THE

AMERICAN NATURALIST

Vou. LV. January-February, 1921 No. 636

GENETIC SEGREGATION?

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

DIRECTOR OF THE JOHN INNES HORTICULTURAL INSTITUTION

Tue later developments of Mendelian analysis have

been in the main an attempt to elucidate the scope and

nature of segregation. Mendel proved the existence of

characters determined by integral or unit factors. Their

integrity is maintained by segregation, the capacity,

namely, to separate unimpaired after combination with

their opposites. Our first aim has been to discover spe-

cifically what characters behave in this way, whether

there is any limit to the scope of segregation, or any

characters or classes of characters which are determined

by elements unable to segregate simply. The second ob-

_-ject has been to decide the time and place in the various

life-ceyeles at which segregation occurs. It is with the

latter problem that I propose to deal more particularly

in this lecture, but a brief consideration of the range of

characters, subject to segregation, is appropriate.

1. THE Score or SEGREGATION

Of the classes of features by which animals and plants

are distinguished, most have now been shown to be de-

pendent-on segregable elements. It is perhaps notice-

able that we have no quite clear proof that the factors

governing differences in number, meristic characters in

the strict sense, commonly behave so simply as those

determining other characteristics. There are examples

of parts repeated in series, such as the extra toe of the

_ 1Croonian Lecture delivered before the Royal Society of London, June

17, 1920, and printed in the Proceedings.

6 THE AMERICAN NATURALIST [Von. LV

fowl (a dominant) and the leaf of the monophyllous

strawberry with a single leaflet (a recessive) which have

a factorial inheritance, but the resulting terms, especially

the heterozygotes, are indefinite. In the polydactylous

fowl, for example, the heterozygote has a great variety

of shapes. The hind toe is most often represented by

two sub-equal digits, but the duplicity may be so slight

as to appear externally only as a division of the nail. It

may also assume another very different form, in which

there is only a single many-jointed digit representing

the usual pair. In the monophyllous strawberry the

homozygous recessive, whether before or after imme-

diate extraction from the heterozygote, shows fluctua-

tions to a 2- or 3-leaved condition.? Such fluctuations are

common among forms distinguished by meristie pecul-

iarities. There are not the uniformity and simplicity

which are such striking features of variations in color

and many other substantive characteristics. The evi-

dence as to meristic differences is, however, still scanty,

and it is too soon to decide what importance should be

attached to this preliminary impression.

With more confidence we recognize that merely quanti-

tative differences seldom, if ever, have a perfectly simple

inheritance. There are two obvious interpretations: (1)

that the factors do not usually segregate clean; (2) that

the number of factors involved is so large that their

effects are masked. Adequate discussion of these two

possibilities could only be given at considerable length.

On the whole, I incline to the former alternative, but the

material for a decision scarcely exists as yet. Certain

examples should be noticed in which, though the most ob-

vious differentiating feature is quantitative, the under-

lying physiological distinction is more probably to be

referred to a qualitative factor. Height in certain

plants is a good illustration. It is ostensibly a quantita-

tive feature, and the talls segregate clean from the

dwarfs. But in various cases, e.g., peas and sweet peas,

the dwarfs are also a darker green. The dwarf of Cam-

* Richardson, C. W., Jour. Gen., Vol. 3, p. 171 (1914),

No. 636] GENETIC SEGREGATION T

panula persicifolia, especially (about 8 inches high), is a

plant so strikingly different from the type (2-3 feet

* high), that it is sold as a species, ‘‘C. nitida.” The

leaves of the dwarf are an intense, dark green.? This

variety is, of course, a recessive and segregates without

intermediates, yet, if the qualitative distinctions were

less evident it might easily be classed as a variety in

quantity. But the critical distinction is certainly quali-

tative and the great difference in size is consequential.

Though in such cases segregation is complete, it may

almost be said to be characteristic of purely quantitative

distinctions that one or other of the original parental

types fails to reappear in its extreme form after a cross.

So usual is this feature in quantitative segregation that

the phenomenon must have special significance.

Another fact is beginning to emerge which must con-

tribute to the shaping of a conception of the physio-

logical nature of segregation. We have learnt that

groups or complexes of factors may segregate whole.

To such a complex the distinction in sex is due, but in cer-

tain cases it may break up. The occurrence of a large

spur in fertile hens (Leghorns for instance) must be re-

garded as due to the absence of that part of the sex-com-

plex which in the normal inhibits the growing of the

spur. In ordinary fowls the whole inhibiting group re-

mains on the female side, but the spur-inhibiting element

can evidently separate from the rest. On the other

hand, when in the cocks of certain breeds (as occasion-

ally in Wyandottes) there is no spur, we may perhaps

conjecture that this element has been transferred to the

male side.*

3 The ovary projects in a curious way above the sepals so much that, were

the plant: not a Campanula, botanists would describe this ovary as superior.

4 After much experiment the geneties of spur-development are still very

obscure. In the course of a long series begun by crossing spurred Leghorn

hens with a spurless Wyandotte cock (giving F, normal in both sexes),

neither form has reappeared in F, with its original: development. Very

rarely a hen with minute spurs has appeared, and occasionally a cock with

the spurs sensibly reduced. Nor from spurred hens X F,¢ has anything ap-

proaching the original types been raised. Conceivably the detached ele-

ment is able to join again with the test of the female complex.

8 THE AMERICAN NATURALIST [Vor. LV

The presence of the characteristically masculine comb

and wattles in the male Sebright which is otherwise

henny, shows that another factor similarly detachable

governs their development.

To the breaking up of large compound factors the pro-

duction of intermediate types, such as occur among the

color-varieties of plants, is in all likelihood due. The

sweet pea and the snapdragon have now an innumerable

series of such color-forms which may be represented as

having arisen by the disintegration of the various antho-

eyanins. That, at least, is the simplest way in which

their origin can be conceived.

To the final result many qualifying elements contrib-

ute, and these may naturally be separate entities. But

change in the amount of the same coloring material, and

diminution in the mere extent of its distribution, are

common features of these graduated series. As the cul-

tural development of the species progresses, more and

more of these quantitative intermediates appear, and are

selected until a practically continuous series is produced.

Although the interrelations of the whole series can be

represented by a factorial scheme, the assumption that

each of these factors had an aboriginal individuality ap-

pears. gratuitous. In Antirrhinum, for instance, the

ordinary self-colored flower segregates as a single unit

from the white. But there are *‘Delila’’ forms having

the ‘‘face’’ colored and the ‘‘throat’’ white. Another

variety has the ‘‘lip’’ colored and the peripheral parts

white and to this again there is an almost exact counter-

part in which the peripheral areas are colored and the

lip nearly white, and between these again there are fur-

ther intergrades. Apart from factors modifying its

quality, the color of the whole corolla, segregating as a

single entity from the white, would without hesitation be

represented as depending on a single factor. Subse-

quent experience that this entity can break up into an

indefinite number of fractions is not evidence that the

original representation was wrong. This reasoning ap-

plies to a great range of phenomena.

No. 636] GENETIC SEGREGATION 9

In view of the chromosome theory of linkage, it is

therefore worth remarking that we do not find linkage

especially frequent among these fractional factors.

Have they, then, been distributed among different chro-

mosomes? If in Antirrhinum, the color of the face and

of the throat were lately parts of a-single factor for the

total flower-color, would not linkage between them be ex-

pected? Nevertheless, in cases of this sort, so far as I

know, linkages have not been found with any special

frequency.

The segregation of a group of differences—and pre-

sumably factors—in combination has lately been shown

by Renner’ to occur with extraordinary frequency in the

(Enotheras, and this peculiarity is without doubt at the

bottom of the difficulties which have beset the genetic

- analysis of these plants. The complexes are in several

forms or species not borne equally by the two sexes of

the same plant, and most of them are unable to exist in

the homozygous state. These discoveries greatly eluci-

date the Œnothera problem. For instance, not only Œn.

lamarckiana, but biennis, muricata, and others also, are

not homozygous types, but heterozygotes of a special

kind. Consequently, the production by them of ‘‘mu-

tants’? is not capable of the simple interpretation

originally applied to them by de Vries. Renner sug-

gests that the mutants arise owing to some interchange

between the complexes which at present we can not in-

vestigate, but whatever be the exact manner of their

origin we can not regard them as genuine examples of

the production of novel forms by a homozygous type.

Before leaving this part of the subject, we may notice

that the supposition that segregation is concerned solely

with characters of a superficial or trivial nature has been

long ago disproved. Baur’s Antirrhinums, the study of

which was continued by Lotsy, were an excellent demon-

stration to the contrary, for they provided many illustra-

tions of segregation in features, the ‘‘specific value” of

which no systematist would question. If further evi-

5 Zts. f. ind. Abstammungs- und Vererbungslehre, XVIII, 1917, p. 121.

10 THE AMERICAN NATURALIST ` Won LV

dence were needed, it may be found in the fine series of

experiments lately published by Heribert-Nilsson® in

Salix, which, contrary to the belief long ago entertained

by Wichura, show that, when F, families are raised in

adequate numbers, very striking segregation occurs in

the species-crosses. Many geneticists are inclined to the

view that segregable characters should be pictured as

implanted on an irreducible base which is outside the

scope of segregation, but no means have yet been devised

for testing the reality of this conception.

2. The Moment of Segregation

The next question is to determine when in the various

life-cycles segregation can occur. Obviously it is a phe-

nomenon of cell-division. If we knew nothing of the

genetics of plants we might confidently adopt the view

which Morgan has so successfully developed, that nor-

mal segregation and redistribution happen exclusively

in the process of reduction. Though unconvinced, I can

not deny that linkage and crossing-over may well be

represented provisionally, as effected during synapsis.

The scheme previously offered by Punnett and myself

as a diagrammatic plan capable of representing these

phenomena is certainly far less attractive. There is evi-

dence that in certain plants, e.g., Matthiola, the composi-

tion of the families derived.from single pods shows very

great and perhaps irregular fluctuations, and the normal

ratios for those families is only found by taking the

average of many, but no sufficient statistical examination

of such cases has yet been made. In some suitable case

estimations of the offspring derived from individual

anthers might be of value in this connection. Renner,

by examining the starches of the pollen grains in

(Hnotheras, has lately made visible that dimorphism, of

which we had previously genetic proof, and perhaps this

novel and striking observation might’ be used for the

purpose of mapping the distribution of such a character

-6 Lund’s ‘‘ Universitets Arsskrift, N.F.,’’ Avd. 2, Vol. 14, November 28,

1918,

No. 636] GENETIC SEGREGATION 11

among the pollen grains. Meanwhile, it must be granted

that no indication that gametic linkage results from

somatic differentiation has yet been obtained.

When, however, we examine the view that linkage of

factors is a consequence of their association in a chromo-

some, we must observe that there is no body of evidence

that the number of linkage-systems agrees with that of

the chromosomes, a primary postulate of Morgan’s

theory. Drosophila is the only example which has been

adequately investigated. The cytological appearances

are not readily consistent with the other postulate of

Morgan’s case, that crossing-over is effected by anasto-

mosis of chromosomes and exchange of materials þe-

tween them. In our present ignorance of the physical

nature of the factors we are not obliged to assume that

an actual transference of material is an essential condi-

tion for an exchange of properties; but since Morgan’s

suggestion is made in that form we are bound to notice

how difficult it is to interpret the visible phenomena of

cytology in accordance with that hypothesis. Without

personal familiarity with cytology no one can have a

confident opinion. I observe, however, that in his most

recent publication on these subjects, E. B. Wilson? gives

a very emphatic ‘‘counsel of caution,” remarking that

writers on genetics have taken too much for granted,

and that for the present the genetic development of the

chromosome theory has far outrun the cytological. To

a layman the visible appearance of chromosomes is

scarcely suggestive of the prodigious material hetero-

geneity demanded, and the general course of cytological

evidence seems to indicate that the rôle of the chromo-

somes is passive rather than active. Though showing

features of regularity, they are capable of very wide

variations without transgressing the limits of viability,

` which could scarcely be the case were every detail in

their organization critical. The appearance of chromo-

somes is not to me suggestive of strings of beads of ex-

treme heterogeneity, but rather of strands of sòme more

7 AMER. Nat., p. 210, May-June, 1920.

12 THE AMERICAN NATURALIST [Vou. LV

or less homogeneous substance; and in so far as numer-

ical and geometrical order is exhibited by them, it would,

in my opinion, be more proper to compare this regular-

ity with that seen, for example, in drying mud or in the

formation of prisms of basalt, than to attribute to it a

more fundamental meaning.

