. Yet these

males could not have received their X-chromosome from

their father, unless non-disjunction (see Bridges, ’13)

had occurred. In fact, to produce the results here given

non-disjunction must occur in one half the females of

the Wild stock. Frequent tests with the Wild stock by

practically all of the students in the laboratory make it

certain that this is not the case. I also tested a consider-

able number of the females by mating them to sex-linked

mutants and found no non-disjunction.

This apparent sex-linkage that does not follow the

‘ordinary rules’’ of sex-linkage must be left for the time

being as one of the still unsolved problems. The only

possibility of explanation that occurs to me is that the

above-described effect would be produced if in the cyto-

plasm of the egg of the Beaded female something were

present which is absent in the egg of the normal female,

and to which the males are more responsive in their

development than are the females. This suggestion has

not a particle of cytological evidence to support it. Mor-

gan (1912d) has suggested that the influence of cytoplasm

may cause certain peculiar results obtained in crosses

between Miniature-winged and Rudimentary-winged flies.

(b) Linkage to Sex-linked Genes

Matings of Beaded flies to flies with sex-linked char-

eaters, including Vermilion and Vermilion-yellow, have

been made and the F, generation raised. No sign of

linkage was observed. The F, figures are given in Tables

X and XI. These eases definitely establish that there is

no gene for Beaded wings in the X-chromosome.

Although no sex-linked gene for Beaded wings are

known, there has arisen in the Beaded stock by mutation

a fly with notched wings (Fig. 13) that proved to be

730 THE AMERICAN NATURALIST [Vou. XLVIII

TABLE X

F, COUNTS FROM THE CROSS VERMILION 2 X BEADED ¢

hese Q rg bg Q pioi oe a Q pyel et 9 sarod

% Bd o No. of Bd. if

| no coupling exists

Bd.V : N.V.= 65: oig a E a, e | ree

ba Red: NeR 55o Talore ae: 28.0

Bd.Total : N Total = "i20: ‘Hii SOP wine ae 27.8 | (120)

TABLE XI

F, COUNTS FROM THE CROSS VERMILLION YELLOW X BEADED

Pens" Normal} Beaded | Normal | Beaded | Normal | Beaded | Normal

Gra Gray Gr Gray Yellow Yellow Yellow | Yellow

Zar 35 | 34 41 34 | aa AT 88

Expected No. of Bd. if

% Beaded a coupling exists

44.1 01.7

Les. 5 N Ro 2068 20l La a e a |

| Bd.G. : N. G. = g ‘ >y IEE POT AE SS |. 43.4 199.1

PMS, NY etc Paa 42.7 92.0

| Bd.Total : N. vor gk one We ee ROE 43.2 | (291.0)

TES

caused by a dominant sex-linked

factor lethal when homozygous.

(See page 754.) It will be dis-

cussed under the name ‘‘Perfect

Notched’’’ and its peculiarities

described in the section on ‘‘ Mu-

tation in Beaded Stock.”

some Genes

For the reasons given, it seems

certain that there is in the group

of sex-linked genes no gene con-

cerned inthe production of Beaded

wings. We shall later bring for-

ward evidence to show that there

is such a gene in the third chromo-

some group. (Sturtevant, 1913.) |

The crosses made with flies showing characters whose

No. 576] VARIATION IN DROSOPHILA 731

genes are in the second chromosome are still perplexing;

for while the second chromosome exerts an influence on

the statistical results, as will be evident from the figures

to be presented, the nature of this influence is not fully

determined.

The second chromosome characters with which tests

have been made are the wing characters, Are, Curved,

Vestigial, Antlered, and Strap,' the body color, Black;

and the eye color, Purple.

Vestigial, Antlered, and Strap stand for wing char-

acters of such a nature that it is not possible to distin-

guish Beaded-winged individuals if any of these other

characters are also present. They are therefore of no

use for determining whether or not there is a second

chromosome gene for Beaded wings. The crosses be-

tween Beaded flies and flies with these characters do not

especially interest us here. It may be said in passing,

however, that in every case in the F, generation between

these flies and Beaded flies, from 60 to 90 per cent. of the

offspring had non-normal wings, and the author was put

to serious straits to classify the new wing types that

appeared. These were similar in all the crosses, however,

and on the whole resembled Beaded wings.

In the F, generation, and in back crosses to Beaded

Stock and to Vestigial Stock further complications arose

with more new types of wings, including a new ‘‘muta-

tion’’ which bred true from the start, and which will be

discussed briefly and described under the name ‘‘Spread’’

in the section that concerns mutation.

We may now return to the crosses between Beaded flies

and Black, or Purple, or Are, or Curved. These crosses

give results that can be used for the study of linkage,

and they present in common a number of distinguishing

1 Strap Wings is a mutant much resembling Extreme Beaded in appear-

ance but its mode of inheritance has not yet been worked out. It may be

that it actually is Beaded plus some at present unknown gene. Beadedness

is suspected to occur also in Vestigial and Antlered stock. This may very

likely be true since Strap and Antlered arose in Vestigial, and Vestigial in

Beaded.

732 THE AMERICAN NATURALIST [Vou. XLVIII

characteristics. Tables XII to XX give the results in

sauces form.

TABLE XII

F, Counts oF THE Cross BEADED 2? X CURVED 3

| Beta Norii Be aded No on | ‘ta Ba. a. | Exp. B

| Curved Curved Straight Straight | Curved | Straight

Type to o] å | 61 | 16 | 246 | 4 | 16

E Pewee. | 2 25 18 105. | 3.6 16.4

One of the most striking characteristics of these second

chromosome crosses is that the F, flies fall into two

classes or possibly into three classes with reference to the

offspring that they produce. These classes I have called

Type 1, Type 2, and Type 3.

In Type 1 there is no linkage between Beaded wings

and the second chromosome character, but Beaded-

winged flies occur with equal frequency in all classes of

TABLE XIII

F, Counts oF THE Cross BEADED 2? X ARC ĝ

Beaded Normal Beaded Normal Exp. Bd. | Exp. Bd.

Are Are Straight Straight Are Straight

Tyee i... | 39 | 200 | 184 | 970 | 98+ | 185

SYDO De. oe sas 41 152 330 452 73 t BGS

offspring. In Type 2 there is linkage of Beaded wings

with the second chromosome characters, so that the

Beaded wings appear more frequently in flies showing

the characters of the Beaded parent. In Type 3, which

occurs only a very few times and is not very marked

except in Table XX, Beaded wings appear to a greater

percentage in the offspring whose other characters are

not those of the Beaded parent. (J. ¢.,‘‘repulsion’’ occurs

between the factor for Beadedness and that for the

second chromosome character with which it entered the

cross.) I do not wish to emphasize Type 3, but concern-

ing the other two it is important to note that about one

half of the F, flies seem to be of Type 1 and one half of

| Typo 2,

In re 2 ae in the cases here adduced the linkage is

No. 576] VARIATION IN DROSOPHILA 733

strongest with the character Are and weakest with Black.

If there is a gene in the second chromosome which aids in

producing Beaded wings, it seems probable that it is

located nearer to Are than to Black and on the side of

Are away from Black.

TABLE XIV

BACK-CROSSES OF F, BEADED-WINGED MALES OF THE Cross BEADED ¢@ X

PURPLE CURVED 2 TO PURPLE CURVED FEMALES OF NORMAL STOCK

Normal Pr. Beaded Red Normal Red Exp. Nor. | Exp. Nor.

Beaded Pr,

Curved Curved | Straight | Straight Bd. Pr. Cv.) Ba. . Strt.

Hopsa So 15 125 | 15 | 154 | 13.6-| 164

Types. Gee ee 9 28 OV We io ia

It remains to consider Type 1, and to find the reason

for the existence in the F, generation of flies whose off-

spring show no linkage between Beadedness and second

chromosome characters, and in the same brood, flies whose

offspring do show such linkage. The most obvious ‘‘ex-

planation’’ would be, of course, that the factor in the

TABLE XV

BACK-CROSSES OF F, BEADED-WINGED FEMALES OF THE Cross BEADED ¢ X

PURPLE CURVED 2? TO PURPLE CURVED MALES OF NORMAL STOCK

Brood 2 is of Type 3 for Purple and of Type 1 for Curved.

-| B, Pr. | N, Pr. Ba, Red| N, R. | Bd. R. | N. R.

| | Us :

SEER ents eee tas

Boots: os od art oe 6 0 B 44} 7

second chromosome was a ‘‘lethal’’ such as the factor 1,

described in an earlier section of this paper. If this were

the case, there should be some flies in the Beaded stock

homozygous for L, the normal allelomorph of this gene,

and also for B’, i. e., B'LB’L. These flies should have

fewer Beaded offspring than those heterozygous for 1,

and none of these F, offspring should give linkage with

second chromosome characters. As a matter of fact, in

the F, results given in Table XVI for Beaded by Black,

no Bbc was observed; but this case is not good evi-

dence, for it was made in the first attempts to solve the

problem of Beaded wings, and I had not yet learned the —

734 THE AMERICAN NATURALIST [Vow XLVIII

value of F, counts, matings in pairs, and back-crosses to

normal. It stands however as the only evidence of its

sort that I can give at present.

TABLE XVI

F, COUNTS OF THE Cross BEADED 9 X BLACK ¢

l :

Beaded glen | on Normal | Expected 5 ais

Black Epo Gray z Bd. Bl. Gar.

Tyl 19 eee 1602 | 19 82

Type 3 is not easy to explain. There are no known

cases of this sort elsewhere in Drosophila and I prefer

not to attempt to answer this question at present.

TABLE XVII

BACK-CROSSES al F, MALES OF THE Cross BEADED g X BLACK 2 TO BLACK

FEMALES OF NORMAL STOCK

Beaded Normal Beaded Normal | Expected nde ty

Black Black Gray Gray No, Bd. Bl. Gray

faan ORS sec Zz 162 9 187 7.3 8.6

EVO 2 pice wee 5 110 40 97 20.5 24.5

In general, it may be noted that technical difficulties

have disturbed the crosses with second chromosome char-

acters. The wing character Are is not always easy to

recognize, as it is very often nearly normal in appearance.

On the other hand, the flies with Curved wings, though

always distinct, occasionally get ‘‘stuck up’’ with the

food and in their bedraggled condition it can not always

be determined whether or not the wings are Beaded as

well as Curved. I was at first inclined to attribute the

apparent coupling (which was discovered for Are and for

Curved before it was discovered for Black) to errors

made in the counts.

As for Black, the F, and later generations give a much

lower percentage of Beaded offspring than do most other

crosses, and this necessitates raising large numbers of

offspring. The results are, pokes: trustworthy when

apie’:

No. 576] VARIATION IN DROSOPHILA 735

The crosses with Purple-eyed flies presented no diffi-

culties but ran smoothly aside from the fact that the

purple-eyed flies had Curved wings, and as remarked

above, Curved wings sometimes get bedraggled.