Leaving these speculative considerations, and limiting

our inquiry to the concrete question, at what moment in

the cycle does genetic segregation occur, we reach a per-

fectly definite answer: that whatever future research

may decide as to the occurrence of segregation in ani-

mals—which, for aught we know, may always be effected

at the reduction division—there is no such limitation in

plants. We are now thoroughly familiar with a large

group of examples in which the genetic properties of the

male and female cells of the same plant are quite differ-

ent. In these, at all events, the reduction-division can

not be the moment of the segregation by which these

characters are distributed.

The first case detected was in Matthiola, where Miss

Saunders’ results proved that in the double-throwing

singles the pollen carries exclusively doubleness, the eggs

being mixed, some single and some double. A similar

condition was shown to exist in regard to the cream and

white plastids, respectively, the pollen grains bearing ex-

clusively cream. De Vries observed a comparable ar-

rangement among the Œnotheras, and Renner has lately

shown that the phenomenon is widely spread in that

group, thereby providing a consistent interpretation of

much that was formerly obscure in the genetic behavior

of these plants. In Campanula carpatica Miss Pellew

proved that the pollen grains of the hermaphrodite form

called pelviformis carry exclusively femaleness, and

preponderantly white flower-color (the plant being

heterozygous for blue). The ease of Petunia investi-

gated by Miss Saunders* is somewhat peculiar in

the fact that in the heterozygous singles the male side

carries tle dominant singleness only, since in those in-

_ § Jour. Gen., Vol. 1, p. 57 (1910).

No. 636] GENETIC SEGREGATION 13

stances to which the conception of dominance can be ap-

plied, it is the male which commonly carries the re-

cessive. Segregation of these characters can not in

plants so organized be supposed to take place later than

the constitution of the male and female organs, and

therefore the reduction division can not be the one crit-

ical moment. The suggestion has been made that germ-

cells of the missing kinds may be in reality formed and

perish before reaching a functional stage. As regards

the Cinotheras, where shriveled pollen grains abound,

this conjecture is very plausible and probably correct;

but when, as in the other cases here quoted, the pollen

grains are uniformly sound, the hypothesis is inap-

plicable and without evidential support. Moreover, even

if it were true that certain classes of germ-cells perish in

one or other of the sexes, that would hardly alleviate

the difficulty, for this differential viability would remain

to be accounted for, nai itself a phenomenon of segre-

gation.

Begonia Davisii? is another curious illustration in

which segregation must occur even earlier. This plant

is a wild, true-breeding species, with ordinary single

flowers. All the pollen grains, however, carry double-

ness, and used on the female flowers of doubles give off-

spring all double (single being the dominant). The

pollen of this plant is as uniform and perfect as that of

any species I have ever seen. We must therefore con-

clude that the segregation by which singleness separates

from doubleness is effected not later than the formation

of the rudiments of the male and female flowers. Cyto-

. logical investigation may no doubt show that the distinc-

tions between the genetic properties of the male and

female are associated with visible nuclear differences,

but I see no reason for anticipating that such differences

must exist. Cells which differ in their genetic poten-

tialities must of course differ in physical constitution,

but there is no reason to suppose that this difference

need be in any way dependent on chromosome structure.

9 Jour. Gen., VIII, 1919, p. 199.

14 THE AMERICAN NATURALIST (Vou. LV

As regards Campanula carpatica ‘‘pelviformis’’ and

Begonia Davisii, experiment has shown that the peculiar

kind of segregation which they exhibit may recur in

their offspring. In the Begonia, if the female of Davisi

be fertilized with pollen of an ordinary double tuberous

Begonia, the doubleness so introduced stays on the male

side just as the doubleness of its own male does, and a

plant so bred has its pollen all double. But if the male

of Davisii be used on the female of an ordinary single,

there is no restriction of doubleness to either sex of the

offspring. The peculiarity of Davisii must therefore be

attributed to the special properties of its female side.

The Campanula case is complex and has not yet been

fully explored, but at least from the female side of pelvi-

formis plants have been raised which retain the proper-

ties of the mother as regards the distribution of the

white and blue colors.

We have at the John Innes Institution been lately in-

vestigating a simliar ease in flax, which; though compar-

able, has some special features. A dwarf flax (Linum

usitatissimum) of unknown origin, presumably a stray

seedling of one of the varieties grown for oil, was fer-

tilized with pollen from our tall fiber strain. Both par-

ents breed true to the fully hermaphrodite condition, with

anthers perfectly formed, and the F, plants were normal

in this respect. F, consisted of hermaphrodites, and a `

recessive form with aborted anthers, generally conta-

bescent and not dehiscing at all, but having occasionally

a few grains of good pollen. The ratio was a normal 3: 1.

The recessive, having occasional grains of pollen, self-

fertilized, gave similar plants with anthers wholly or.

almost wholly aborted. The normal F, hermaphrodites

gave in F, families which showed that some of the F,

plants were pure normals, others heterozygous in the

ordinary way. But when the recessives were fertilized

with pollen from three several varieties of tall fiber flax,

only recessives were produced. These tall flaxes there-

fore are normally heterozygous for the recessive or ‘‘sub-

female’’ condition, and this in segregation is perman-

No. 636] GENETIC SEGREGATION 15

ently relegated to the male side of the plant, while the-

female side takes the hermaphrodite factor. Segrega-

tion in regard to the same recessive may take place in

one of two ways. It may be unilateral, as it is when in

heterozygous association with the female of the tall

flaxes, or it may be ambilateral and unrestricted to

either sex when it is in association with the female of

the oil flax. We must infer that the female halves of

the two types differ in some critical respect which decides

the manner of the segregation.

Unilaterality may also show itself as a difference in

the closeness of linkage on the two sex-sides of the same

plant, and no doubt this fact may have a bearing on the

interpretation of the foregoing cases. The late R. P.

Gregory discovered the first case of this, in Primula

sinensis. He found that the linkage between magenta

color and short style was closer in the eggs than in the

pollen. Recent work on a larger scale has given 10.9: 1

as the female linkage and 6.4:1 for the pollen. A sim-

ilar difference has been also found for the linkage be-

tween green stigma and ‘‘reddish’’ stem (as opposed

to dark red), the value being 29.8:1 for the eggs and

41.7:1 for the pollen. In both examples, individual

families show wide fluctuations, and these values should

for the present be regarded as approximate only. What-

ever be their meaning, they show that some segregation

has occurred in the formation of the two sets of sexual

organs, such that the process of gametic differentiation

is not the same in both.

Besides these examples of differentiation between the

male and female sides, there are others proving that

segregation may occur at other stages in somatic devel-

opment. The most obvious examples are the variegated

plants. I have discussed this subject elsewhere in con-

nection with reversible periclinal ‘‘chimeras’’ of white

over green which produce shoots having the white in-

closed with the green.° To these must now be added

the cases in which the plants arising from adventitious

10 Jour. Gen., Vol. 7, p. 93 (1919).

16 THE AMERICAN NATURALIST [Vou. LV

buds differ from the plants which produce them. I have

described one of these examples in Bouvardia. The

pinkish white ‘‘Bridesmaid’’ gives the red-flowered

‘‘Hogarth’’ from its root-cuttings. Three similar oc-

currences have been found in fancy Pelargoniums. The

root-cuttings of a white-flowered variety, ‘‘Pearl,’’ give

a red-flowered form very like ‘‘Mme. Thibaut.” ‘‘Mrs.

Gordon,’’ which is a full rose-pink, with whitish edges,

gives from its root-cuttings flowers in which the two

posterior petals are marked with dark red, not unlike the

variety ‘‘Cardiff.’’ A more striking case is that of

“Escot,” which gives from its roots plants with bright

pinkish red marks, those of the original being purplish

red. The most curious feature of this case les in the

increased size both of the plant and the flowers coming

from the roots, and it is scarcely possible to see the

petals of ‘‘Hscot,’’ which are characteristically rolled

back, side by side with those of the root-form, which are

not only larger, but also flat, without surmising that this

roiling back is an expression of the greater size of the

larger petal contained within the smaller, causing a want

of correlation between the growth of the inner and outer

tissues.

Buckling or crumpling of leaves through want of cor-

relation was a conspicuous feature of some of Winkler’s

‘‘oraft-hybrids,’? made from Solanum nigrum and S.

lycopersicum, when the larger tomato was inclosed

within the smaller species. We have had a precisely

similar example in a salmon-fringed Pelargonium bred

by Mr. Jarman, of Chard. The leaves are obviously

buckled, the petals are laciniated, and the female parts

aborted, though the anthers are perfect. This male and

deformed flower is proper to the outer tissues only; for

on two occasions the plants have produced shoots with

large flat leaves and normal hermaphrodite flowers with

their petals entire. Obviously, this normal plant was

inclosed within a skin of the fringed type.

In all these examples, a somatic segregation has oc-

curred which determines the genetic potentialities. The

No. 636] GENETIC SEGREGATION 17

interpretation that they are periclinal chimeras is prob-

ably correct for the most part. The fringed Pelargonium

is obviously of this nature. Nevertheless, the fact that

a root-cutting consistently produces a certain type of

, plant which is not the original does not prove that the

distribution is periclinal. Another possibility is well

illustrated by the case of a variegated Spiræa ulmifolia,

having the stem, petioles, and (basal) centers of the

leaves without chlorophyll... The growing point has

the power of laying down green tissue in the lateral

areas only, the internodal regions being albinotic. Root-

cuttings from this form give albino plants which die

after the development of two or three small leaves. Now

in this case we can see the distribution of the green and

white, respectively, and we recognize that the roots give

albino plants because they belong wholly to the albinotic

area. On similar lines it is possible to interpret the

Bouvardia and other cases. The distribution of the two

types in the same plant may be such that one is limited

to the root and internodes, while the other is in the nodal

structures.

That considerations of this kind are not fantastical is

proved by the genetical phenomena seen in the case of

‘‘rogues’’ in culinary peas, which Miss Pellew and I

have been investigating for a number of years.'* The

rogue is a peculiar, wild-looking plant, differing in

various ways from the type, chiefly in having pointed

leaflets. Crosses between it and the type give plants

which in their lower parts are intermediate, though turn-

ing into rogues as they develop. The self-fertilized off-

spring of rogues and also of these F, plants are always

rogues, and evidently the type-characters introduced

11 This is somewhat like the Pelargonium named by Messrs. Cannell

‘t Freak of Nature,’’ in which the chlorophyll has a closely similar distri- —

bution, and it, like the Spiræa, is sterile on both male and female sides. In

this Spiræa I have never seen pollen, but very rarely a fruit is formed, which,

no doubt, is due to an occasional development of a bud in the green area,

an occurrence frequent in variegated plants. Whether these fruits contain

viable seeds is not yet known,

12 Roy, Soc. Proc., B, Vol. 91, p. 186 (1920).

18 THE AMERICAN NATURALIST [Vor. LV

from the type parent are left behind in the lower parts.

Such a case may perhaps be compared with the condition

seen in the variegated Spirea, and we may fairly con-

jecture that if it were possible to raise root-cuttings

from the F, peas, they would produce types.

A more gradual exclusion of the type-elements in the

lower parts is seen in certain intermediates. Thesé may

scarcely differ from types in the lower parts, though

changing to rogues, sometimes abruptly, sometimes

gradually, as the series of flowering nodes is developed.

Reciprocal crosses between the successive flowers of such

plants and the flowers of types has shown that, together

with the gradational change in the somatic structures,

there is also a gradational change in the proportion of

gametes bearing the rogue and type characters respec-

tively. This proportion and the rapidity of the change

differ on the male and female sides. Of the egg cells

in the lower flowers, up to about the tenth flowering node,

rather more than 50 per cent. carry the type-characters

—or at least the non-pointed leaflets—but above this

level the proportion of types declines. Of the pollen in

the lowest two flowers only about 20 per cent. is type-

bearing and the proportion diminishes rapidly in each

successive flower above the level.

In all the examples given hitherto the segregation is

in diploid tissues, but a comparable phenomenon has

been proved by Collins to occur in the haploid axis of a

moss (Funaria). In a dicecious moss, as the Marchals

have shown, sex-segregation occurs at spore-formation,

the deviation in which reduction is effected. This, of

course, agrees with cytological expectation, though, so

far as I know, the details have not been observed. But

from the leaves of mosses placed in nutrient fluids new

plants may be raised without great difficulty, and Collins

found that the (perigonial) leaves surrounding the male

organ thus propagated, produce exclusively male axes?

He has since raised similar cultures from the (peri-

13 Jour. Gen., Vol. 8, p. 145 (1918-19).

No. 636] GENETIC SEGREGATION 19

chetial) leaves surrounding the female organ, and, as

related in his recent paper (Genetics of Sex in Funaria

hygrometer), from them monecious plants resulted. The

proof is thus complete that in a haploid tissue a segre-

gation of sex can occur.

The inference may be drawn that the factors for other

characters may similarly be liable to segregate in the

haploid state. In this connection I may mention a case

which, though as yet obscure, perhaps fulfils this expecta-

tion. In botanic gardens a variegated maidenhair fern

(Adiantum capillus-Veneris) is grown which has wedges

of white tissue irregularly distributed in the segments.

This plant produces spores freely,’* and these give rise

to prothallia which in several cultures raised here have

always been entirely green. But when ferns arise from

these green prothallia by the sexual process they are of

three kinds, green, white, and variegated like the parent

plants. The fact that the prothallia should be all green

is entirely unexpected and creates a distinct problem,

but it is evident that segregation must occur either in

some of the cell-divisions by which the prothallia pro-

liferate, or in those by which the gametes are formed; in

either case in haploid tissue. This segregation is essen-

tially different from that by which the differentiation of

organs, such as the archegonia and antheridia, is accom-

plished, inasmuch as it relates to elements determining

the characters of the next generation.

From the evidence given it is clear that in a wide view

of living things segregation can not be exclusively a

property of the reduction-division, and for the present

it should be regarded as a possibility which may occur

at any division in the life-cycle.

14 T have not satisfied myself that spores are produced in sori on the white

areas. i

FACTORS IN THE RESISTANCE OF GUINEA PIGS

TO TUBERCULOSIS, WITH ESPECIAL

REGARD TO INBREEDING AND

HEREDITY

DR. SEWALL WRIGHT, 8.D., AND DR. PAUL A. LEWIS, M.D.

Bureau OF ANIMAL INDUSTRY, U. S. DEPARTMENT OF AGRICULTURE,

Wasuinaton, D. C., AnD THE Henry Puirps Institute, THE

UNIVERSITY OF PENNSYLVANIA, PHILADELPHIA, PA.

In the pre-bacteriologic era physicians were quite gen-

erally of the opinion that heredity played a large part in

resistance or susceptibility to tuberculosis. Following

Koch’s demonstration of the part played by the Bacillus

tuberculosis in the etiology of the disease it was recog-

nized that the hereditary influences might be only ap-

parent; the disease being established in any family there

was evidently an immeasurably increased chance of inter-

familial infection; and as a consequence of this uncer-

tainty of interpretation it has become quite customary to

regard hereditary influences, properly considered, as a

negligible factor.

Pearson’? and Goring? compared the correlation be-

tween parent and offspring, in incidence of the disease,

with that between husband and wife. The effects of un-

favorable conditions and the chances of infection might

be about as great in one case as in the other, while

heredity would be a common factor only in the first case.

Pearson dealt with upper class families of which one

member was being treated in a certain sanitarium. The

correlation between parent and offspring came out about

.50, in the usual scale, in which 1.00 is perfect correla-

1 Pearson, Karl, 1907, ‘‘A First Study of the Statisties of Pulmonary

Tabereulosis, ’? Dulau & Co., London. 26 pages.

2 Pearson, Karl, 1912, ii fdberoalosik, Heredity and Environment.’’

Dulau & Co., London. ges

3 Goring, Chiles, 1909, ‘*On the Inheritance of the Diatheses of Phthisis

and Insanity.’’ Dulau & Co., London. 28 pages

No. 636] GUINEA PIGS AND TUBERCULOSIS 21

tion, after making certain assumptions in regard to the

frequency of the disease in the whole population, and

making allowance for the age of the children, many of

whom had not reached the age of greatest danger. This

correlation is about the same as that between parent and

offspring in characters such as height, weight, eye color,

etc., which are largely or wholly determined by heredity.

The correlation between husband and wife was only .24

and even much of this, the author found reason for at-

tributing to assortative marriage.

Goring’s study was made with the families of British

convicts. He also found a correlation of about .50 be-

tween parent and offspring but no correlation (— .01)

between husband and wife. The only assumption that

had to be made in this case was a correction for the age

of the children.

We are acquainted with no experimental work bearing

on this question.

MATERIAL

The present paper describes the results of experiments

on certain inbred and crossbred stocks of guinea pigs.

All of the animals used were a by-product of experiments

on the effects of inbreeding which have been carried on

by the Bureau of Animal Industry of the U. S. Depart-

ment of Agriculture since 1906. One of the authors

(Wright) has been in charge of these experiments since

1915. A detailed report on the results is soon to be pub-

lished. A brief summary will suffice here.

Twenty-three families of guinea pigs were maintained

for a number of years by mating exclusively brother

with sister among the descendants of twenty-three orig-

inal pairs. In sixteen cases (including families 2 and 13

of those tested for resistance to tuberculosis) both of the

original parents came from a stock which had already

been maintained for twelve years without the infusion of

fresh blood, by the Experiment Station of the Bureau of

Animal Industry. In the remaining families, including

families 32, 35 and 39, the original females came from

22 THE AMERICAN NATURALIST (Vor LV

the above mentioned stock, while the original males were

purchased from a local dealer.

In 1911 a number of animals were selected from the

stock of the Experiment Station to start a control ex-

periment. The mating of even second cousins has been

avoided in this stock, which has been called Experiment B.

Eighteen of the inbred families were still on hand in

1916, having then on the average about eleven genera-

tions of brother-sister matings back of them. At this

time most of the families were disposed of in order to

make room for cross-breeding experiments and to obtain

larger numbers from the five families which it was de-

cided to retain. These families, numbers 2, 13, 32, 35

and 39, are the ones which have been tested for resist-

ance to tuberculosis. They were retained in part because

they already occupied many pens, but largely because of

the possession of contrasting characteristics in size, fer-

tility and coat pattern. Each of these families is at

present very homogeneous in heredity. Family 2 is de-

scended wholly from one mating in the sixth generation —

of inbreeding. Families 13, 32, 35 and 39 come from

single matings in the seventh, eleventh, twelfth and

eighth generations respectively.

The total number of animals involved in the experi-

ment on inbreeding and cross-breeding has been about

30,000.

Tue EFFECTS OF INBREEDING ON VIGOR

It is noteworthy that there has been no very obvious

decline in vigor, although the families are now on the

average in the fourteenth generation of brother-sister

mating and one of the most vigorous (No. 35) has

reached the twenty-first generation.

There has, however, been some decline in vigor in all

respects which have been studied. The decline is most

marked in fertility, including both frequency and size of

litters. It has been so great in this respect that it

would have to be recognized even though the decline in

No. 636] GUINEA PIGS AND TUBERCULOSIS 3 23

other respects were assumed to be due wholly to less

favorable environmental conditions.

That there has been a real genetic decline in the inbred

stock in all elements of vigor is shown by comparison

with the control stock B, which has been superior in every

respect. Still better evidence has been obtained by com-

parison of the inbreds with the young from crosses þe-

tween the different families raised at the same time and

under the same conditions.

THE EFFECTS OF CROSSING

In interpreting the effects of crossing, the character-

istics which depend on the hereditary make-up of the

young must be distinguished from those which depend on

the dam or sire. In studying these questions, inbred

females have been mated with inbred males of another

family (experiment CO) and with crossbred males (ex-

periment CA). Crossbred females have been mated with

brothers (experiment CI), unrelated crossbred males

(experiment CC) and inbred males of an unrelated fam-

ily (experiment AC).

Size of litter appears to depend wholly on the dam.

There is little or no improvement in the experiments in

which the dam is inbred (CO, CA). There is, however,

a marked increase, 10-30 per cent. depending on condi-

tions, in the litters produced by crossbred females (CI,

CC, AC). 3

The record of an experiment in frequency of litters

depends on the age of maturity as well as on the regu-

larity thereafter. Males mature considerably later than

females, so that the age of maturity of the male is the

controlling factor in this respect in matings, made as in

the present experiments, between immature animals.

The frequency of litters after maturity appears to de-

pend largely on the dam. There is no improvement in

the record of the first cross (CO) over the inbreds. There

is, however, marked improvement in the other experi-

ments in which either the sire or dam or both are cross-

24 THE AMERICAN NATURALIST (Vou: LV

bred. Crossbred males mature earlier and the cross-

bred females have litters more regularly than inbreds.

The percentage of the young born alive depends almost

wholly on the dam. There is little or no improvement in

experiments CO and CA, but an increase of 6 to 8 per

cent. where the dam is crossbred. The percentage which

are raised to 33 days of age of the young born alive de-

pends both on the dam and on the heredity of the young.

There is a marked increase, 9 to 12 per cent., in all of the

crossbreeding experiments mentioned above.

Somewhat similarly, birthweight depends largely on

the dam, while the gains between birth and 33 days de-

pend to a considerable extent, though far from wholly,

on the heredity of the young. Guinea pigs become inde-

pendent of the dam at a very early age. There is an in-

crease of 2 or 3 per cent. in experiments CO and CA, but

one of about 7 to 10 per cent. where the dam as well as

the young are crossbred. In the gain between birth and

33 days, there is an improvement of about 16 per cent. in

the first cross, which is somewhat increased in the young

produced by crossbred dams. There is an increase of

15 to 20 per cent. in the adult weight in the first cross

(CO). This is maintained in the young produced by

matings of unrelated crossbreds (CC), but half of it is

lost where the parents, though crossbred, are brother

and sister (CI). The influence of the dam does not ap-

pear to extend to the adult weight.

A loss in the improvement brought about by vey cross-

ing becomes apparent in the second generation of in-

breeding following a cross (experiment C2) in those

cases in which it is not apparent in the first generation.

COMPARISON OF DIFFERENT FAMILIES

A comparison of the different inbred families with

each other has revealed persistent differences in color,

pattern, tendency toward polydactylism, tendency toward

production of monsters, mortality among the young,

weight and both elements of fertility. It was found that

No. 636] GUINEA PIGS AND TUBERCULOSIS 25

the differences in these respects could not be interpreted

merely as differences in general vigor. Vigor above the

average in one respect was as likely as not to be found

associated with a subnormal record in another respect,

the correlation between the records of the families in two

respects coming out in most cases substantially zero.