TABLE XVIII

BACK-CROSSES OF F, FEMALES OF THE Cross BEADED £ X BLACK È ro

BLACK MALES OF NORMAL STOCK

Beaded No — | Beaded Normal | Expected | Net Ba.

Black Blac | Gray Gray |Nọo, Bd. Bl.

Typs o. 7 5 Lge og | 91 | 3.3 | 3.7

Type 2 bee ee) a57 k a8 493 16.5. 175

(d) Linkage to Third Chromosome Genes

We have said tentatively that there was perhaps a non-

sex-linked lethal gene for Beaded wings in the second

chromosome, and that possibly the cytoplasm carried by

the egg disposes males toward or away from Beadedness

according to whether the fly that bore the egg was or was

not Beaded. These relations are not securely deter-

mined, and the data are still incomplete. The relation of

Beaded wings to characters whose genes are in the third

chromosome is much clearer. All crosses that bear on

this problem point to one fact, namely, that there is a

gene for the production of Beaded wings in the third

chromosome, and that this gene is very closely linked to

Ebony, and very loosely linked to Pink. Tests have been

made between Beaded and the third chromosome char-

acters, Maroon, Sepia, and Pink eyes and E OnT body

color.

TABLE XIX

F, COUNTS OF THE Cross PINK BEADED gd X BLACK 2

Pink Pink Pink Bd.! Pink N.| Red Bd. | Red N. | Red Bd.| Red N.

Bd. Bl. | N. BL Gray Gray Black Black Gray Gray |

Type 1 (?) 2 25 20 57 0 45 3 278 |

Expected No. Bd. if no.

coupling occurs |

Bd.Pink : N.Pink = Fant "3 OREO RGA ON nse rene eet 6 |

Ba Rod = Ned. = 3:9238.. rer eis 19

Bd.Black : KAA. = T 1 E a A rae ee |

Bd.Gray : N.Gray = pa Ver E EE A 21 |

T Ay N.Total = 25 : ba festa VS (25) |

736 THE AMERICAN NATURALIST [Vou. XLVIII

In the cross of Beaded by Maroon-eyed flies, 1,369 flies

were raised in the F, generation. Fifty-seven of these

flies had Beaded wings; only one of the Beaded-winged

flies had Maroon eyes, while fifty-six were red-eyed. (See

Table XXI.)

TABLE XX.

BACK-CROSSES OF F, FEMALES OF THE Cross PINK BEADED ¢@ X BLACK 9

TO PINK BLACK MALES FROM NORMAL STOCK

Pink Bd. | Pink ar Pax Bd.! Pink N. e ep Red N. ae Bd.| Red N.

Black | Blac Gra Black | Gray | Gray

Type tens: 5 98 8 | 113 5 92 1 | 122

aeaa aio cs. 12 68 6 | 58 7 78 3 | 76

Sa d ae ee eee eee del a l

Expected No. Beaded if no

upling occurs

Ba Piik: N Pank St Bele a ps oe vse eee 23

Baked : N Rod = 16 : 30B. oe Saye 24

Bd- Black: N.Biack = 20 2336. ena 23

Bd .Gray : N Oray = 418 2 360) Pcs. i ns a es

Bg oul : N TO ae aro OD Oe ee | (47)

F, males of the cross Sepia by Beaded were back-

erossed to normal Sepia females. Inasmuch as cross-

overs probably do not occur in the male (Morgan, 1912c),

no Beaded Sepia flies should occur in the offspring of this

cross. Table XXII shows that none occurred. The num-

bers are not large, but since they are entirely in accord

with the other third chromosome results, it was not

thought worth while to increase them. That apparent

cross-overs may very rarely occur will appear possible

when we consider the results of crossing Beaded by Pink,

and the probable significance of the phenomenon will be

considered.

TABLE XXI

F, RESULTS OF THE Cross BEADED Ọ X MAROON ¢

1 | 318 | 56 | 994 13 44

In F, counts of the crosses involving Beaded and the

body color Ebony, totaling 4,417, in which 1,205 Beaded-

winged offspring occurred, not one had the body color

_ Ebony, and only eleven had Pink eyes. Repeated attempts

No. 576] VARIATION IN DROSOPHILA tor

to obtain Ebony flies with Beaded wings have failed. The

possibility that for some ‘‘inherent peculiarity’? an

Ebony fly can not have Beaded wings has suggested itself,

TABLE XXII

BACK-CROSSES OF F, BEADED MALES OF THE Cross BEADED g X SEPIA 2 TO

SEPIA FEMALES OF NORMAL STOCK

Bd. Sepia | N. Sepia | Bd. Red | N. Red

0 | 134 | 9 | 132

and although this would seem very improbable, it may

nevertheless be the fact. At any rate, it appears that

Beadedness either depends on genes which in the presence

of the Ebony body color are completely recessive, or that

the third chromosome gene for Beadedness, B’, lies so

close to that for Ebony that cross-overs are extremely

rare even in the female.

TABLE XXIII

F, RESULTS OF THE CROSS BEADED 9 X EBONY ¢

Beaded Ebony | Normal Ebony | Beaded Gray | Normal Gray

0 | wi ae 17 | 525

Very extensive experiments were carried out with Pink-

eyed flies. The important facts brought out are presented

in condensed form in Tables XIX, XX, and XXIV-XXIX.

In Table XXIV are shown the F, counts for Beaded by |

Pink Ebony. From the eleven Pink Beaded flies obtained

a new stock was derived, which was ‘‘purified’’ by a few

TABLE XXIV

F, RESULTS OF THE Cross BEADED Ẹ X PINK EBONY ¢

es g Bå. P. Gray

LS ee

N. P. Eb.

847

“Bd. R. Eb. | N R. Eb.

N. P. Gray Ba. R. Gray N. R. Gray

0 f 1%

167 | L77 | LS%

Expected No. Bd. if no coup-

| ling occurs

738 THE AMERICAN NATURALIST [Vou. XLVIII

CHART IV

Family 1.

Bd 9, 86; 25.6%

Bd 9, 226 ; 25.7%

I l

Bd Ẹ, 110; 27.3% Bd g’, 405; 19%

Bd g’, 135 ; 8.9%

Family 2.

Bd , 460; 28.9%

Båg Bd g Bd Bd oB d

d l | | d

F € Bdo Bið Bas? Båg Bde

154 248 129

47.4% 106 88.86 14.9% 12.9% 5.4%

135 186 192 161

23.7% 14.7% 81.26 289%

] | |] | l

Big Big Bde Bdg Big BAF BAF Ba? BAP BAF BAF Bag Bag Ba

158 177 162 147 22 47

46.8% 38.4% 8.69

143 231 2

8.6% 42.9% 2 ae 15.54 37.66 11.9% 2824 03.54 13.49% 1624 134

Py i4

Bi a Big Bd g

106 27 63

19.8% 33

Bid, 26

Bd o, 193 Bag, 237

50% 16.6%

Bd 3, 190

9.7% T

Bdg a Bdg Bi Bdg Ba Bdg Bda Bdg Bdg Ba Ba as Bdg an Bia Bda Big Bdo

168 185 21S. 185 105.

5T 41.5% 36.94 154s 6.3% 134 al. p 17.8% 16.1% 19% 13.5% 18.7% 20.5% 60.6% 87.26

Family 4.

Bd 9, 48; 4.2%

Ba ð, 137 ; 1.5%

Bd A, 209 ; 2.4% ag

Bd f, 237; 4.2%

Family 5.

Bd 9, 82; 15.9%

Ba i 814; 7.6%

Ba d, 175; 9.7% Bd 7, 144; 24,3%

Bd d, 50; 32% Bd of, 82; 20.7%

o back-erosses of Beaded X Wild in successive generations, show-

ing sex of — pori number of offspring, and percentage of offspring

No. 576] VARIATION IN DROSOPHILA 739

generations of selection, and now gives approximately 100

per cent. Beaded offspring, though no selection has been

practised for nearly a year. This stock has been used in

one series of crosses to supplement another series in

which Pink and Beaded enter the cross from opposite

parents. The results in each case are essentially similar,

and show that when Beadedness enters with Red it comes

TABLE XXV

F, RESULTS OF THE CROSS PINK BEADED X WILD

Bd. Pink

wo | 47

Normal Pink | Beaded Red | Normal Red ‘Exp. No. Bd. P.|Exp. No. Bd. R.

| }

o ee 213 | 366 | 964

out more with Red than with Pink. They show that in

the F, female crossing over occurs almost independently

of Pink, so that almost the same percentage of Beaded-

winged individuals appears in each class, though usually

the class that is similar to the Beaded parent is consider-

ably the largest. In Table XXVII, however, a record

is given in which a very considerable ‘‘repulsion’’ oc-

curred, and the high Beaded class is not Pink Beaded,

as is there expected, but Red Beaded. The results from

back-crosses of the brothers of these females to Pink

normal stock show that no mistake was made in record-

ing the cross, which therefore, though somewhat surpris-

ing, must stand.

TABLE XXVI

BackK-CROSSES OF F, FEMALES OF THE Cross PINK BEADED X WILD TO PINK

MALES oF NORMAL STOCK

<< oo bo gl. gas 2 $37 | 25 | 2.5

Tie Y... | 71 332 58 369 | 626 | 66.4

teal. | za | w 60 706 “ae

The tables show also that in the males, crossing over is

of very rare occurrence, if, indeed, it occurs at all. The

records show that out of 566 Beaded flies (Tables XX VII

and XXIX) which occurred as the offspring of an F, male

740 THE AMERICAN NATURALIST [Vou. XLVIII

back-crossed to Pink normal stock, six flies of the cross-

over class appear. For reasons to be mentioned, it is

improbable that these represent cross-overs, however, but

rather they may be due perhaps either to the presence of

the second chromosome gene, l, which usually does not

manifest itself in the absence of the third chromosome

gene, or to mutation, or to some unknown abnormality.

Through carelessness only one of these males was tested

TABLE XXVII

BACK-CROSSES OF F, MALES OF THE CROSS PINK BEADED X WILD TO PINK

FEMALES OF NORMAL STOCK

heamana. ia T sean a S un se? rote — TAAT? ER AE REA rar UN ne ae See SR GO AEA IY eam aaa on Sena BR ty a

Bd. Pink | N. Pink | Bd.Red | N.Red |Exp. No. Bd. P. Exp. No. Bd. R.

| |

56 | 710 | 5 | 805 b> ae | 32

or used further in breeding. They were very slightly

Beaded, and had only a very slight ‘‘nick’’ at the tip of

the wing, even smaller than that shown in Fig. 3. The

single Pink Beaded male mentioned in Table XXIX was

mated to several females but was sterile. Another test

is also possible, and was made as follows. Pink normal

males and females from Table XXIX, which of course

should not carry the third chromosome gene for Beaded

TABLE XXVIII

REPEATED BACK-CROSSES OF F, FLIES OF THE Cross PINK BEADED X WILD

TO PINK FLIES oF NORMAL Stock

Exp. Exp.

ae ee i GA

Hel MPN ot Type ea 1 228 1 272 1 1

Bred XPINE Oo Type A.o. 24 379 79 383 48 55

ORK. CS ee EA AE E 5k oe os 25 607 80 655 49 56

DOP oss ick ee 7o 193 0| 135 | 35 185.

wings, were then mated together, and among their 374

offspring three males with slight ‘‘nicks’’ at the tip of

their wings, exactly like those of the Pink Beaded male

before mentioned, were produced. One of these males

= was sterile. One of the remaining two was fertile, but

No. 576] VARIATION IN DROSOPHILA T41

gave no Beaded offspring either in the first generation or

in the F, generation, although nearly one thousand of his

grandchildren were carefully examined. The remaining

male was abundantly fertile and had one son exactly like

his father in appearance (with a slight nick at the tip of

the wings). The rest of his offspring were normal. This

son was sterile.

TABLE XXIX

BACK-CROSSES OF F, FLIES OF THE CROSS BEADED X PINK TO PINK FLIES

OF NORMAL STOCK

Beaded | Normal | Beaded Normal Exp. Bd. | Exp. No.

Pink | Pink | Red Red | Pink Beaded R.

Fic! XPink 2 .. 1 | 580 | 223 | 282

F192 XPink:ot i: 70 114 | 114 106 | 84 | 100

The results of these tests with five of these supposed

‘‘cross-over’’ males show clearly that they were not nor-

mal Beaded flies. As said, they might represent muta-

tions, or the dominance of the gene 1, or some abnormality.

These are mere guesses, but since there are no authentic

cases on record in Drosophila of crossing over in the male

sex in those cases where the mutants dealt with are well

known genetically, i. e., since the only apparent cases

occur in the Beaded wings and some of the other not

well-known and peculiar mutants of Drosophila, we are

not justified in assuming that such crossing over takes

place here.

III. THE EFFECT OF ENVIRONMENTAL CONDITIONS UPON THE

PRODUCTION OF BEADED WINGS

A. QENERAL STATEMENT

If we have so far interpreted the evidence correctly we

may formulate the following statement as a provisional

hypothesis. A gene B’ located in the third chromosome

near that for Ebony is directly responsible for the pro-

duction of Beaded wings. By itself in the homozygous

condition, the fly bearing it may have normal wings,

though it usually will have wings somewhat Beaded. In

the heterozygous condition, it is rarely, though sometimes,

742 THE AMERICAN NATURALIST (Vor. XLVIII

dominant. The conditions so far presented which cause

it to be dominant are two. (1) The presence of a gene |

in the second chromosome which can not exist in the

homozygous condition. (2) The influence, particularly .

noticeable in the males, of non-chromosomal constituents

of the egg from which the individual arose, so that if the

mother had been Beaded, the appearance of Beaded wings

in her sons would be increased, and if the mother had been

normal the appearance of Beaded wings in her sons would

be reduced.

Certain facts already brought out (namely, those pre-

sented in Tables II and IIT) show that the tale is not yet

told. Our hypothesis does not explain the fact that from

definite numbers of eggs laid at different periods in the

life of an individual very different percentages of Beaded-

winged offspring arise, and these differences do not form

a definite series progressing to or from a high percentage

as the individual grows older, but are extremely irregular.

We have not gained control over this phenomenon, but

the evidence we have to present points strongly to the

suggestion that the environmental conditions are the final

determiners of the percentage of the Beaded-winged off-

spring. This environmental control might lie in three

distinct methods: (1) The destruction of a certain class of

offspring by their differential viability. (2) In the case

of Table III the results might be explained on the theory

that Beaded flies had a shorter life cycle. This supposi-

tion has, however, been disproved as follows. Five non-

virgin females from Beaded-winged stock and five non-

virgin females from normal-winged stock were put to-

gether without males in the same bottle. When the off-

spring began to hatch they were examined daily. During

the first three days 73 flies hatched, of which 11, or 15 per

cent., had Beaded wings. During the following five days

261 flies hatched, of which 54, or 20 per cent., had Beaded

wings. Since I was particular to take Beaded flies several

days old as the parents of these Beaded offspring, the

experiment shows that if there is any difference in the

length of the larval life, that of normal-winged flies is

No. 576] VARIATION IN DROSOPHILA 743

the shorter. (3) The determination of whether or not a

fly of a given germinal constitution shall have Beaded

wings. The first of these effects is probably not the

significant one, in view of the following facts.

Although as a rule F, normal flies give few Beaded off-

spring, and F, Beaded-winged flies relatively many,

nevertheless, as has been said, at times normal flies give

a high percentage of Beaded offspring and, occasionally,

Beaded flies give a low percentage. This can only mean

that the dominance of the factor B’ is variable, and con-

sidering the large number of times that it shows itself as

a recessive, it must be that this varying dominance has a

marked effect on the percentage of Beaded-winged off-

spring that appear.

The possible amount of variation in the environment

surrounding a brood of Drosophila developing under

laboratory conditions is enormous, even when the attempt

is made to keep conditions constant. These variations

depend upon the exact ripeness of the bananas used as

food, the length of time the food has been fermenting, the

amount of food and filter paper used, the size of the bottle

in which the larve are developing, the tightness of the

cotton plug, the temperature of the laboratory, ete. Due

to these causes there arise very great differences in the

relative moisture content and carbon dioxide content. If

the food is not properly prepared it may rot instead of

fermenting, or it may mould, or the reaction may be in

one bottle quite alkaline and in another very acid. A

perfect control thus becomes an impossibility, and there-

fore the experiments to be described must be considered

as trials only, and not as decisive tests.

In all the experiments on this subject, Beaded flies of

pure stock were mated to normal flies of Wild stock in

order to learn the effect of particular environments on

the percentage of Beaded offspring in the F, generation.

On our hypothesis, the pure Beaded flies from stock should

be of two kinds, viz., those with the lethal gene 1 (i. e.,

B’IB’L), and those without 1 (i. e., B’LB’L). Correspond-

ingly there should be two types of offspring in the F,

744 THE AMERICAN NATURALIST [Vou XLVIII

generation, one of which (B’lb’L) should have a consider-

ably higher percentage of Beaded offspring than the

other (B’Lb’L). If it is possible, however, that B’ should

be dominant in the heterozygous condition and in the

absence of 1, then it should also be possible theoretically

to produce an F, generation every individual of which

should have Beaded wings, while those with | as well as

(constituting one half the progeny) should have a

more extreme form of Beading. In practice it is not

usual even under the best of conditions to get more than

40 per cent. of Beaded-winged flies, while, as has been

seen, the average amount is about 25 per cent.

B. Tar Errect or RELATIVE MOISTURE

Table XXX and Charts 5 and 6 present the data for

TABLE XXX

PERCENTAGES OF BEADED-WINGED FLIES IN THE F, GENERATION IN RELA-

7 Y WET AND Dry Borties. Nor DONE IN PAIRS, BUT

EACH BOTTLE CONTAINED SEVERAL PAIRS

Dry Bottles Wet Bottles

No.1 No. 2 No, 3 No. 4 No.5

No, No. No. No, No.

ist count.: 27%: LG 82:1 48:7 | 26 26.9 | 14-286) 32

2d count. <. ss-a 122 | 14.7 68 | 17.2 79 | 32.9 BT | 20.7 (IB 207

Sd pount, e a 34 | 20.6 19 | 21.1 24 | 41.6 16 | 37.5 37 | 37.9

4th ecunt.<.. 3. oo N 34 8.8 17 | 41 43 | 41.7

Sth oount. cs 59 | 20.81 72 250i 35 |314] 30 | 86.7 | 53 189

6th count. ..:.:. 95 | 10.5 37 | 30.1

LORE oc ei oe 281 | 17.8 | 252 | 24.6 | 203 | 33.5 | 164 | 28 290 29.3

Total Dr... 533 % Bd. 20.5 Total Wet, 657 | % Bd. 30.3

Counts not made every day.

Bottle No. 2 was very dry and the flies very small during time of last

two counts.

this test. The parents were put into bottles of similar

size with plenty of food. In three of these bottles the

food was very wet and from time to time juice was added

in sufficient amount to keep the food saturated. The

other two bottles were made relatively dry by putting a

No. 576] VARIATION IN DROSOPHILA 745

CHART V

% “I S =

> 2 o 2 2

5 5 2 $ -

Ea : : =

50%. ee = as “ag

40 h

30%

Wet

10%

Effect of Relative Moisture in Food on Percentages of Beaded-winged

Flies in F, Generation of Beaded X Wild.

large amount of filter paper into the bottle at night and

removing it the following morning. After two or three -

days of this treatment the bottles were so dry that I did

not venture to carry the process farther; the flies from

CHART VI

a a a Š % 3

D 37 > 2 p o

E g £ £ £

a4 | * 3 S 3 5,

y m

20%

ean Wet

10% Dry!

Effect of Relative Moisture in Food on Production of Beaded Wings, as

shown by Individual Bottles.

746 THE AMERICAN NATURALIST [Vow XLVIII

these dry bottles were rather small and in bottle No. 2,

they were extremely small in the last two counts.

From Chart 6, where the records are given of the indi-

vidual bottles, it will be seen that there is a good deal of

irregularity from day to day.

Special attention should be called to the curve of pro-

duction of bottle 2, which beginning with a high per-

centage of Beaded Sifepeine gives fewer and fewer for

the first four counts (about six days) and then the per-

centage rapidly mounts again. The offspring given dur-

ing the last two counts were of surprising minuteness and

gave as high a percentage of Beaded individuals as the

average of all the bottles on the first day. It has been sug-

gested that it may not be wetness or dryness or any one

specific thing that brings out the Beadedness, but condi-

tions that are unfavorable to the organism as a whole,

resulting in poor nourishment. It has frequently been

TABLE XXXI

THE INFLUENCE OF ACID, ALKALINE, AND FRESH Foop ON THE DEVELOPMENT

OF BEADED WINGS

Oo |] & we} wi) os D ne . a

diale y Eka 2a ERSS

Food So

Mother Beaded; Father normal........ 26 40 151 129 14.7 23.8 19.1

Father Beaded; Mother normal........ 9 7 | 60 61 13.0 10.3 11.7

ood Fresh

Mother Beaded; Father normal........|23| 25; 85 92 21.3 21.4 21.3

Father Beaded; Mother normal........ 54 | 15/147) 164 26.9 8.4 18.2

Food Alkaline

Mother Beaded; Father normal ........ 3641| 40; 41 | 46.8 | 50.0 | 48.7

Father Beaded; Mother normal ....... 28 | 16 | 57 58 32.9 21.6 27.7

noted that those bottles which gave very tiny flies gave

also a higher percentage of Beaded individuals than the

bottles whose flies were of average size. On the other

hand, the first flies of a brood are almost invariably larger

than the later ones, and yet, as has been seen, they are

more Beaded. This is a paradox, but the behavior of

bottle No. 2 suggests that as a hatch proceeds and the

bottle becomes drier, there may be a certain optimum

~ point for the production of normal winged offspring, and

No. 576] VARIATION IN DROSOPHILA 747

that this point is so low that the flies are poorly nourished

for lack of water, though they can survive an even

greater water reduction.