Examples of these family differences will be brought out

later in connection with the effects of inoculation with

tuberculosis.

EXPLANATION OF THE RESULTS OF INBREEDING AND CROSSING

These results harmonize well, on the whole, with those

found by other investigators. It is believed that they can

be explained as consequences of the current theory of

heredity without recourse to the rather mystical ideals

which once prevailed in regard to inbreeding. There

appear to be independently inherited factors which af-

fect frequency and size of litter, ability to bear the young

successfully, vitality and growth as well as for color,

pattern and the other characters in which the families

differ. There seem to be surprisingly few factors which

act on all of these characteristics. The concept, hered-

itary vigor, thus becomes merely an expression for the

sum of a number of independently inherited qualities

and not an entity.

The factors which cause reduced vigor in any respect

appear to be in general recessive. The primary effect

of inbreeding is to render homozygous a random group

of the factors present in the original stock. Some combi-

nation of factors, good, bad and indifferent, thus be-

comes fixed in each inbred line. As the recessive factors,

tending toward lack of vigor, are as likely to become

fixed as the dominant ones there is on the average a de-

cline in vigor in each respect. Moreover, owing to the

likelihood that many factors for vigor will be linked

genetically with factors causing weakness, it is to be ex-

pected that vigor in all respects will be found in very few

lines, even where there is careful selection. In the pres-

26 THE AMERICAN NATURALIST [ Vou. LV

ent case, there was very little conscious selection but a

considerable amount of natural selection was, of course,

unavoidable.

On crossing two inbred lines, each, as a rule, supplies

some of the dominant factors lacking in the other, with

the consequence that there is increased vigor in so far as

the character in question depends on the heredity of the

animal itself. In the next generation if brother-sister

matings are made, there should be a decline as compared

with the first generation in characters which depend

wholly on the animal itself. The decline from this cause

may, however, be balanced or more than balanced by the

improvement due to the influence of the crossbred dam.

EXPERIMENTS ON Resistance To TUBERCULOSIS

Since April, 1919, all surplus individuals in five inbred

families (2, 13, 32; 35 and 39) and a random selection

from certain crossbreeding experiments (CO, CA, AC,

and B) have been shipped from Washington to Phila-

delphia to be tested for resistance to tuberculosis. The

animals were shipped in lots of 30 to 60 every two or

three weeks. They were from 5 to 8 weeks old at the

time of shipping. They usually arrived in good condi-

tion but with some loss in weight. The first two lots

were not tested because of heavy mortality before inocu-

lation due to feed conditions. Lot 6 was not tested be-

cause of heavy mortality due to excessive heat during

shipment. Data on the shipment and inoculation are pre-

sented in Table I. The deaths within 15 days after in-

oculation are probably not attributable to tuberculosis and

are not considered in the later work. The experiments

were considered closed when all but two or three of the

animals had died. The last column gives the number of

days after inoculation at this time. The next to the last

column gives the day by which 50 per cent. of those which

passed 15 days had died. It agrees fairly well with the

average number of days calculated on the assumption

that the few survivors of the experiment died on the day

following that given in the last column.

No. 636] GUINEA PIGS AND TUBERCULOSIS

The cultures used were of the human type of tuber-

culosis with the exception of lot 3 in which the type was

bovine.

TABLE I

g=] | | i=)

I 42/73 pes «a

È E EIIE $

8 a fs Inoculation (E8 | Sa ef ag

8 A= | ire - >a

ee sg| | 22/3 | 2°

Z8 z Ag

3 |May 4 June 2, 1919 1/10 mg. culture 07.8 intraperi-| 2 | 32 43.9 44 | 62

| toneal |

4 June 4 Ee W Ro a 465 és 5 | 30-| 25.9) 21 0

5 , uly 3, 1/10 ys 465 S 73 874340 31

7 \July 17 Aug. 25 “ ‘1/5 DG n 3 | 37 256 24 | 43

8 oe | é é t ae s: 1 28 25.2 24 38

9 |Aug. 14| “ AETA +t ahd Au 3 | 44 |25.3| 24 | 39

Op 26 Oot. 2 oT gs “* subcuta- | 6 | 46 | 49.7) 47 | 84

` | | us |

EE peptic. oo | p s T g 3 | 31 | 54.2) 54 | 83

T2 oe 25.. ék ék ae oe éé ae oe > 30 49.5 46 83

13 (Oct. 9 eed Boh AE A as S h 3 | 42 |45.2| 40 | 83

ee oS A al a ries Moa ay = 2 | 35 | 42.3! 39 | 83

Lots 7, 8 and 9 were given the same inoculation on the

same day. The results are so similar, in spite of the

differences in the age and weight of the animals, that

they can be dealt with satisfactorily as one experiment.

Lots 10 to 14 were all given the same inoculation on Oc-

tober 24 and 25 and can also be dealt with as one experi-

ment. The same quantity of the culture, inoculated

intraperitoneally in lots 7 to 9, was inoculated subeu-

taneously in lots 10 to 14, with the consequence that the

average length of life was much greater in the latter

group of lots. Lots 3, 4, and 5 were inoculated separately

with different cultures or different amounts of the same

culture. The numbers in each case are rather small. for

separate treatment. The mortality curves for lots 4 and

5 are not, however, very different. That for lot 3 can be

made similar by multiplying the days of survival after

inoculation by 3/5. These three lots have been combined

in this way for comparison with the later and more satis-

factory experiments.

28 THE AMERICAN NATURALIST [Vou. LV

Sex as A Factor IN RESISTANCE

In attempting to analyze the difference in the length

of life following inoculation, we will consider first a num-

ber of possible factors other than heredity. It will be

convenient to start with sex. Chart 1 shows the per-

|e? 12 s_|se las lve Pa . ro e br

maet DAYS Liven

ro N

‘ TS

š he

vs BA

> w

pF aS

DA \ my

à \ M

es Jt

a he

2 = cai i

Se A

>. ST isaac an

Bea ie &

2 Sex L 10-14. Ni iea AN

ae! SR Pe

° HART |/

. The mortality curves of the males and females of lots 10 to 14.

This and the later charts give the percentage alive at the beginning of each

five-day period starting with 100 per cent. on the fifteenth day after inoculation,

centage of the males and females alive at the beginning

of each day in lots 10 to 14, starting with 100 per cent. on

the fifteenth day after inoculation. The males have some

advantage in lots 10-14, but it is too small to be of more

than doubtful significance. In lots 3 to 9, moreover, the

two mortality curves keep crossing each other in such a

way as to indicate that sex makes no appreciable differ-

ence in resistance.

Tue AGE or THE Dam as a Factor

Some rather dubious indications of a relation between

age of mother and susceptibility of the offspring to tuber-

culosis were found by Pearson among human beings.

No. 636] GUINEA PIGS AND TUBERCULOSIS 29

The guinea pigs in lots 10-14 were divided into two

groups, those containing at least 25 per cent. of the blood

of family 35 and the remainder. Each of these groups

was classified by birth rank with results given in Table II.

TABLE II

AVERAGE NUMBER or Days SURVIVAL AFTER INOCULATION AMONG THE YOUNG

IN DIFFERENT BrrtH Rangs, Lots 10-14

; Animals with Blood of Family Miscellaneous

Order of Litter

No, Av. No. Av.

De ee eye 16 58.9 17 35.6

r E oes 21 60.3 29 43.1

Dae oe ee T 7 70.6 23 43.3

Bee ee ae 12 63.3 26 39.8

Pic T a be 4 8 6 35.8

Ee Sie areas ae il 45.9

E aaa a SB | 5 371

ae E 3 33.3

a a Glee oie 4 30.8

These figures suggest an increase in resistance of the

young born in litters up to the third and a decline in re-

sistance in young born of aged dams. They must be con-

firmed by larger numbers, however, before much signifi-

cance can be attached to them.

THe RELATION or AGE, WEIGHT AND RATE or Garn TO

RESISTANCE

Mere bulk must be considered as likely to be a factor

in resistance to tuberculosis. Other things being equal,

one would expect to find that the same inoculation given

to a St. Bernard and to a toy spaniel would have a more

disastrous effect on the latter. In lots 10-14, some of

the guinea pigs were over three times as heavy as others

at the time of inoculation (variations between 120 and

440 grams).

The rate of gain is a good indication of the condition

of the animals and is thus another factor which one would

expect to find correlated with resistance to disease. In

this connection, we have for study the birth-weight,

weight at 33 days and the rate of gain between 33 days

30 THE AMERICAN NATURALIST [ Vou. LV

and inoculation. The birthweight varies between 40 and

120 grams, the weight at 33 days between 90 and 360

grams, and as to rate of gain after 33 days, some lost

weight while others gained over 3 grams a day.

ge is a factor which must be considered apart from

weight and rate of gain. In lots 10-14, there were varia-

109 c s e af yo Cis o lea 2 7e F Szi 3s

Ng Days Liv

NS, Ye

a? NSA

+ ANS k;

x AN Wesker AT Tw Ty: revere Days

N N LOTS 10-44.

S N c

23 S.J

en NIN

Q =

= AN kes

s 204/50 6. has N :

= YONE

3 W

$a LSO-2ICRAMS —-— eS Ee S

re pS Bee?

2 210-2708 RAMS mmm at has T

N BE TAURS Bee

2 ovek 210 GRAMS a ti

G umaa eie, T T

. CHARTS.

CHART 2. The mortality curves of the guinea-pigs of lots 10-14, grouped ac-

rding to weight at weaning.

tions between 35 and 105 days at the time of inoculation.

In lots 7-9 the variation was between 45 and 85 days.

The correlation between the length of life after inocu-

lation and each of these factors, including also size of lit-

ter, is given in Table III for lots 7-9 and lots 10-14. All

of these correlations were calculated by the usual product

TABLE III

tors Correlated Lots 7, 8, 9 Lots 10-14

Days ved Size OF Witter: vei es + .090 + .064 + .029 + .050

weight fe ess — .044 + .064 + .059 + .050

aa at 33 days...... + .022 + .065 + .056 + .050

—Rate of gain age )-- + 120 + .064 + .223 + .047

—Age at inoculation...... + .025 + .065 + .180 + .048

—Weight at biota ag ..- + -092 + .064 + .219 + .047

No. 636] GUINEA PIGS AND TUBERCULOSIS 31

moment method. Perfect positive correlation is +1,

perfect negative correlation is —1, while 0 indicates the

absence of correlation.

Inspection of these figures reveals that all of them are

surprisingly low. In lots 7-9, the factor most closely

correlated with length of life is the rate of gain between

weaning (33 days) and inoculation, and even this is not

certainly significant in view of its probable error (+ .120

+ .064). Larger correlations are to be expected in lots

10-14, owing to the greater heterogeneity in age and

weight, but even here the correlations are all small. Size

of litter, birthweight and weight at 33 days are again of .

no value as indicators of resistance to tuberculosis.

There, are, however, significant correlations between

length of life and rate of gain (+.223+.047), age

(+ .180 + .048) and weight of inoculation (+ .219 + .047).

The degree to which variation is determined by a given

factor is measured by the square of the coefficient of cor-

relation. On this basis, only 5 per cent. of the variation

in length of life (i.e., 5 per cent. of its mean square de-

viation) is determined by the most important of the above

factors. The degree of determination by all of these

factors combined can not be found by merely adding the

squares of the correlations, because the factors are not

independent of each other. For example, rate of gain is

an element in determining the weight at inoculation. Any

effect which it has on resistance should be reflected in a

correlation between weight and length of life as well as

between rate of gain and length of life. The correlations

among the more important of these factors are given in

Table IV.

TABLE IV

Lots 7, 8, 9 Lots 10-1

E at inoculation—Rate of gain..... + .357 + .056 + .553 + .035

Te rere rors + .420 + .054 + .742 + .022

_weight at 33 days. + .780 + .025 + .670 + .027

Rate of gain MO Sy ae ee + .323 + .058 + .592 + .032

—Weight at 33 Peo — .153 + .063 — .040 + .050

Weight at 33 days —Age .........+++- .012 + .065 + .160 + .048

32 THE AMERICAN NATURALIST [Vou. LV

~ Weight at inoculation is of course completely deter-

mined by the weight at 33 days and the rate of gain and

age at inoculation. The correlations between weight at

inoculation and these three factors are accordingly high.

They are rather different in the two groups of experi-

ments, but this is to be expected. There was much less

we 2 Ś g o £ r ç = < s z ee

Pee DAYS LIVED

F RN eee

AN ee,

ry oN \\,

* E N \N s,

2 x \ a ie Wetter at [mogut ATIO.

x k N X LOTS 10-14.

S z =

“sel N,

q R

& à 9n-J/99CRPAMS \ hy : bA :

+. gail

N poo-219¢ RAMS. ——-—— a z z

i p - é. =. a eee ho a j O

e DVER IGP GRAMS —-----[—-----— E fes. sad SN,

è K e R 2

e x AE

N be) Me

o CHART = 3 ; ST EAN

CHART 3. The E curyes of the guinea-pigs of lots 10-14, grouped ac-

ording to weight at inoculation

variation in age in lots 7-9 than in lots 10-14 and varia-

tion in age accordingly plays a less important part in de-

termining the variations in weight than in the latter case.

The rather high correlation between age and rate of gain

is due to the fact that nearly all of the guinea pigs suf-

fered a loss in weight following shipment at a little over

33 days of age. Thus the older they were at inoculation,

the greater the time in which they had to recover from

the effects of shipping. ‘The rate of gain before 33 days

is generally an indication of the probable later rate of

growth. In the present case, however, the slight nega-

tive correlatiop between the rates of gain before and

after 33 days seems to indicate that shipping delayed

No. 636] GUINEA PIGS AND TUBERCULOSIS ' 33

more the growth of those guinea pigs which had been

gaining rapidly than of those which had more growth to

make up. The small positive correlation between weight

at 33 days and age at inoculation reflects a decline in the

condition of the stock, due probably to the use of an in-

ferior quality of hay, which took place during the sum-

108 . i f] r r °

Pt K DAYS | LINED

= > NE as

4 E tá

NN Sa moe

¥ N Vy

* \ S

Ns Ko Te

cu p% \ k Imapi: atro.

N SN E Lors solr.

. k 3.4 .

N Ns j

aÀ N ENR

X T y“

EN a

= N NON

Fa a A

pa l a

3 mAN A

N re Aaa

des 49 bys = x Kae ae 4 ~

$ S0-69 DAYS p-o N Bo gy

» 0-39 wf TELLIN see r ~% oo

sy N a s “

deire ECR ioen te eras YE

` SEn j

Pa en Se

o CHART” 4 —

CHART 4. The mortality curves of the guinea-pigs of lots 10-14, —— ac-

cording to age at inoculation

mer of 1919, especially during the time in whidh lots 10-

14 were being raised.

With the help of these correlations, we can calculate

the degree to which variation in length of life is deter-

mined by each factor separately and by all of them com-

bined. For the last purpose, Pearson’s coefficient of

multiple correlatiòn can be used. This coefficient comes

out +.136 in lots 7-9 and + .251 in lots 10-14 for the cor-

relation between length of life and rate of gain after 33

days, age and weight at inoculation combined. The

degree of determination is measured by the square of

this ċoefficient. These three most important factors

combined, therefore, determine less than 2 per cent. of

34 THE AMERICAN NATURALIST [Vou. LV

the variation in length of life in lots 7-9 and less than

7 per cent. in lots 10-14. The separate contributions of

the various factors can be calculated with results piye

in the following table.

TABLE V

DEGREE OF DETERMINATION OF DAYS LIVED AFTER INOCULATIION BY AGE,

WEIGHT AND RATE OF GAI

ctors Lots 7-9 Lots 10-14

‘Direct effest Gf ago oces: ketur wis 0015 0005

ee aaa oes .0049 0229

Taio OF PAM ees eel as .0116 0233

ge and weight ....... — .0023 — .0051

age and rate of gain ... — .0027 — .0041

weight and rate of gain. Me .0054 + .0256

Total degree of determination ..... 7 Otsi -0631

wo 5 as ys è b f 20 a So om

ah... DAYS| LIVED

E me ees he

eee

X N S \

a in 5

N ES

£ Ñ. S p

@ NTR

è N w

p ù as c oF.

Š Da N + 33 DAYS TO /NOCULAT IOV

LS mee “s LOTS 10-14%.

a N M K

ee see,

X AUS

id è GAIN OVER 2| GRA. PER |DAY. Ns a ae

= i in Oy YN

a PAINIȚ2 GRAMS PER DAY —/—-—--— KO A

2 ARIN ġ-1 AM PER DAY = d Pai XN.

LOST END EE E h aa E ai

1° Nox = fe te ayes a

Ngee ees Fa epi, OE. WR

o CHART "S.

CHART 5. The mortality curves of the guinea-pigs of lots 10-14, steed ac-

cording to the rate of gain between vweamnmg and inoculati

While these figures can not be trusted in detail, owing

to the large size of the probable errors of the correla-

tions on which they are based, they indicate that age

within the limits of one to three months has virtually no

direct effect whatever on resistance to tuberculosis.

There seems to be a slight direct effect of absolute weight

No. 636] GUINEA PIGS AND TUBERCULOSIS 35

and a slightly greater direct effect of rate of gain, pre-

sumably as a measure of the condition of health.

NOTE

It may be of interest to readers who are not familiar with this method of

analysis to compare the degrees of determination shown in Table V with

those in a case in which we know a priori that there is complete determina-

tion of one variable by a number of others. The data for such a case have

already been presented. The weight at inoculation is ieee determined

by weight at 33 days and the gee tie of the rate of gain and interval be-

tween: 33 days and inoculat he coefficient of aaa correlation be-

tween weight at foëolatión aa pa three factors combined comes out

.9884 in lots 7—9 and .9628 in lots 10-14. The degrees of determination are

as follows:

TABLE VI

Lots 7-9 Lots 10-14

Direct effect of weight a 83 PES AE N 7106 -3689

AET EE AR AA ais 0907

AE bie 4 E E Serene eiy Whe 930 2177

Joint effect of ek at 33 days and rate of gain. — .0500 — .0073

weight at 33 days and age ........ — .0031 0907

rate of gain and age... 6005205. Ag _ 0763 1663

9769 9270

The departure of the multiple correlations and the sums of the degrees =

determination from unity are of the order usually met in such cases.

method of analysis applies strictly only where the effects of the factors are

combined by addition and the correlations are linear, conditions which are

not eanit met in the present case.

The facts brought out by the method of correlation are

presented graphically in Charts 2, 3, 4 and 5. These

charts show the decline in numbers, on the basis of 100

per cent. alive on the 15th day after inoculation among

groups of guinea pigs of lots 10-14, classified by weight

at 33 days, weight at inoculation, and rate of gain.

As a result of the foregoing considerations, it must be

_ concluded that the apparent condition of a guinea pig at

the time of inoculation and a knowledge of its past his-

tory give exceedingly little indication as to its probable

' length of life after inoculation. Over 98 per cent. of the

variation in length of life in lots 7-9 and over 93 per

cent. in lots 10-14 is caused by factors other than those:

discussed. This leads us to a consideration of hered-

36 . THE AMERICAN NATURALIST [Vou. LV

itary differences as one of the possible causes of this

variation.

HEREDITARY DIFFERENCES IN RESISTANCE

The average length of life in each inbred family and

crossbreeding experiment in lots 3-5, 7-9 and 10-14 is

given in Tables VII. VIII and IX together with other

TABLE VII

AVERAGES OF THE CHARACTERISTICS OF THE. GUINEA PIGS TESTED FOR RE-

SISTANCE TO TUBERCULOSIS IN Lots 3, 4 AND 5

The days lived in lot 3 are multiplied by 3/5 to make the average about

the same as in lots 4 and 5.

Family or Number Size of i Birth | Wt. at 33 | Wt. at In-| Age at In-| Av. Days

Experiment Tested Litter Weight | Days oculation | oculation Lived

oe 28 2.85 73.3 | 201.1 | 209.8 | 546 | 23.1

ey 15. 3.13 83.4 | 939.7 [957.7 49.7 20.3

Eo n 2 | 250 | 805 | 2560 | 215.0 | 53.0 | 210

BO. Pies 9 2.56 86.5 252.0 | 236.7 50.5 25.7

SO eee ees, 1 4.00 77.0 57.0 140.0 49.0 16.0

8s T A ACEEA Ae 10 2.70 82.1 236.5 243.5 51.5 26.0

GA Oe henry 24 A 85.5 247.4 247.7 52.0 25.7

AC Lee 21 3.76 ie: 230.2 | 223.6 51.3 27.4

Pid aro 9 2.33 101.2. | 277.9 | 280.6 54. 25.

Iod oc: 55 2.89 78.5 | 219.2 218.0 52.4 22.6

Crossbred. ... 64 3.14 84.6 | 244.4 | 243.7 | 52.0 26.3

TABLE VIII

AVERAGES OF THE CHARACTERISTICS OF THE GUINEA Pics TESTED FOR RE-

SISTANCE TO TUBERCULOSIS IN Lots 7, 8 AND 9

Bi i -| Av.

Lots 7, 8,9 No. Sane Weight bb Dars Dihneenphsd cat gr

Wa he Seat 13 2.46 80.2 197.1 243.1 64.2 22.8

AB eC SOS 16 2.81 91.0 257. 63.6

Boer eee 10 2.40 86.6 224.4 273.0 71.8 20.6

ios E 9 2.56 0 242.5 282.2 61.2 27.1

CO- er ee 13 3.00 80.2 226.2 286 67.8 31.5

CASE as 14 2.93 80.1 213.1 251.4 58.1 24.0

AG Nee 21 3.38 74.6 213 257 61.9 25.

Pe i gga: 13 3.38 3.7 249.5 320.0 69.2 24.8

Inbred: . 22... 2.58 86.0 231.3 278.3 65.0 23.5

Crossbred. ... 61 3.20 79.0 223.6 275.6 63.8 26.8

data. The distribution of the deaths is given in Tables

X, XI and XII. As already stated, most of the lots were

brought to a close while a few animals were still living.

This introduces an element of uncertainty into the aver-

No. 636] GUINEA PIGS AND TUBERCULOSIS 37

ages, but the assumption that these survivors died on

the following day is unfair only to the superior stocks.

TABLE IX

AVERAGES OF THE CHARACTERISTICS OF THE GUINEA PIGS TESTED FOR RE-

SISTANCE TO TUBERCULOSIS IN LoTS 14

Size of Birth | Weight at | Weight at | Age at D.

Lots 10-14 No. Litter Weight 33 Days bpewaiics Soin | Lived

Seabees Mae ee ET" Co ar 2.56 73.1 175.0 254.4 76.9 45.3

bs Reread OA 18 2.67 84.2 29 a | 301.1 91.3 35.7

32 5 2.20 88.0 223.2 228.0 63.0 35.4

i a MERE NAE 22 2.40 83.8 231.6 301.4 72.8 58.5

Col oe eee 26 2.81 82.8 210.9 274.6 66.8 52.7

A ie ee 33 3.45 (23 195.6 274.8 69.8 56.5

r ASPR RATA 31 3.26 75.3 213.7 273.9 68 45.9

Boe ee ci 22 09 79.5 220 67.7 40

DEOL a te EUERE 2.61 80 207.2 278.6 73.3 46.2

Crossbred 112 3.18 77.0 75.7 68.2 49.4

TABLE X

THE NUMBER OF DAYS FROM re TO eesti IN EACH EXPERIMENT

IN Lots 3, 4 AND

The intervals in lot 3 are multiplied by 3/5 to make them equal on the

eien to those in lots 4 and 5. The deaths before 15 days are given in

Ta The animals in experiments CO, CA and AC are also classified by

the ge of blood of family 35 at the bottom of the table. The last col-

umn includes two animals in lot 3 which survived at 61 days, one in lot 4

which survived at 42 days, and 3 in lot 5 which survived at 30 days.