It is, perhaps, needless to say that an effort has been

made after these experiments to keep the moisture con-

tent high and fairly uniform in cases where other envi-

ronments were being tested.

C. THE EFFECTS or COVERING WITH PARAFFINE THE MOUTH

OF THE BOTTLE IN WHICH THE FLIES ARE DEVELOPING

On observing that the proportion of Beaded to Normal

offspring was lowered as a hatch continued, it seemed

possible that this might be due to one or to both of two

causes: (1) The diminishing water content. This matter

has already been considered. (2) To a changing carbon-

dioxide content. When a brood is first counted the cotton

plug that has been for several days in the mouth of the

bottle is removed, and in removing the flies the air within

the bottle is very apt to bè much changed. With this pos-

sibility in mind a number of bottles were supplied with

food and flies, and after ten days (when the larve were

beginning to pupate) the parent flies were removed, a

little new food put into the bottle and a paraffine cap

melted over the cotton so that the bottles were tightly

TABLE XXXII

Counts or SEVEN BROODS WHICH HATCHED DURING Two PERIODS, THE First

oF WHICH WAS SPENT IN A BOTTLE SEALED WITH PARAFFINE,

AND THE SECOND IN A BOTTLE COVERED WITH CHEESE

CLOTH. BOTTLE No. 7 WAS NOT SEALED WITH

PARAFFINE BUT HAD BEEN LIGHTLY

STOPPERED WITH COTTON

| Bottle 1} Bottle 2 2 Bottles 3 Bottle4 Bottle5 Bottle 6 Bottle 7

sd gliosis cal g 3 33| 3168] g 58 n 4

ZE Se) Mel a ZE] ZE] a E| a ZE S

S go ea |75 |36 82/40 54/44 51/45 59/29 18|39 87 | 28

seescvsstvcess.[64| 032] 3/17] 6/22/18! 36) 0/15/13| 52) 2

Total, first count, 426, per cent. Bd., 36.

Total, second count, 228, per cent. Bd., 6.

Total, both counts, 654, per cent. Bd., 25.

748 THE AMERICAN NATURALIST [Vou. XLVII

sealed. At the same time other bottles were very loosely

covered with a light cotton plug. The bottles remained

covered till flies had been hatching for four or five days

and then the plugs were removed and the flies counted.

The paraffine plugs were not replaced; after carefully

renewing the air in the bottles, they were covered with

cheese cloth and their brood counted again in four days.

The results of this test are given in Table XXXII. The

results are striking enough at first sight, but I do not

know just what their significance is. They show exactly

the same phenomenon that is described earlier and illus-

trated in Table III. They are more striking than any

case I have yet found of the sort, and yet the first infer- —

ence drawn, viz., that the markedly higher percentage of

Beaded flies in the first count is due to these flies having

undergone their late development in a ‘‘close’’ atmos-

phere, must be qualified by the statement that ‘‘close’’

does not refer to the carbon-dioxide content.

At first suspecting this to be the case, I made an appa-

ratus by means of which fresh air could be drawn through

a bottle during the entire development of the brood. By

this means the carbon-dioxide content could not become

very high. In order to prevent drying out, a large amount

of food was put into the bottle and the air which was to

enter the bottle was first passed through water. The

hatching period was prolonged in the cool sink. The re-

sults were decisive. One hundred and sixty-nine flies

were hatched in the first four days, of which 32 per cent.

were Beaded. One hundred and eighty-four flies were

hatched in the next four days, of which 10 per cent. were

Beaded.

The same flies that were the parents of this brood were

in the meanwhile transferred to another bottle, which was

covered with paraffine. The first four days of hatching

gave 108 flies, of which 15 per cent. were Beaded.

This case shows conclusively that the carbon-dioxide

content of the bottles is not the feature of the closed

bottles that determines whether or not a fly shall have

Beaded wings. It leaves the question still unsettled as

No. 576] VARIATION IN DROSOPHILA 749

to the effect of moisture, but corresponds to the results

obtained in the study of moisture effects.

D. THE EFFECTS or ÅCIDITY AND ALKALINITY OF THE F'oop

Normally the reaction of food at the time of putting it

in the bottles is acid, the degree of acidity depending upon

the length of time it has been fermenting. This sourness

usually passes gradually away as the larve grow older,

and by the time a brood begins to hatch the reaction is

frequently quite alkaline, unless fresh food has been put

recently into the bottle.

On the other hand, if the acidity of the food is neutral-

ized at the beginning with sodium hydrate or carbonate,

or if the reaction is made alkaline while yet there remains

a good deal of unfermented banana, the acidity will re-

turn for a time if not carefully guarded against. There-

fore to keep the reaction acid or alkaline is a difficult

matter, and requires occasional stirring of the food to

make the reaction uniform; this operation is likely to

prove disastrous for the developing pupe.

In the tests here recorded I used food that had been

fermenting for one month, so that it had a very acid

reaction that lasted till hatching time. For studies of the

effect of alkalinity I used food that had been fermenting

about one day and mixed with it sodium carbonate,

sodium hydrate or ammonia. The results were unsatis-

factory and the reaction did not remain constant in spite

of my efforts, though on the whole it remained alkaline,

and became strongly alkaline, and also slimy towards the

end of the experiment, and not a great many flies hatched.

I also used food that had not been allowed to ferment

at all, and although I do not know its reaction, it was

certainly not so alkaline as the last mentioned, nor so

acid as the first. It was soon attacked by mold (Bread

mold). I refer to it here as fresh food.

The results are given in Table XX XI, but may be more

briefly summarized here.

Of 483 flies raised on sour food, 17 per cent. were Beaded.

Of 605 flies raised on fresh food, 19.3 per cent. were Beaded.

Of 317 flies raised on alkaline food, 38.1 per cent. were Beaded.

750 THE AMERICAN NATURALIST [Vou. XLVIII

In other words, a high percentage of Beadedness came

from flies raised on alkaline food, a low percentage from

flies raised on acid food, and intermediate amount from

flies raised on fresh food.

A careful study of Table XXXI will reveal the curious

partial sex-linkage of which I spoke on pages 15 et seq.,

and here, too, the explanation suggested there seems to

apply as in other cases of the sort. It is not a little

peculiar that in all of these food tests this phenomenon

should have occurred, though I consider this purely a

coincidence. In any case, if we can draw any conclusion

at all from its appearance, it would only be that the re-

action of the food has nothing to do with the occurrence

of the phenomenon rather than the reverse.

E. Tue Errects or RELATIVE TEMPERATURES

No evident effect was produced by rearing the F, gen-

eration in an ice-chest, but ratios were as varying as when

the flies were raised at room temperature. Ratios of 15.4

per cent., 19.2 per cent., 10.3 per cent., 20 per cent. of

Beaded offspring are examples of those given by broods

raised at low temperatures. The cold does, however,

lengthen greatly the larval life and flies were in the case

of the brood last mentioned twenty-eight days in hatching.

The brood consisted of 312 normal and 77 Beaded-winged

flies,

Similar results were obtained in experiments with

heat, except that here the larval life was correspondingly

shortened and was at times reduced to eight days. It

was not found practicable to keep the flies at higher than

30°-33° Centigrade, as they soon died at higher tempera-

tures.

F. Tue Errects or Darkness

Flies were raised in complete darkness and sister

broods in full daylight, but no differences appeared in the

offspring. Of 484 flies raised in darkness 30 per cent.

had Beaded wings. Of 360 flies raised in the daylight,

29 per cent. had Beaded wings. This experiment seems

No. 576] VARIATION IN DROSOPHILA 751

to show conclusively that light and darkness do not influ-

ence the.percentages of Beaded-winged flies.

IV. THE EFFECT OF SELECTION ON THE PRODUCTION OF

DIFFERENT TYPES OF BEADEDNESS

Just how much can be accomplished by selection in

Beaded stock was one of the first questions that arose.

Morgan (1911a@) describes the origin of pure Beaded stock

as having occurred through the selection of Beaded flies in

the early generations after its first appearance. He says

the first Beaded fly found arose in a culture of Droso-

phila that had been exposed to radium. Mated to his

sisters, 1.6 per cent. of the offspring were Beaded. When

these Beaded flies were inbred 3 per cent. of the offspring

were Beaded. These inbred gave 8.5 per cent. Beaded

offspring.

The same process continued through many generations has finally

produced stock that gives in certain cultures nearly 100 per cent.

Beaded wings.

In continuing these selection experiments, he says more

extreme forms of Beaded wings appeared, and at the

time of publishing (March, 1911) he was attempting ‘‘to

fix some of these extreme variations.’’ While engaged in

this work other wing forms arose, most of which are

among the best-known mutants of Drdsophila. Among

these are Truncate, Miniature, Rudimentary, Vestigial

and Balloon wings, and the Black and Yellow body colors.

Most of these forms have been ‘‘purified’? now and

Beadedness never appears in them though it can still be

found in Vestigial stock. All of the above-named forms,

by the way, with the exception of Truncate and Rudi-

mentary bred true from the start. The Truncate case is

not yet published and Rudimentary has proved (Morgan

and Tice, 1914) to be due to a single Mendelian factor.

The Rudimentary flies were at first self sterile and highly

non-viable, and therefore gave peculiar results in breeding

tests.

When I first began work with Beaded flies (Sept., 1912)

the stock gave 100 per cent. Beaded-winged offspring.

752 THE AMERICAN NATURALIST [Vou. XLVIII

So soon that I did not realize it, nor think to count the

generations, I had one stock that gave offspring much

more extremely Beaded than the ordinary stock, and this

stock is the one on which most of this report is based.

About December, 1912, I started one stock bottle to form

the basis for a ‘‘No selection ”’ test. The parents of this

brood were ‘‘pure stock Beaded’’ males and females.

The first generation, no normal-winged flies appeared.

The generations following were made up by shaking at

random from the bottle of the generation before a dozen

or two flies into a new bottle.

The second, third and fourth generations gave three

normal-winged flies to 325 Beaded. The sixth, 3 normal

to 100 Beaded. In later generations I occasionally found

normal flies. The stock is in its 27th generation now,

the 25th generation having given rise to a large brood

of which I counted 541 flies (284 2 and 257 g), all of

which had Beaded wings of a type averaging like those of

Figs. 4-6. Itis very apparent that the stock is not under-

going any marked change, though I can not guarantee

that it would give exactly the same results in other

respects as the extreme (selected) Beaded stock that I

have used in the linkage tests.

On the other hand, I have not been able thus far to

increase the Beadedness of the selected stock beyond a

point which it apparently reached many generations ago.

The Figs. 1-12 (excepting 2 and 4), which are here re-

produced, were made under Dr. Morgan’s direction long

before I took up the work, and the forms he had drawn

then are as extreme as any that I now have.

If this extreme stock be allowed to go without eden

for two or three generations, it ‘‘reverts’’ to a less ex-

treme form, from which it can apparently be recovered

by one mass selection. I feel confident that in selecting

the extreme forms one merely selects a large percentage

of individuals that are heterozygous for l, and of course

when the stock is not selected for a while, LL forms be-

come relatively more numerous. This would account for

all the facts here recorded.

No. 576] VARIATION IN DROSOPHILA 753

On the other hand, selection for less extreme Beading

is also rapidly effective and normal-winged forms appear

soon, but this effect soon reaches its limit apparently,

and a normal strain or even a strain throwing a high

percentage of normals has not yet been obtained. I am

not yet certain that it can not be done. I selected in each

direction for eleven generations without marked success

beyond that here recorded.

V. MUTATION IN BEADED STOCK

A, GENERAL STATEMENT

As will be gathered from statements made in the last

section, the Beaded stock has been prolific in giving muta-

tions. There has been no especial attempt made to see

how many different mutants could be obtained from the

stock, and yet a goodly number have appeared. Most of

these have been marked types showing little variation and

coming out regularly and distinctly in Mendelian propor-

tions in crosses with other types. They have in general

bred true from the start without further selection.

A few of these have been of a sort to confuse for a time

the study that I have been making, because of their re-

semblance to certain types of Beaded flies. The criterion

in every case as to whether or not a fly was an ordinary

Beaded fly or a new ‘‘mutant’’ was its genetic behavior,

and the cases to be here described have, with the excep-

tion of Stumpy, shown themselves to be due to a single

gene conforming in general to those of other well-known

mutants of Drosophila.

B. Prrrect NorcHep Wines

In the beginning of my work on Beaded wings I thought

it might be possible to isolate definite types from the

Beaded stock by crossing out to Wild and extracting the

F, types that appeared; or by back-crossing the F, forms

to Wild again and extracting new types, ete. Several

thousand flies were raised in the hope of accomplishing

this, but the ‘‘types’’ found did not breed true, but con-

tinued to behave like ordinary Beaded flies, from whose

754 THE AMERICAN NATURALIST [Vow. XLVIII

many original types none were distinguishable. Finally

a genuine new ‘‘type’’ appeared, with both wings alike

and definitely ‘‘notched’’ (Fig. 13, p. 730). This female

which was at sight named Perfect Notched, was mated

to Wild. Her ancestry was as follows: :

The grandmother came from pure Beaded stock, and

the grandfather from Wild stock. Their offspring con-

sisted of 18-Beaded and 69 normal flies.

A Beaded female of this generation was mated to a

normal brother and gave 100 Beaded offspring, male and

female, and one ‘‘perfect notched’’ female.

This female and her descendants behaved in a very

different manner, genetically, than the Beaded stock

from which she arose.

She was mated to a Wild male and gave 62 Beaded off-

spring and 112 normal offspring. Of the Beadeds, 50

were notched in a way resembling the parent and of the

50, 49 were females. Several other peculiar wing types

appeared among the remaining 12 Beaded flies of this

generation, but did not breed true and were later dis-

carded.

The notched male gave ordinary Beaded and normal

offspring and never gave in either the first or later gen-

erations any ‘‘notched’’ offspring. He was probably an

extreme variant of a common Beaded type (Fig. 4).

Of the normal offspring of the Perfect Notched female

four pairs were made up. Seven hundred and forty-nine

normal sons and daughters appeared, and no notched.

Of the notched daughters of the perfect notched female,

two were mated to normal brothers and two to Wild

males. Their progeny was:

By normal brothers ...... 53 0 79 69

By wild males. :......-.: 56 0 47 46

Six of the notched females of this generation were

mated to normal brothers and gave

Notched 9 Notched @ Normal 9 Normal g

126 0 144 120

At this time, June, it was necessary to leave New York.

_ In traveling, the Perfect Notched stock was lost. Enough

No. 576] VARIATION IN DROSOPHILA "(155

had, however, been done to show definitely the nature of

the mutation involved. It is clear that the perfect notched

wings owed their appearance to a dominant sex-linked

gene, lethal for males. This accounts for the fact that

the males are only half as numerous as the females, and

none of them notched, while notched and normal females

occur in nearly equal numbers. It also accounts for the

fact that the normal females of these generations gave no

notched offspring.

Other sex-linked lethal genes have appeared from time

to time in the crosses of Beaded flies with others, but

none of them were dominant, and therefore they made

themselves evident only by preventing the development

of one half of the males. I have not worked out the

inheritance of these cases.

C. Spreap Wines

Comment has already been made on the extreme num-

ber of wing types that appeared both in the F,, F,, and

back-cross generations of the cross between Beaded and

Vestigial flies. Most of these forms gave results too com-

plex to be analyzed at present. However, among the off-

spring of a considerable number of the F, females there

were flies with wings perfectly normal in appearance save

that they were held at right angles to the long axis of the

body. In all, 60 flies with Spread wings appeared. One

of the 60 had wings very slightly Beaded. Some of them

were mated together and produced only spread-winged

offspring with no sign of Beadedness. Spread-winged

males were mated to Pink Black females in order to test

the linkage of Spread. (Pink is in the third chromosome

group, and Black in the second.) The F, generation gave

only flies with red eyes, gray bodies and normal wings

(neither Spread nor Beaded). In the F, generation were

Black flies, Gray flies, and Red-eyed flies with normal

and with Spread wings, but none of the Pink-eyed flies

had Spread wings, though a large number of F, Pink

normal flies appeared. The Pink-eyed flies were also

mated inter se, but no Spread-winged flies appeared in

the F, generation. This definitely places the gene for

756 THE AMERICAN NATURALIST [Vou. XLVIII

Spread wings in the third chromosome group. Beaded

wings have not appeared in the stock bottles of Spread

which breeds perfectly true.

D. Stumpy Wines

Recently a new non-lethal sex-linked character has ap-

peared in the offspring of the cross of an F, Beaded male

to a Wild female. Its nature has not yet been worked

out, since only males have thus far appeared. The flies

have wings resembling those of Vestigial, save that they

are not held at right angles to the body, but in the normal

position. Vestigial is not a sex-linked character.

SuMMARY

The character under consideration is that of Beaded

wings in Drosophila ampelophila. All gradations of

form between that of normal wings (Fig. 1) and those

shown in Figs. 2 to 12 oceur in the stock bottles, though

certain selected strains of the stock give no normal-

winged offspring.

When a Beaded fly is mated to a fly of a stock not

carrying genes for Beadedness in its germ plasm a vary-

ing percentage of the F, offspring is Beaded. If the male

parent is Beaded the majority of the Beaded offspring

are usually females; and if the female parent is Beaded,

the majority of the Beaded offspring are usually males.

A female Beaded fly however gives a larger percentage

of Beaded daughters than does a male Beaded fly. This

phenomenon is repeated from generation to generation,

no matter whether a given Beaded fly has come from a

male or female Beaded parent, and this shows that the

phenomenon is not caused by a sex-linked gene.

This phenomenon is not caused by non-disjunction of

a sex-linked gene, for tests of both the Beaded and Wild

stocks showed non-disjunction to be a rare phenomenon.

The only explanation suggested was that the male off-

spring were somewhat influenced to or away from Beaded-

ness by the nature of the cytoplasm that was brought in

with the egg, while females were not readily influenced

in pie way. oo .

No. 576] VARIATION IN DROSOPHILA 757

A study of the F, generation shows that the majority

of the normal F, offspring differ from the majority of

the Beaded F, offspring genetically in that normals give

fewer Beaded offspring in the F, generation than do the

Beaded flies.

Beaded wings showed no linkage to any sex-linked

character.

Approximately one half of the flies of the F, generation

of a cross between Beaded flies and flies with characters

whose genes were in the second chromosome, showed

linkage in the following generation to second chromo-

some characters, while one half of the flies did not show

such linkage. The cases where linkage did not occur

gave a slightly lower percentage of Beaded offspring

than did those where linkage was present. An explana-

tion of these phenomena is sought in the suggestion that

there was in the second chromosome a gene, here called 1,

that was recessive but that in the heterozygous condition

intensified the dominance of another gene, called B’,

which was not in the second chromosome. This gene 1

behaves as a lethal factor preventing the development of

any fly that carries it in a homozygous condition.

All of the F, offspring of the crosses of Beaded flies

by flies with characters caused by genes in the third

chromosome showed linkage in the following generation

between Beaded wings and the third chromosome char-

acters. This was taken to signify that there was in the

third chromosome a non-lethal gene concerned in the

development of Beaded wings. This gene was called B’.

This gene was shown to be the essential germinal factor

in the production of Beaded wings. It is sometimes

dominant and sometimes recessive.

The determination as to whether B’ should be dominant

or recessive seems to lie in several possibilities: 1st, the

nature of the egg cytoplasm; 2d, the presence or absence

of the gene l; 3d, the nature of the environmental con-

ditions.

With reference to environmental conditions, it was

shown that a larger percentage of the F, generation had

Beaded wings when the culture was wet than when it was

758 THE AMERICAN NATURALIST [Vov. XLVIII

dry; and more when the food was alkaline than when it

was acid. No other environmental factors were discov-

ered which influenced the production of Beaded wings.

Selection of more or less extreme Beaded flies very

quickly moves the average Beadedness of the offspring

in the direction of the selection, but this selection appar-

ently becomes further ineffective in a very few genera-

tions.

Mutation is of very frequent occurrence in the Beaded

stock and the new mutants obtained have in most cases

shown themselves to be produced under the influence of

one normally Mendelizing gene.

I acknowledge with pleasure the kindly interest and

suggestions made from time to time by Dr. A. H. Sturte-

vant and Mr. H. J. Muller. These have been of much

assistance to me. My thanks are also especially due to

Dr. T. H. Morgan whose advice and criticisms at critical

points have never failed to aid in clearing up the situation.

BIBLIOGRAPHY

Baur, E. 1912. Einführung in die experimentale Vererbungslehre

siete C..B.° 1913. PE EET of the Sex Chromosomes of Dro-

i Jour. Exp. Zool., 5.