Experiment

15 | 16 | 17 | 18 | 19 | 20 | 21 22 | 23 24 | 25 | 26 | 27 | 28 | 29 | 30 |30+| Total

DALITA 4 5 OS ae A a, ae 28

120 2 PS AR Tad Be C Ree 15

Soc. Ks 1 1 2

Dorie 1 2 1 bri 2 1 9

1 pea 1 1

ee es a Sy ie Fe ew 212 1 x 9

COs: Die tf aoa le 2 10

CA yi 1 bs ia a | cs eas ae ed Bere a 5 24

ACs. 211/121 Se eats PS te 5 21

0 (35). SRS Lee 1 ee ree Te 6 37

4 (35). 1 Pi242 4 1 6

2 (35).| | 1 1 | S73 7214 Si if

THE EFFECT OF [NBREEDING

In each case, the crossbreds (CO, CA, and B) lived a

little longer on the average than the combined inbred

families. This result is in line with the results of cross-

38 THE AMERICAN NATURALIST [Vou. LV

TABLE XI

THE NUMBER OF DAYS FROM INOCULATION TO DEATH IN EACH EXPERIMENT

OTS 7, 8 AND

Experiment

| | 41/42/42) Total

parc | | | ss

ee 3} 1) |1) |3|2|2 | 1 13

See 1/214 114) |2 11 16

Se ee 1} 1} 2) 2} 1 3 | 10

rees OE i 2/3 41 1 9

eooo 1| 1| 5} 4| 1 | Hoyos

COR 3} | 2). /1 ja} | 3 1 1/1] 138

res wreeraee 2| 3).2) 7] | | | 14

BO eek 1| 2| 2! 3| 2} 3| 1 uia |1 1} -|i 21

0 (35). 1/2} |4/6|1/5| |1| j1} |2| |1 24

1 (35).. 1| 1| 2| 2| 2 11 10

} (35)..... 1 2.3 Poke 1] ‘| 3|1 1/1] 14

TABLE XII

THE NuMBER OF Days FROM INOCULATION TO DEATH IN EACH EXPERIMENT

IN Lots 10 To

Experiment

15 | 20 | 25 | 30 | 35 | 40 | 45 | 50 | 55 | 60 | 65 | 70 | 75 | 80 (82+ Total

Beaver Sts |e eee ba eek E i Da Se se ee 27

IS es Cee Poe ek 18

Boies ees 1 pie as Md ee Zen | 5

BO vic ticws ge be: SN Ged Ms BS ie fe eas ee Bad yk 22

AE Ce Roe Ge t ee au ere 22

CO Seas Let EPS CS se) 3 Se 72212 3 26

CAA 2) a re eS By 2S x ae ae’ wee 33

BO aR: SVS (2, 4@i 8 | 61s 2) 2 ail 31

0 (35).. 315141815 |10|6;4/6 1 52

4 (35)... bs His re Pd eS a olar “a a ; ETA 13

4 (35). ja a e, A Ce eee ie ee E 25

ing on vigor in other respects. In the present case, how-

ever, the difference does not appear very striking. In

particular, the superiority of the control stock B over

the inbreds seems very dubious. An earlier test of this

question was made by Mr. E. H. Riley and Dr. E. C.

Schroeder, of the Bureau of Animal Industry, when the

inbred stock was in the sixth and seventh generations of

inbreeding. They found that the inbred was distinctly

inferior to the control stock in resistance. At that time,

however, eighteen families were on hand. It is not un-

likely that the four families tested in the present experi-

ment are a selected lot in respect to resistance.

No. 636] GUINEA PIGS AND TUBERCULOSIS 39

Returning to Tables VII, VIII and IX, there are no

differences between experiments CO (parents unrelated

inbreds), CA (sire crossbred, dam inbred) and AC (sire

inbred, dam crossbred), which indicate an influence of

the breeding of the dam on the resistance of the young.

There are rather large differences, it is true, but these

are not consistent and must be attributed largely to

heredity from particular families rather than to the

system of breeding.

' DIFFERENCES IN RESISTANCE AMONG INBRED FAMILIES

Passing now to a comparison of the different inbred

families with each other, we come to results which ap-

pear much more striking than the differences between

inbreds and crossbreds. In all three groups of experi-

ments, one family, 35, stands out as distinctly more re-

sistant than the others. It leads the average of the

others by 16 per cent. in lots 3-5, 19 per cent. in lots 7-9

and 43 per cent. in lots 10-14. The more striking result

in the last case is probably due to the weaker inoculation.

In spite of the large amount of variation in each case,

the probable errors put the superiority of Family 35

beyond question. Among the other families, family 2

is on the whole the most resistant. Families 13 and 32

are about equally susceptible. Only one animal from

family 39 was tested. This one was one of the first in

its lot, no. 7, to die, indicating low resistance in this

family also so far as conclusions can be based on oat

a slender basis.

INHERITANCE OF RESISTANCE AMONG CROSSBREDS

If there are hereditary differences in resistance, one

might expect to find differences among the crossbreds de-

pending on the families which went into their ancestry.

A preliminary test of this point was made as follows:

The length of life of each crossbred (CO, CA or AC)

was entered under each of the four grandparental inbred

families. Thus an animal in experiment CA, whose sire

40 THE AMERICAN NATURALIST [Vou. LV

was a cross between families 35 and 2 and whose dam

was of family 13, was entered once under each of the

former families and twice under family 13. The aver-

ages of the entries under each family in this tabulation

are given in Table XIII. It will be seen that in each

group of experiments, family 35 has a distinct lead over

TABLE XIII

AVERAGE NUMBER OF DAYS LIVED BY THE ee DESCENDED ' FROM

cH INBRED FAMIL

Each crossbred is entered under the family of each grandparent.

Lots 3, 4, 5 | Lots 7, 8, 9 Lots 10-14

Family

No. Av. No. Av. No. Av.

Sa eas 30 27.9 38 30.3 64

OSE ae Grae: 25.9 56 26.8 116 51.4

fs ges ey aE AN: 45 24.8 49 26.0 72 48.7

De eX Che 33 24.4 21 23.9 51 49.0

co La rari areas 28 26.0 16 24.9 39 48.0

DIR oa 19 26.0 12 24.2 19 48.8

the others as ancestral to resistant crossbreds. Family

39 and the other families which entered into the ancestry

of the crossbreds appear to rank with the more sus-

ceptible families 13.to 32. It will be seen that the rank

of families 35, 2, 13 and 32 as ancestors of resistant

crossbreds is the same as their own rank in resistance.

These results are brought out more clearly in a tabula-

tion in which all of the crossbreds are classified as half-

blood, quarter-blood and zero-blood of family 35. The

last class may be divided into half-blood, quarter-blood

and zero-blood of family 2. The results in comparison

with those for family 35, 2 and the others combined are

given in table XIV.

The half-bloods of family 35 are distinctly superior

to family 35 itself, and thus much superior to their other

ancestral families.* In the three groups of experiments

4 The rong study of another large series of animals (lots 15 to 21)

shows that w the order among the erossbreds, as related to the presence

of the blood a fanta 35 has been maintained, ‘the advantage of the half-

bloods over family 35 is absent, the curves being almost the same with the

half-bloods slightly less. The order of the inbred families it may be added

remains exactly as described, -

` No. 636] GUINEA PIGS AND TUBERCULOSIS 41

TABLE XIV

THE AVERAGE NUMBER OF Days LIVED BY THE CROSSBREDS WHICH ARE }

Buoop, } BLOOD AND HAVE NO BLOOD or FAMILY 35, THE LAST CLASS

BEING CLASSIFIED SIMILARLY WITH RESPECT TO FAMILY 2

The averages for Families 35, 2 and the other inbred families combined

are given for comparison

Lots 3, 4, 5 Lots 7, 8, 9 | Lots 10-14

No. | Av. No. Av. No. | Av.

Bed T E i aos 12 | 289] 14 | 322] 25 | 65.6

ere, G a es ee ee 6 |287 | 10 | 248] 13 | 56.5

No Bleed GIN. ii cde wi ie ee 37 | 25.0 | 24 | 244 | 52 | 43.9

o ($ blood Buns eseesssenrees 16 | 25.9 | 11 | 244 | 32 | 46.0

blood (2 blood 2)..............:. 16 | 244 |. 10 | 24.0 | 13 | 39.2

(38) (No blood REIR A 5 | Ssi 8 | 25.7 43.6

Famil GD a i 9 | 25.7] 9 | 27 58.5

VAD ak ee es. 28 | 23.1 | 13 | 228] 27 | 453

Other Tibrods a eana e a 18 | 204! 26 | 2261 23 | 35.6

3-5, 7-9 and 10-14, these half bloods exceed family 35

by 13 per cent., 19 per cent. and 12 per cent. respectively

in duration of life, and exceed the other inbred families

by 31 per cent., 42 per cent. and 60 per cent. respectively.

We have here more decisive evidence of the improve-

ment in vigor in this respect which may follow a cross,

than in the comparison previously made between the

total inbreds and total crossbreds. The most probable

explanation is the same as that applied to the improve-

ment in fertility, weight and death rate following crosses,

viz., that each inbred family tends to supply dominant

factors, favoring vigor, which are lacking in the other.

Applied to the present case, this means not only that

family 35 possesses dominant factors for resistance,

lacking in the other families, but that some or all of the

latter may possess such factors lacking in family 35. By

inbreeding among the crossbreds, it should be possible to

develop a strain even more resistant than family 35 pro-

vided that the linkage relations do not interfere with the

fixation of the factors for resistance in one strain.

The quarter-bloods of family 35 confirm the conclu-

sions derived from consideration of the half bloods.

They are about intermediate between the half-bloods

THE AMERICAN NATURALIST [Vor. LV

——

17 i 20 mi > g ¢ igs” ITA t ¥ 2

cu Rk Eds N, aN

ESN a an Die Pa DA i eer ee

ae Teih N X BA ARED STodKs

s ahaa be ze AND

‘7 oh N SA + CERTAIN Crosses

s ef. N, y \ ati Amo M

i % N X% \. oy L076 3-

~ %e, Me VN “he,

we = 3 4s, =

à \ z ee, Rt `N ie EPR Ie

5 = Ly FEN AP SEEDED $ A “od nA

= FAMILY 2 pesecescesven N w N SE

ru a Amily /2 s- eohbos--od \ > O Pen RA

4, a me N

u N, Po., . e]

a OSSES on . Poa wee

I WBLogn as Ls te ae Wa

= pena, be 4

op 3 DDD as 3% oo

Laie =a

Wo 8100035 ——- oN tes

; pire pa

z "s o

\ "

o CH R oe À

. The mortality curves of the guinea-pigs of lots 3-9, grouped by

bree ae Four inbred families. The crossbređs grouped according to the

amount of blood of the most resistant inbred family, No. 35.

yao

val

n DAYS LIVED

ra es Iweern STOC e

AN, a a AND

. wh N Dg N

on LAAN `$ Sr Amone THEM.

-> EE D. a LOTS 10414.

ind ed yN

v2 eh ee +

x ; a Aa

a ek `

g FAM 2c > X EN \ NI oe

3 FAMILY Z segece cadence $ s a SE R

FJ re canniiorz 4 \ P N. xt

= AML a2 if h \ .. ts, W

Pal osse ads moO N \

pa > *. WN

a BLOOD 3S A eel: w a

$ 2 odas ‘I \ ae Pee, “J el

BLDOD 3 “ea ra, on ley

iw r T z % ts

\ "nes, Toti

ú p F p i Po, Pe,

i HART Ť7 u A `-

CHART 7. The mortality Sba of the guinea-pigs of lots 10-14, grouped by

reeding as in Chart 6.

No. 636] GUINEA PIGS AND TUBERCULOSIS 43

and zero-bloods in lots 10-14 and in a combination of

lots 3, 4, 5 with lots 7, 8, 9.

Curiously enough the crossbreds with no blood of

family 35 do not show the improvement over their an-

cestral inbred families shown by the half and quarter

bloods. They exceed the average of families 2, 13 and

, 32 by only 13 per cent. in lots 3-5, 8 per cent. in lots 7-9

and 7 per cent. in lots 10-14. They are not consistently

superior to the best of these families, family 2. There

seems to be dominance of resistance on the whole but no

supplying of complementary factors for resistance such

` as was indicated in the crosses of family 35. This indi-

cates that families 2, 13 and 32 are in the main sus-

ceptible because of the same genetic factors. The rela-

tions of the factors present in the different families is a

question on which it is hoped more extensive evidence

can be presented later (Charts 6 and 7).

INHERITANCE OF RESISTANCE AND SEX

The erosses involving family 35 give some evidence on

the inheritance of resistance from sire and dam. The

average length of life in reciprocal crosses is given in

Table XV.

TABLE XV

THE AVERAGE LENGTH OF LIFE OF MALES AND FEMALES FROM RECIPROCAL

Crosses INVOLVING FAMILY 35

No. Av, Days

d (85) XD (mise) . eis eee ce wc ese cesses 7g

5? 59.4

12° 61.9

g (35 X mise.) X Q (mise.) ...---- eee eee e ees 6 64.0

Total, (35) om sire’s side ...........-+-+-+00% 18. 62.6

No. Av. Days

g (mise.) XQ (35) -riitse oses orren tenni 73 73.6

69 63.7

: B 69.0

3 (mise.) X Q (35 X mise.) ...-. es S 7 50.1 4

Total, (35) on dam’s side .:.....-++++++++++ n+ 20° 62.4

The results are irregular, as might be expected with

such small numbers, but they show that there is trans-

44 THE AMERICAN NATURALIST [Vor. LV

mission of resistance from each parent to offspring of

_ each sex. The total figures—62.6 days where family 35

is on the sire’s side, 62.4 where on the dam’s side—indi-

cate that there is probably equal inheritance from each

sex.

DEGREE oF DETERMINATION OF RESISTANCE By HEREDITY

We have shown that the length of life after inoculation

was determined less than 2 per cent, by weight, age and

rate of gain combined, in lots 7-9 and less than 7 per

cent. by these factors in lots 10-14. It is interesting to

compare the degree of determination by heredity with

these figures. Such a comparison can be made for the

- erossbreds by finding the correlation between length of

life and the amount of blood of family 35. The correla-

tion in lots 3-5 comes out + .319 + .082, in lots 7-9,

+ .539 + .069 and in lots 10-14 + 572 + .048. For rea-

sons which have been mentioned the latter two groups

are much more satisfactory than the group of three mis-

cellaneous lots, 3,4 and 5. The average in lots 7-9 and

10-14 is + .560 = .039. The square of this coefficient,

.314 indicates that over 30 per cent. of the variation in

length of life is caused by heredity, neglecting such

heredity as may be due to differences among families 2,

13; 32 and 39.

If we assume that some 10.per cent. of the variation

is due to differences in condition, weight and age and

over 30 per cent. to heredity, we still have over 50 per

cent. of the variation due to unknown causes. From

the nature of the case, however, a large amount of acci-

dental variation is to be expected.

RELATIONS BETWEEN RESISTANCE AND OTHER /

HARACTERISTICS

‘It is clear that family 35 is markedly more resistant

to tuberculosis than the other inbred families, and that

among these, family 2 is somewhat more resistant than

families 13, 32 and probably 39. The question arises

No. 636] GUINEA PIGS AND TUBERCULOSIS 45

whether these differences in resistance are related to any

of the other characteristics in which these families differ.

We may dismiss color with a few words. Families 32

and 39 produce only animals of the primitive golden

agouti color; families 2 and 13 produce only blacks, a

color which differs from golden agouti by a single re-

cessive factor; family 35 is composed of yellow agoutis.

This color is recessive to golden agouti, depending on an

allelomorph of albinism. We can not attribute the high

resistance of family 35 to the yellow agouti color since

the even more highly resistant crossbreds between family

35 and the other families are all of the golden agouti

color also found in the susceptible families 82 and 39.

In all of the families there are varying amounts of

white in a piebald pattern and of red or yellow in a tor-

toise-shell pattern. No influence can be attributed to the

amount of white unless it is assumed that an inter-

mediate condition is superior to either extreme. Family

39 has the least white, averaging less than 20 per cent.;

families 13 and 32 have the most, averaging over 80 per

cent.; while families 2 and 35 are intermediate with

about 70 per cent. and 60 per cent. white, respectively.

It may be added that the correlation between the amount

of white in the coat and length of life in family 35, lots

10-14, was found to be virtually zero. Similarly, family

35 is intermediate between family 39 with least red and

family 2 with most red in the colored parts of the coat.

The records in size, fertility and death rate among the

inbred families and cross breeding experiments during

the year 1919, in which all of these tested animals were

born, is given in Table XVI. In this table, the weights

and mortality records are corrected for the important

effects of size of litter by calculating separately the aver-

ages in litters of 1, 2, 3 and 4 and finding an index in

which these averages are weighted 1, 3, 3 and 1, respect-

ively. This means practically that all of the records

are reduced to the basis of an average size of litter of 2.50.

Because of the method of averaging, the figures are

46 THE AMERICAN NATURALIST [Vor LV

not strictly comparable to those presented in Tables VII,

VIII and IX for the animals actually tested. However,

it is safe to conclude from a comparison of these tables

that the animals tested were a fairly random selection

from their respective families and experiments.

It may be noted in passing that the figures in Table

TABLE XVI

AVERAGES OF THE CHARACTERISTICS OF GUINEA Pigs Born IN FIVE INBRED

FAMILIES (2, 13, 32, 35 AND 39) AND ites CROSSBREEDING Ex-

PERIMENTS IN 9

CO is the first cross between inbred families, In CA, the sire is cross-

red, dam inbred, AC is the reverse of CA. B is the control stock, The

mortality and size characters are indices in which the averages for litters of

1, 2, 3 and 4 are weighted 1, 3, 3 and 1 respectively.

Re K Sep 7

be 5> | 482 £ Ba | ee 3 glun H be 24

amtyor | £3 | 83 | ESS] 83 | 58 gA gaa 3 | ES | 3 | 25

Experiment | 5S | va | ogg] sd | a6 | S26| 395| #4 | ig | ae | Bs

zm | £3 |E2s| E= | 2 | gle) sad) @ 7R] ga

= a O GA

Aa 331 | 80.3 | 88.9 | 71.4 | 76.9 | 3.40 | 189.1| 2.43 | 3.82 | 9.30| 6.46

E 9| 87.1 | 85.2 | 74.2 | 88.4 | 4.76 | 245,4; 2.85 | 3.79 | 10.79| 6.67

3 a. 146 | 90.3 | 78.8 | 71.2 | 82.2 | 4.13 | 218.6| 2.25 | 3 ‘81| 5.30

35 258 | 80.2 | 84.7 | 67 .57 | 239.1| 2.37 | 3.68 | 8.72| 5.64

aa 72| 93.7 | 80.7 | 75.6 | 77.9 | 3.79 | 203.1| 2.40 | 2:97 | 7.13) 5.15

Tot. Inbred. |1,066 | 84.0 | 85.1 | 71.5 | 82.8 | 4.12 | 218.9| 2.47 97) 5.96

E 280 | 91.1 | 89.4 | 81.4 | 81.5 | 4.52 | 230.5| 2.55 | 3.90 |- 9.93| 7.8

OA ees, 302 | 85.8 | 94.5 | 81.1 53 | 233.8) 2.72 | 4.24 | 11.53} 9.20

AOL ES ox 419 | 87.7 | 90.9 | 79.7 | 88.5 | 4.54 | 238.2) 3.27 | 4.16 | 13.60| 10.78

ae 3 0 76.4 | 90.2 | 4.94 | 253.2 2.90 | 3.76 | 10.90) 7

XVI give a good illustration of the statements made in

regard to the effects of crossing, with the exception of a

few records based on inadequate numbers

From this table it will be seen that laiis 35, the most

resistant to tuberculosis, held a rather jów position

among the inbred families in most other respects. It

is actually the poorest in mortality among the young,

and fourth in size of litter and first in nothing. There

was thus no close relation between high resistance to

tuberculosis and vigor in other respects in 1919.

However, the rank of a family in a single year is not

always a safe indication of its true position genetically

No. 636] GUINEA PIGS AND TUBERCULOSIS 47

in characters in which environmental conditions are of

very much more importance than heredity. The rank

of the families has been calculated for the periods 1906-

1910, 1911-15 and 1916-19, making due corrections for

the effects of size of litter. As previously stated, high

correlations were found between the ranks in the first

two periods in most respects among the 23 inbred

families. Table XVII shows the ranks of the 5 surviv-

TABLE XVII

RANK OF FAMILIES IN RESISTANCE TO TUBERCULOSIS IN 1919, AND RANK IN

OTHER CHARACTERISTICS DURING Two PERIODS, 1911-15 AnD 1916-19

Mortality and size characters are corrected for the influence of size of

litter and (in 1916-19) for seasonal differences.

leg |£ | 2 | | = A + t | | | 3

is S ise 21. 3 | E $ oo E] = A | & | k

| $3 | Se | N 2 83] 2 ŝel £1 4] og] &,/ 38

remty| 38 | $5 989 Sa hetae aa] Els | gf] sf | Hs

qe |o° eee" | gia is | 21 218 |g | 28

Mie pa ra) = < a | | | be

Sep N ie, | | | |

5...., 1 | 33|31/31/|22/ 22| 22/22] 1-2 | 22] 1-1] 1-1

Gee T | 5-5 | 5-5 5-5 | 54 | 44 | 1-1 42 3-2

B 3 oa fe ee | 44d an 2-2 11-1 1 9-4 1 8-4 2828

32....| 3 |42 |55 |65 |33 |43 |43 |45 | 55 | 43/54 | 54

39. g lii 1-3 | 44/3-4/|34/3-3133/65/35|45

ing families in the second and third period. The reality

of the differences among these families is evident.

Taking the following 6 characters—percentage born

alive, percentage of these raised, birth weight, gain to

33 days, frequency and size of litter—the correlation

between the ranks in the two periods is + .83.

The positions of the families agree in the main with

those for the single year 1919. The position of family

35, however, is much better in most respects, a point

which will be discussed later.

The difficulty of classifying the families in the order of

general vigor is shown by this table. Families 2, 13 and

39 present curious combinations of high vigor in certain

respects with weakness in others. It is true that there

is perfect correlation between the ranks in size of litter

and adult weight, but the order in which this places the

45 THE AMERICAN NATURALIST [ Vou. LV

families, 13, 35, 39, 2 and 32, is far from being the order

in frequency of litter or the mortality records. Taken

as a whole, family 2 seems to have been rather the easiest

family to keep going. Its regularity in producing litters

and success in raising the young which are born alive are

factors in this, but even more important seems to be

another factor, probably correlated with that last named,

the longevity of the animals after the matings are made,

in which it far surpasses all of the others. Families 13

and 35 have also been relatively easy to maintain. This

leaves families 32 and 39 as those most difficult to keep

up to a desirable strength.