Hoge, M. A. 1914, The Talons of Temperature on the Development of

a Mendelian Character. Jour. Exp. Zool. In press

MacDowell, E. 1914. a Factors in Mendelian Inheritance.

Jour. Exp. Baat, Vol.

Morgan, T. H. 1911a. ah ore of Nine Wing Mutations in Drosophila.

Science, N. S., Vol. 33.

Morgan, T. H. 1911b. A year Sex-limited Character. Proc. Soc.

Exp. Biol. and Med., Vol.

Morgan, T. H. 1912a. The teint of a Mendelian Result a the Influ-

ence of the Environment. Proc. Soc. Exp. Biol. and Med., Vol.

gic bi H. 19126. The Rieaiaawibons of a New Sex-ratio in Dresbphita.

Science, N. S., Vol. 36.

aac 7. mi 1912. E Linkage in the Second Chromosome of the

Male. Science, N. S., Vol. 36.

Morgan, T. H. 19124. "i Modification of the Sex-ratio and of other Ratios

in grays through Linkage. Zts. ind. Abst. u. Vererb., Bd. VII.

Morgan, T. H., and Tice, S. C. 1914. The Influence of the Havivamaue

on the Size of ‘Expected Classes. Biol. Bull., Vol. 26.

Shull, G. H. 1914

Sturtevant, A, H "1913. A Third Group of Linked enag T in K Teisi tta

ampelophila. aoe N. 8., Vol. 37.

SHORTER ARTICLES AND CORRESPONDENCE

ON THE PROGRESSIVE INCREASE OF HOMOZYGOSIS

BROTHER-SISTER MATINGS

Ir has been brought to my attention that the note concerning

inbreeding, written at the request of Mr. Phineas W. Whiting to

add to his paper on ‘‘ Heredity of Bristles in the Common Green-

bottle Fly, Lucilia Sericata Meig.,’’ which appeared in the AMER-

ICAN NATURALIST for June, 1914, might be taken to mean that my

data had been sent by Dr. E. M. East to Dr. Raymond Pearl by

whom it had been published as his own. I wish to make it clear

by a statement of the facts herewith that no such interpretation

should be placed upon the note. I was seriously ill at the time

and did not submit my manuscript to Dr. Castle or Dr. East for

revision, as I should ordinarily have done. In that case no

doubt, any ambiguity of statement would have been pointed out

to me. ;

Mendel, in his original paper, showed that if equal fertility

of all plants in all generations is assumed, and, furthermore, if

every plant is always self-fertilized, then in the nth generation

the ratio of any allelomorphic pair (A,a) would be 2”—1 AA:

2 Aa:2"--laa. This statement was generalized in 1912 by East

and Hayes! for any number of allelomorphie pairs. ‘‘The prob-

able number of homozygotes and any particular class of heterozy-

gotes in any generation r is found by expanding the binomial

[1 + (2"—1)]" where n represents the number of character

pairs involved. The exponent of the first term gives the number

of heterozygous and the exponent of the second term the number

of homozygous characters.’’ A little later Jennings independ-

ently showed how homozygotes are produced from heterozygotes

by self-fertilization.?

East and Hayes? published no generalized formula for calcu-

lating the reduction toward homozygosis through any other type

of mating, but that this was thought to be a proper conclusion

deducible from the above is shown by the following quotation

(p. 21):

1U. S. Dept. Agr., Bur. Plant Ind., Bull, No. 243.

2 AMER. NAT., August, 1912.

3 Loc. cit.

759

760 THE AMERICAN NATURALIST [Vou. XLVIII

Close selection, of course, tends toward the same end (homozygosis),

but not with the rapidity or certainty of self-fertilization.

This idea is further shown by their statements under the head-

ing ‘‘Extension of Conclusions to the Animal Kingdom”’ (pp.

39-43). i

A little later Mr. Whiting had occasion to work out the results

of random matings of brothers and sisters, in connection with his

work at the Bussey Institution. He found that the amount of

heterozygosis was reduced one eighth in matings of the F, gen-

eration and from this concluded that the remaining heterozygosis

was reduced one eighth in each succeeding generation, so that in

the nth generation the number of matings which would produce

at least some heterozygous offspring would be (7/8)"*. He

showed these figures to Dr. East, who agreed with the general

conclusion (tendency toward homozygosis), but thought that the

ratio would not hold for offspring after the F, generation. Dr.

East, however, after a casual examination was not able to show

Mr. Whiting the fallacy in his work and did not go into the

matter further.

In the AMERICAN NATURALIST for October, 1913, Dr. Raymond

Pearl criticized the extension of the conclusions for self-fertilized

plants to the animal kingdom.* He applied the figures of Pear-

son, 1904,5 for random matings, which show that the relative

number of homozygotes and heterozygotes remains constant in

a population where all factors of fertility, virility and environ-

ment have the same effect upon each individual in each genera-

tion. Dr. Pearl’s error, as he has since recognized, lies in the

fact that in the F, generation random mating involves only

brothers and sisters, while in all subsequent generations it also

involves other ‘relationships.

en I read Dr. Pearl’s article in October I naturally won-

dered why there was such a difference of opinion between Dr.

East, Mr. Whiting and Dr. Pearl. Before finishing the article I

computed the amount of homozygosis in the F, generation as 0

per cent.; in F,, 50 per cent.; F, 50 per cent.; F,, 62.5 per cent.,

and F,, 68.25 per cent. As soon as possible after that I figured

other generations until the heterozygosis would be reduced to

one half of one per cent. of the maximum of heterozygosis in the

4 East and Hayes, 1912, loc. cit.

5 Phil. Trans. Roy. Soc. (A), Vol. 203, pp. 59 and 60.

No. 576] SHORTER ARTICLES AND CORRESPONDENCE 761

F, generation and found that this was accomplished in the F,

generation, the amount of heterozygosis in each generation being:

Fe, 75.000 per cent. Fs, 94.312 per cent. Fi», 98.710 per cent.

F,, 79.687 per cent. Fy, 95.398 per cent. Fa, 98.956 per cent.

Fs, 83.594 per cent. Fs 96.277 per cent. Fa, 99.155 per cent.

Fə, 86.719 per cent. Fis, 96.988 per cent. F», 99.317 per cent.

Fy, 89.258 per cent. Fin, 97.563 per cent. Fa, 99.447 per cent.

Fu, 91.309 per cent. Fs, 98.029 per cent. Fæ, 99.553 per cent.

F.» 92.969 per cent. F», 98.405 per cent. Fæ, 99.638 per cent.

With the approval of Dr. Castle and Dr. East I prepared to pub-

lish these figures.

Shortly after this Dr. Pearl wrote to Dr. East asking for an

opinion upon his article. Dr. East, in the meantime, by a method

differing from mine, had worked out the ratios independently.

Before answering Dr. Pearl’s letter, however, Dr. East compared

his results with mine. They agreed. Dr. East then wrote to Dr.

Pearl, giving a short rebuttal of Dr. Pearl’s arguments, enclosing

some of his own figures and adding that a student of Dr. Castle’s

(myself) was thinking of publishing the complete figures. Dr.

Pearl immediately acknowledged his mistake and -very gener-

_ ously asked if he should wait until I had published my article

before he published a correction. Dr. East replied that he could

see no, reason for delaying the correction and advised me of this

reply.

Since it seemed proper for Dr. Pearl to correct his previous

article, I decided to withhold my own figures and incorporate

them later in a paper bearing also upon other matters. Dr.

Pearl’s second article came out in the AMERICAN NATURALIST for

January, 1914, and this paper together with the third article in

the same journal for June, 1914, shows that his work was en-

tirely independent of Dr. East’s or my own. |

When Mr. Whiting asked me for a note giving the figures

showing what might be expected in the way of an automatic in-

erease in homozygosity when brothers were mated with sisters in

successive generations, as Mr. Whiting had done with his flies, I

naturally was pleased to have him accept my figures as correct-

ing his own, and at the same time give me an opportunity to ac-

knowledge my indebtedness to those who furnished the idea upon

which my figures were based.

H. D. Fisa

BUSSEY INSTITUTION,

FOREST HILLS, Mass.

August 18, 1914

NOTES AND LITERATURE

MENDELIAN FLUCTUATIONS?

WHEN the observed proportions, say of dominants and reces-

sives, in any Mendelian experiment are worked out for small

groups, such as individual litters or the seeds on individual plants

in individual fruits, considerable fluctuations round the expected

proportions may be observed. In the present note the magnitude

of these fluctuations is compared with the magnitude to be ex-

pected if the fluctuations were the result merely of chances of

sampling—corresponding to the fluctuations that would be ob-

served in drawing, say, samples of black balls from a bag con-

taining white and black balls in the proportion of 3 to 1. In so

far as there is good agreement, this is additional confirmation of

the Mendelian process holding good in its simplest form: if

the fluctuation observed is markedly greater than this theory

would indicate, some source of disturbance is certainly present,

but whether this disturbance arises from irregularities in the

distribution of the gametes or merely from extraneous circum-

stances (varying death-rates or difficulties of sorting) can not,

of course, be determined from the data alone. For albinos in

individual litters of mice (Darbishire’s data), and for numbers

of ‘‘green’”’ or ‘‘wrinkled’’ in Mr. Bateson and Miss Killby’s

crosses of peas I find exceedingly good agreement, at least if very

small plants are omitted. Lock’s data for maize give good agree-

ment for the DRX DR cross, but poor agreement for the

DR X RR cross. Some data given me by Miss E. R. Saunders

for seed characters in the individual fruits of stocks show rather

irregular results. Further comparisons on similar lines would

be of interest, especially for the DR X RR cross, for which very

few data are available. For the case to afford a good test the

sorting should be clear and there should be nothing in the data

to suggest differential death rates obviously.

GUY.

1¢‘¥Fluctuations of Sampling in Mendelian Ratios,’’ G. Udny Yule (Proe.

sci no dows, Soc., XVII, 425).

762

INDEX

NAMES OF CONTRIBUTORS ARE PRINTED IN SMALL CAPITALS

Age, Mean, at Marriage of Men

and Women, Variation and Cor-

relation, J.

‘Physiological

Climatic Reae

FREEMAN, 356

Atlelomorphism, Multivle, W.

CASTLE,

Allelomorphs Multiple, in Mice, T

MoRGaN, 449; and

, GLOVER M., Pattern Devel-

opm ent in Mammals aud Birds,

885, 467, 550

Analysis of a Case

ariation in sophi ag

paad of its Linkage Balarins,

N S. DEXTER

and Plent, ” sessile and

RD, 641

Ants, Gynandromorphous, described

during the a ade a Pe

WILLIAM MORTON WHE

Apple, A We of Variation | in pan

Wade: 595

Apterous Prosenik and its thie

sella Bursa-pastoris, HENRI Hus,

193

erage H. L., The Meadow Jump-

Mouse (Za apus Hudsonius)

porny regarding Hibernation,

Behavior, Genetic, An Pite a

Drosophila and its, CHARLES W.

METZ, 67.

Bessey, CHaRLES E., A Study of

Desert Figetation £

Biology of the Thysanoptera, A.

lopm

, 885, 467, 530

Br reeding, Experiments = Grass-

hoppers, Nabours’s,

TETI,

Ken

317; Alfalfa, Se

763

ARTHUR HARRIS and |

| Primula

E. | Bristles, Heredity of,

Gre

Close Li nk- |

The p at Distinction |

M c Wi

e Goa |

“Ye Climatie

F. FREEMAN,

Correlations

ctions in, GEO.

B., The Chromo-

sis of Linkage ap-

| ical

| Rea

| 356

| BRIDGES, CALVIN

| some "Hypothe

plied to Ciao in Sweet Peas and

e Com

een-bottle wry, pike

HITING,

British Islands, ° The Endemic Mam-

mals of the, T. D. A. COCKERELL,

177

| Calf, Double- ig a Osteology

of, . M.

| Capsella pe arachnoidea,

| The Origin of X, Henri Hus, 193

| CasE, E. estoration a Edapa-

osaurus crueiger Cope, 1

| CASTLE , Some New Nase

| of Rats and Gann -pigs and their

| relation to Problems of Color In-

heritance, 65; Yellow Varieties of

Rat

Multiple Allelomorphs and Close

Linkage, 503

bil-

Castle and Phillips on the Varia

| ity of Genes, The Bearing of the

| Selection Eo iai of, HEr-

. MULLER, 567

Asa C., The Effect of

| eg of Distribution on Specia- -

tion, 129 !

| Chromosome Hypothesis of Linkage

esate to Cases in Peas

rimula, CALVIN B. BRIDGES,

Citrus, Hybrids, Swingle on Varia-

tion in F,, and the ry of Zy-

Clima tic Rosetions | in Alfalfa Breed-

ing, GEO. F.

CocKERELL, T. D. A , The \demi

Mammals of the British Islands,

177

764

CoLLINs, G. N. and J. H. KEMPTON,

Inheritance ot Endosperm Tex-

a in t X Waxy Hybrids of

Mai 584.

Color p Re Some New Vari-

eties of s and Guinea-pigs,

and yas Relation to Problems of,

CooK, 2 Marad relating to Gen-

eric Types, 308

Correlation, ‘and Variation in the

Mean Age at Marriage of Men

Women, J. ae HARRIS

Roxana H. Viv 5

Correlations, Physiclogtea and Cli-

tic Reactions in Alfalfa Breed-

ae EO. F. ;

Porapon ies and Shorter Articles,

122, 177, 248, 308, 383, 446,

191, 567, 635, 693, 759

Davis, BRADLEY Moore, Stomps’s

Œnothera ep

., 498

Deri Vegetat A Study of,

sophila by a oon y of its Linkage

Relations, 712

Differential Mortality with Respect

to Seed Weight occurring in Field

RTHUR HARRIS,

aes ip in bhatar nar Galton

a GGLES GATES, 697

Distribution. The Effect of Extent

irds and Mammals,

GRINNELL, 248; A Study of Pact-

tors governing, "Punzas E. WHr

ING, 33

‘‘ Dominant’? ‘t Recessive”?

Spotting in Mice, C. C. LITTLE, 74

Drosophila, The Redupication Hy-

pothesis as applie

NT, 535; hn y E

tions, JOHN S. TER,

Duplicate venes, SEWALL WRIGHT,

638

E. M., and H.

produced

ments vith Tobacco, 5

K. HAYES, A

e Chan

THE AMERICAN NATURALIST

[Vor. XLVIII

genap of Mammals, VERNON

N KELLOGG, 257

apioa “eas lone Res-

s 117

he “Inheritance of

ing Somat e Variation in

ze, 8

OLLINS and J. H

KEMPTON, 584

Environmental Work, Humidity—a

Neglected Factor, FRANK

LUTZ,

Errors, Probable, A Short-cut in the

omputation of Certain, HOWARD

Ether, The Fa ilure of, to produce

Mutations in Dros ophila, ToB.

FERIER “705

Evolution, and Taxonomy, X, 369

IsH, H. D., On the Progressive In-

crease of Homozygosis Brother X

Sister Matings, 758

Fluctuations, Mendelian, G. UDNY

YULE, 764

Fly, tha Common Green-bottle, oa

redity ra ote tles in, PHINE

W. WHI

Formule for. the Results of Inbreed-

ing, H. S. JENNINGS, 6

FREEMAN, Physiological

Correlations and Reaz

tions in Alfalfa Breeding, 356

FROST, rt-cut in

the

paige | of Certain Prob-

able Errors, 696

GATES, R. RUGGLES, Galton and Dis-

continuity in Variation, 697

eg AE cane, Terms relating to, O.

00K

Genes, The Paria ni x5 see geting

Experiments of Cas Phil-

- on the Variability pa

ANN J. MULLER, 567; Du ists,

BEWA WRIGHT, 638

Genetic, Analysis of the Changes

produced by saat in Experi-

pret with Tobacco, ST

Genetics, aryen Terminologies

. E. Cas 83

in,

GERO EROULD, JOHN T ’ Species-building

A Hybridization and Mutation,

No. 576]

Grasshoppers, Nabours’s racer

= iments with, JoHN 8

Gregory” s ie i oAinter tera A

ess de Sei egation in,

Hark NN J. on, , 508

Gant. JOSEPH, Pania to Dis-

tribution as regards Birds and

Mammals, 248

Gataka bigs, and Rats, Some New

Varieties and their Relation to

Probl ea of Color Inheritance, W.

E. CASTLE, 6

GULICK, sous T., Isolation = Se-

lection allied in Principle,

D 1903-1913,

WILLIAM MORTON WHEELER, 49

HAGEDOORN, A. C. and A.

Another Hypothesis to account for

Dr. Swingie’s Experiments wit

Citrus, 440

HAGEDOORN, A. L., Repulsion in

Mice, 6

Harris, J ae Ugeran

eed

Correlation in the

Marrıage of Men and Women, 635

s, H. K., and E. M. East,

Genetic "Analysis of the Changes

produced by Selection in Exper

ments with Tobacco,

‘The Influence of

onarchs ode E; 265; of

Bristles in the Common reen-

bottle Fly, PHINEAS W. WHITING,

Hibernation, tig on the Meadow

Jumping Mouse, L. H. BABCOCK,

485

r Ti Brother-Sister Ma-

The Progressive Increase

Fisu, 758

Hamidiiy—a N eglected Factor in

Work, FRAN

Environ K E. ,

+ Luz,

Hus, gee The Origin of X Cap-

pec a-Pastoris Arachnoidea,

Hurcueson, T. B., Th — Years

of Wheat Selection, 4

ph ridization and "Station, Spe-

aa x Mendelian Population,

RAYMOND PEARL, 57; Notes on

INDEX

765

RAYMOND PEARL, 491; and Rela-

COLL

Internal Relations of Terrestrial i.

sociations, ARTHUR G. VES

Tsolation and Selection F in

Principle, JoHn T. GULICK, 63

JENNINGS, H. S., Formule fi the

Results ‘of Inbreeding, 693

Woods on Heredity and

horas ‘Influence of Monarchs,’’

255

KELLOGG, VERNON gerry Ectopar-

asites of Ma mmals,

MPTON, and Aa X. COLLINS,

Inheritance of Endosperm Tex-

ure in Sweet X Waxy Hybrids

of Maize, 584

Linkage, in the Silkworm Moth, A.

H. STURTEVANT, 315; and Mis-

one Terminologies in Genetics,

W. CasTLE, 383; Close, r

Continu

sophila by a Study at its, JOHN

S. DEX 711

Literature and Notes, 185, 255, 315,

50 62

’ and

< E., Humidity—a Ne-

ected F Factor in Environmen tal

ork,

Maize, The Inheritance of a Recur

ring Somatic Variation in Varie-

gated Ears of, R. A. E

766

87; Inheritance of Endosperm

Texture in Sweet X Waxy Hy-

brids of, G. N. COLLINS and J.

H. KEMPTON, 584

Mammals, The Endemie, of the Brit-

ish Islands, D. KERELL,

177; and birds, Barriers to Dis-

RINNELL, 248;

tern Development in, GLOVER

ALLEN, 385, 467, 550

Marriage of Men and W. omen, Vari-

and Correlation the

Matings, Homozygosis Brother X

pees, The Eropa Tans

-H D. Fisu, 759

Maiioi Janion Mouse (Zapus

Hudso n especially see

Hibernation, H. L. -BAB 5

en ap Population, Oa. the Re-

nbreeding a, RAYMOND

Fluctuations, G.

s. W., An Apterous

Drosophila and its Genetic Be-

havi

Mice, A ponani and ‘‘Recess-

ive,’’ Spo TTLE,

tting in, C. U.

74; Multiple Allelomorphs in, T.

H. Morean, 449; Repulsion in, A.

; RN, 699

‘í Monarchs, The Influence of,’’ and

eredity, V. L. K., 255

Morgan, T. H Multiple Allelo-

s in T

inked I 577; The Fail-

= of Ether to pebaae Mutations

, 705

Mortali ity, rabaan with respect

o Seed W t occurring in Field

Cier of Pisum sativum, J. AR-

THUR

Moth, Silkworm, ys in the,

A. L STURTEVANT,

Motile and Sessile ine and Ani- |

mals, comparison of the Re-

sponses of, VICTOR E. SHELFORD,

e of the Selecti ion ` Experiments

THE AMERICAN NATURALIST [Vou. XLVII

of Castle and rio on the

Triru: ity of Genes, 56

Alle Eier ee

Multiple, WwW. E.

CASTLE, 383; -Allel phs in

Mice

GAN W

Mutation, and Hybridization, Spe-

cies- e by, JoHN H. GER

OUL

Mutations, se in @Œnother

biennis P

Y M Dii

sophila, The Failure of Ether to

Produce, T. H. Moraan, 705

Nabours’s, Breeding Experiments

i Grasshoppers, JOHN S. DEX-

rasshoppers, W. E.

Fig 383

Notes ane Literature, 185, 255, 315,

505, 639, 762

Œnothera — L., Parallel Muta-

cares = J. pe 494;

SH Dana

Osteology “of 2 Double headed Calf,

. M. REESE, 701

Pattern Development in Mammals

and Birds, GLOVER M. ALLEN, "385,

467,

ARL, RAYMOND, On the Results of

ionship ppa aa 513

proes. and Castle,

Selec Experi im aoa

Variability 4 Genes,

HERMANN

Physiological’ ar i and Cli-

aeg etions in Alfalfa Breed-

g, GEO. F N, 356

Pisum pe EN Differential Mortal-

with Resp o Seed

eigh

po en in. Pela Onara of, J.

ARTHUR HARRIS,

Plants and Animals, Sessile and Mo-

tile, A Comparison of the Re-

E. SHELF

sponses of, VICTOR ORD,

Primulas, Tetraploid, vie bel s, A

Segregation in,

HERMANN J. MULLER, So

Rats and Guinea-pigs, Some New

arieties of, and their labian to

Problems of See ene Inheritance, W.

Fe w Varieties

of, W. E. CASTLE, 254

‘í Recessiv 6 and ‘í Dominant,’

No. 576]

Spotting in Mice, C. C. LITTLE, 74

Reduplication Hypothesis <a applie ed

to Drosophila, A. H. STURTEVANT,

aol

REE A. M., The Sage eat of a

Double- headed Calf, 701

ep Variation and Corre-

lation in Thyone, JoHN W. Scort,

Repulsion in Mice, A. L. HAGE-

s and Animals,

ciger Cope, E. C. CASE

INDEX

Scorr, Joun W., Regeneration,

enun and ’ Correlation in |

Thyone, 280

Seed Weight oceurring in Field Cul-

tures of Pisum sativum, Differen-

tial Mortality with Respect to, J.

ART s, 83

a cee A New Mode of, in

regory’s sg ye Primulas,

rane ANN J. MULL 50

Selection, in AN with To-

b Genetic rg fps of the

E and H. K.

OHN T. G

P Thirteen Years

Variability E Genes

NN J. MUL 567

Sessile and Motile Plants and Ani-

mals, A Comparison of the

ea of, VICTOR E. SHELFORD,

Sex “init and Sex-linked Inherit-

ee s

. H. Morean, 5

Peers Victor E., A Comparison

of the ” Responses ‘of Sessile and —

and Animals, 641

Computation of Cer-

obable Errors, Howarp B.

Shorter Articles and Corr rrespond-

ence, 57, 122, 177, ae 308, 383,

446, 491, 567, 635, 693, 9

LIN, ; Biology of the

LER, 129 : ;

Species-building by Hybridization

767

and Mutation, JOHN H. GEROULD,

oe in Mice, ‘‘Dominant’’

d ‘‘Recessive,’’ C. C. LITTLE,

STOMPS, THEO. J., Parallel Muta-

STURTEVANT, Linkage in the

Silkworm Moth, 815; The Re-

duplication Hypothesis as applied

to Drosophila,

Sweet Peas and Pimi Chromo-

some Hypothesis of Linkage ap-

plied to Cases in, B

BRIDGES,

Swingle on Variation in F, Cit

Hybrids and the Theory of lat

185

other Hypothesis to accou

and A, L. HAGEDOORN,

xonomy and Evolution, X, 369

A REE Associations, Internal

Relations of, ARTHUR G. VESTAL,

413

Thyone, si, Se ation, Variation -

and Correlation in, JoHN W.

Scorr, 280 me

Thysanoptera PAN Bs the

FRANKLIN SHU 236 |

Primi ra Genetic Five Ceret of the

Changes produced by Selection in

Experiments ion E. M. East

and H. K. HAYES, 5

Types, Generic, maa relating to,

O. F. Coox, 308

popie of bR The Bearing

f the Selection Experiments of

Castle and Phillips on the, HER-

uity in eer

697 ; Continuous ia OL A

Case in “som by a Study of

768

its Linkage Relations, JOHN S.

Desert, CHARLES E.

639

VESTAL, of Sata G., Internal Rela-

tions of Terrestrial Associations,

Vivian, Roxana H., and J. ARTH

riage of Men and Women, 635

Wheat Selection, Thirteen Years of,

pe 459

WHEELER, WI

LLIAM Morton, Gy-

during the Decade 1903-1913, 49

WHITE, OR E., Swing n

of Bristles in the Common Green-

THE AMERICAN NATURALIST

(Vou. XLVIII

bottle Fly, Study of Factors

Governing Distribution, 339

Woods oF Heredity and the ‘‘In-

e of Monarchs,’’ V. L. K,

WRIGHT, SEWALL, Duplicate Genes,

638

X, Taxonomy and Evolution, 369

x Capsella Bursa-pastoris loa

noidea, e Origin of, HENR

Hus, 193

Yellow Merges of Rats, W. E.

a 254

| You "Wy. J., A Study of Varia-

Hok ‘in the Apple, 595

|) gene Swingle on Variation in

F, Citrus Hybrids and the Theory

of, ORLAND E. WHITE, 185

Two Important Biological Journals

The Journal of Genetics

Edited by

W. BATESON and R. C. PUNNETT

A periodical for the publication of records of original research in heredity, variation, and allied topics,

The JOUENAL also ram from time to time, articles summarizing the PRS state of knowledge in the

various branches of genetics. The aaa are fully illustrated, and, when the subject-matter angga it.

colored plates are freely employed. The JOURNAL i s issued in parts as material rent umulates, a volume

(appear aring, so ~~ oe annually) consists ay four such parts. Sahewigiion price, $7. sa a walaiih:

single copies, $2.5

Biometrika

Edited by KARL PEARSON

A journal for the statistical study of biological problems, pps about four times a see in volume

ontaining about 500 pages, with plates and tables, is issued annually. Volume X beginsi e

nine preceding volumes form a veritable encyclopaedia of biological statistics in the fields of E n.

craniometry, medicine, heredity, etc. Subscription price, $7.50 a volume ; single copies

Prices on back numbers of these journals vary, and postage from London is charged on back volumes

and single copies

Published in America for the Cambridge University Press of England by

The University of Chicago Press

Chicago, Illinois

University Control

By J. MoKxen CATTELL, Professor of Psychology in Columbia University

Together with a series of Two Hundred and Ninety-nine Unsigned Letters by Leading Men of Science

holding Academic Positions and Articles by Josera Jastrow, Guorem T. Lapp, JOHN J. Sravenson, J. E.

Crucutron, J. McKuen CATTELL, Gmoram M. STRATTON, STEWART PATON, JOHN Jay Cuarman, James P,

Munros and Jacos Goutp SCHURMAN.

A great variety of questions pE general university administration are dealt with in an original and

heipful way.—Nature.

These quotations and examples are taken from Professor Cattell’s informed and thorough

of ergot and standing as the champion of an

rag ‘coma Date which a man enter — PERPER his self-respect.—

Sentences E a aa a a highly-trained mind, Sn a

fresh a d tell us that a new day is about to dawn in en writing.— Boston Evening Transcript.

SCIENCE AND EDUCATION

A series of volumes for the promotion of scientific research and educational progress

VOLUME I. The Foundations of Science. By H. Porvoard. Containing the authorized

by Gmorcs Brocom Hatsrep of “Science an@ Hypothesis,” “The Value of Science,” and “ Science

Method.” Price, $3.00 net.

Wuuasm H. Wats, W. H.

and

VOLUME II. Medical Research and Education. By Rıcmaro M. Pmanom,

Howeut, P. Mau, Luwentrs F. Cuantms S. Mrvor, W. B. Cannon, W. T.

Coun 4 pap Surra, G. N. Srawazt, C. M. Jackson, E. P. Lyon, Jamms B. Henner, JonN

= Cele ee ee , W. W. Kasan, Haner H. Domaup-

N, CHRISTIAN

VOLUME III. Uurali Oe Control. Now Ready. Pages z+48§. Price, $8.00 net.

GARRISON, N. Y. THE SCIENCE PRESS | LANCASTER, PA,

SUB-STATION 84, NEW YORK

The American Naturalist

A Monthly omen established in 1867, Devoted to the

ecial Reference to the Factors of Organic Evolution and Here

Advancement of the Biological Sciences

ity

CONTENTS OF THE JUNE NUMBER

See rr by Hybridization and Mutation. Pro-

or John H. Gerould.

eee of Bristles in the Common Greenbottle Fly—

A pent a rusai governing PIR ENIA Phineas

Tpi bre Aia and wwen Reactions in

Alfalfa Breedin. eo. F. F

Taxonomy and Evolution. By =

ion: Nabours’

Misleading Terminologies in Genetics. Professor

Castle.

CONTENTS OF THE JULY NUMBER

Pattern Development in Mammals and Birds.

Glover M. Allen

Internal Relations of Terrestrial Associations.

Arthur G. Vestal.

Shorter ak and Discussions Another Hy-

pothesis to Account for Dr Swingle’s Experi-

ments with Citrus, A.C. and A, L. Hagedoorn,

CONTENTS OF THE AUGUST NUMBER

a RET in Mice. Professor T. H

Tunen Years of Wheat Selection. T. B, Huteh-

Pattern Development in Mammals and Birds.

Glover M. Allen.

The iabea Jumping Mouse. Dr. H. L. Babcock.

Shorter ae and D cag t

Studies on Inb:

Dr.

Parallel Mutations i. i Agee bien L. Dr.

J. Stomps, Dr. Bradley M. e Theoret-

ical Distinction Leis EA Alema ang

and Close Linkage. Professor T. H. Morgan,

fessor W. E. — e.

Notes and Litera

opa etrics. s furion a A

nd P

of Segregation in vaa "8 Tetreplond } Prhouiase

Hermann J. Muller

CONTENTS OF THE SEPTEMBER NUMBER

Studies on Inbreeding. Dr. Raymond Pearl.

The Chromosome ig ap = pein ed to

a weet Peasand Primula. Calyin B.

Brid

The Reduplieation a as applied to Droso-

phila, Dr.

Pattern Development pa ghee and Birds. Dr.

Glover M.A

Shorter Articles oak Correspo:

The Bearing of the iene Experim ts of

Castleand Phillipson the Variability oy pan

Herm:

CONTENTS OF THE OCTOBER NUMBER

Sex-limited and Sex-lMnked Inheritance. Professor

THEM

CONTENTS OF THE NOVEMBER NUMBER

A Ta of the Responses of Sessile and Motile

Prof ‘ord.

organ

ENP EAE of Endosperm Texture in Sweet x Waxy Plants and Animals. essor Victor E. Shelf

ne tn in aes EOS usps: a An as PrE and its Genetic Behavior.

Charles W. Me

a stay ot v of Speen En oe ss W.J. Young. Shorter Articles and Discuss ie n = = u

nbreedin: essor

Variation and Coie in the Mean Age at ig g PE inhi my re Resulta ot Inbre int he Computation

aa k yira Dopisi Gaek Caran Probable Errors. Howard B. Fro

= eee : P: piae e: Galton and Discontinuity in Variation. Dr. m

Sewall Wright. Rugg The Os f a Double

Notes and Literature headed Cal À.

A Study of Desert Vegetation. Professor Charles

E. Bessey.

Single N Number 40 Cents

Yearly Subscription, $4.00

four months — UIG Unas

THE SCIENCE PRESS

Sub-Station

84:

NEW =

Lancaster, Pa.