In spite of the better record of family 35 in the four

years, 1916-19, as compared with the single year 1919,

there is still no close relation between resistance to tuber-

culosis and vigor in other respects. In size, in both ele-

ments of fertility, and in the percentage of the young born

alive, family 35 is still inferior to families which are dis-

tinctly more susceptible to tuberculosis. It may, how-

ever, be significant that family 35 led in the percentage

raised of the young born alive and that family 2, second

in resistance, is also second in this respect.

It may also be significant that while family 35 does not

stand out in any particular element of vigor, unless in

that last named, it stands relatively high in all, and so is

the best family in the number of young produced per

year by a mating, the product of frequency and size of

litter, and is also the best family in the total percentage

of the young raised, the product of the percentage born

alive and the percentage of these raised. Moreover,

family 2 is second in both respects as in resistance. These

results suggest that while resistance to tuberculosis is

not related to the most important factors which deter-

mine the various elements of vigor, it is a contributing

factor to a sufficient extent to make the total efficiency of

a resistant family higher than that of a susceptible

family. |

In this connection, the low standing of family 35 dur-

No. 636] GUINEA PIGS AND TUBERCULOSIS 49

ing 1919 requires some consideration. The general con-

dition of the stock was much better in 1919 than in any

of the three preceding years. The principal cause of the

high death rate, small and infrequent. litters and slow

growth during 1916, 1917 and 1918, especially during the

first half of each year, was probably an insufficient sup-

ply of green feed during late winter and early spring.

Symptoms, such as lameness, swollen and bleeding gums,

were noted rather frequently. These were probably in-

dicative of scurvy. A form of pneumonia was also rather

common. Whether tuberculosis was present in the stock

at this time is not certainly known. It is doubtful, as

guinea pigs seldom take the disease unless directly inocu-

lated. However, this may be, it seems probable that

family 35 has a special ability to withstand exceptionally

adverse conditions. It is not unlikely that this character-

istic may be connected directly or indirectly with its re-

sistance to tuberculosis. Under good conditions, on the

other hand, there seems to be little if any relation between

this form of vigor and apparent vigor in other respects.

CONCLUSIONS

_ There is little or no relation among guinea pigs be-.

tween resistance to tuberculosis and sex. The present

data indicate a possible superiority of the males, but one

which is too slight to be certainly significant.

The data suggest a slightly greater susceptibility among

the progeny of very young or very old females but this

also is of doubtful significance.

Size of litter, birthweight and rate of gain up to wean-

ing give virtually no indication of the probable length of

life after inoculation.

The rate of gain preceding inoculation and the age

and weight at that time all together determined less than

7 per cent. of the variation in length of life in a very

heterogeneous lot of guinea pigs and less than 2 per cent.

in a somewhat more homogeneous lot.

Marked differences in resistance were found among a

number of inbred families of guinea pigs.

50 THE AMERICAN NATURALIST [Von LV

The high resistance of one of these families was trans-

mitted by each sex to the offspring of each sex in crosses

with other inbred families.

In crosses involving this most resistant a the

progeny were superior even to this family itself, indicat-

ing not only the dominance of resistance over suscepti-

bility but possibly also the presence of se adn ti

factors among the families.

In crosses among the more susceptible families, the

progeny were little, if any, superior to the family of the

better parent, indicating dominance but not complemen-

tary factors in this case.

Over 30 per cent. of the variation in length of life after

inoculation among the crossbreds was determined by the -

amount of blood from the best inbred family. Allowing

10 per cent. of the variation as due to age, weight and

condition, 50 to 60 per cent. remains as due to accidental

and unknown causes.

The factors which determine the resistance of a family

to tuberculosis are not closely related to the other ele-

ments of vigor, including rate of growth and adult

weight, frequency and size of litter, the percentage of the

young born alive and the percentage of these raised to

weaning. There is some evidence that they are contribut-

ing factors to a sufficient extent to give the most resistant

family a record above the average in each element of

vigor and so give it the highest or nearly the highest

total efficiency. Even this relation of resistance to other

elements of vigor appears to be present only under ex-

ceptionally adverse conditions when it may play a direct

part in determining the health of the stock.

The results in regard to resistance to tuberculosis are

like those for other characteristics as regards the differ-

entiation among families brought out by inbreeding, the

improvement resulting from crosses between inbred

families and the independence genetically from other ele-

ments of vigor.

GAMETIC AND OBSERVED RATIOS IN DROSOPHILA

DR. CALVIN B. BRIDGES

THE populations and families with which the geneticist deals

are not the real objects of his investigation; for him, the distri-

bution of characters is only an index of the preceding distribu-

tion of genes in gametes. But the whole course of embryonic

development, with heavy mortality possible at every step, has

intervened between the individuals that he classifies and the

gametes from which they came. The observed classes correspond

accurately to the original gametic series only in case this mor-

tality is indiscriminate—thaf is, only if there is no differential

viability.

In the breeding work with Drosophila there has been a con-

tinual effort to eliminate distortion in the ratios, which depends

largely upon: (1) the extent of the mortality involved, this being

characteristic in amount for each mutant type and character

combination, (2) the suitability of the culture media and condi-

‘tions, and (3) the competition when the number of developing

individuals is in excess of the optimum number for the available

food supply.

The problem of over-crowding (3) is simplest of solution,

though over-crowding was thé largest source of disturbance in

most of the early work, as well as in some of the later. The rem-

edy is, in the first place, to limit the number of eggs per culture

to the output of a single female. No mass-cultures should be

raised in experiments in which the ratios among the offspring are

of importance. In the second place, as the larve grow larger and

also increase in number with each day’s output of eggs, the com-

petition becomes intensified throughout the later stages of the

culture. To meet this increasing demand, there must be fresh

supplies of food, or enough food’ must be provided at the start

so that even at the end there is sufficient for free development of

all larve. In point of economy it is better to concentrate on a

few cultures that are liberally supplied than to raise a greater

52 THE AMERICAN NATURALIST [Vou. LV

number that would mean optimum conditions for none and aai

concerning the reliability of all.

The main problem in connection with the environment (2) is

to find a kind of food that will allow full development of even

very weak classes. It is in this field that the greatest changes in

method have been made. For some years—from 1910 to 1916—

some modification of the fermented-banana method of preparing

food was followed. Ripe sound bananas were peeled, and the

pulp left for about 24 hours in a liquid containing yeast. This

liquid was usually the fermented juice from the previous lot of

bananas. About 25 grams of this fermented banana was put in

the bottom of a culture bottle and covered with absorbent paper.

It was suspected that the real food of the larve was not pri-

marily the banana but was rather the yeast cells and perhaps

also the bacteria, the banana being mainly the culture medium

for the yeast. This has been established by the work of Northrop,*

of Loeb and Northrop, and of Baumberger.* In July, 1916, in

consultation with Northrop, I started experiments with a view to

using as a culture medium standardized solutions, instead of

banana. The solution was absorbed and held in a cake in the

bottom of the culture bottle by shredded paper toweling, which

offered extensive surface for the growth of the yeast. This

method was unsuccessful; the flies laid few eggs and these were

often overgrown by the yeast and killed. Esters and other chem-

icals with fruit odors did not lead to greater egg production.

Perhaps better success with culture solutions would be obtained

in supplementing and modifying banana methods.

The banana method was modified with a view to discouraging

the growth of moulds and putrefactive bacteria by mild anti-

septies or correctives, such as benzoate, thymol, formaldehyde,

alcohol, powdered marble for neutralization of excessive acidity,

ete. Good results were obtained with alcohol, where several ex-

tensive sets of comparative tests seemed to show that about 1.5

per cent. of alcohol in the food was desirable. The most success-

ful alcohol method was roughly as follows: The pulp of sound

ripe bananas was weighed and put with an equal number of c.c.

of 3 per cent. alcohol in a shallow, covered dish.. No yeast was

added, since enough wild yeast was usually present. The food

1 Jour. Biol. Chem., 1917, pp. 181-187.

2 Jour. Biol. Chem., 1916, pp, 309-312.

3 Jour. Exp. Zool., 1919, pp. 1-28.

No. 636] GAMETIC RATIOS IN DROSOPHILA 53

was at its best when it had fermented for about 24 hours. The

optimum amount of drained banana was found to be about 25

grams per bottle. This was put upon the bottom of the culture

bottle and one gram of paper-toweling strips (about 5. <.7 em.)

was matted down on the top. Pint culture bottles gave a greater

output per pair than halfpints. The alcohol method was used

~ more successfully than the old method during the fall and winter

of 1916

In the spring of 1917 considerable work was done in testing

out various media containing starch, sugar, peptone and salts.

This method gave good results except that trouble with moulds

was greatly increased.

In the autumn of 1916, Dr. R. W. Glaser told me of certain

culture-media experiments that Mr. Baumberger and he were

carrying out with banana infusions and agar.* Dr. Glaser later

sent me directions for preparing these media and also some pre-

pared tubes. My tests of the method showed that the amount of

food was inadequate for general use, although sufficient for the

small number of flies that they wished. I increased the con-

centration of the media by the addition of sugar, banana flour,

ete., but principally by grinding up and using all the pulp of

the bananas, instead of using simply the strained juice. A.

comparison of fresh banana with banana that had been fer-

mented before incorporation showed that the fresh banana was

superior. Likewise fresh banana was superior to banana raised

to the boiling point at any stage of preparation. It was found

that yeast should not be distributed throughout the media.

Experiments showed that it is advisable to have a very light

seeding of yeast confined to the surface of the solidified media.

Also it is well to keep the yeast from the margin as much as

possible, since fermentation at the sides and beneath the cake

makes the cake break loose and rise. The amount of agar was

found to be adequate at 1 per cent.

It was some months before this method was improved so far

that it gave better results than those given by the old method or

the aleohol method. In the spring of 1917 it was worked out

well enough so that it was substituted for the old method in

my regular work. By the winter of 1917 it had become quite

4 Science, 1917, pp. 21-22.

54 THE AMERICAN NATURALIST [Vou. LV

generally adopted in the laboratory. Several points have been

improved since, so that the procedure at present is as follows:

1. Use bananas that are thoroughly ripe or over-ripe.

2. Peel the bananas and weigh the pulp (100 grams of pulp

provides for about four culture bottles).

3. Weigh agar-agar, 2 per cent. of amount of banana.

4. Measure as many c.c. of water as there are grams of banana. ©

5. Add agar-agar to water and heat until the agar has dissolved.

(Complete solution is hastened by the addition of a small

amount of fresh water soon after the boiling point has

been reached.)

6. While the agar is heating, press the banana through a potato

masher or a coarse sieve, and place in readiness the

bottles (which should have been previously washed and

also preferably steam sterilized). Get ready yeast (Magic

Yeast ground up) and paper (absorbent paper, paper

toweling cut into 4-fold squares 3” X 2”) and cotton

(stoppers may be reused, but should be dry sterilized by

enclosing over formalin. Cotton stoppers are better if

made rather tight and covered with very soft cheese-cloth).

7. Stir banana into hot agar solution. Mix thoroughly. Mix-

ture should not be heated any. longer.

8. With ladle and funnel pour about 50 c.c. of the media into

each half-pint or pint milk bottle. (The media should

be at least ł in. thick to stick well.)

9. Sprinkle top lightly with dry yeast.

10. Put in contact with media a 4-fold square of absorbent

paper. |

Stopper with cotton.

Use same day. Best to use as soon as cool. Not good after

two days.

Flies can be mated in vials and then transferred to the cul-

ture bottles at the end of the day. A little food may be kept

going’ by the alcohol method’ for use in vials, for covering over

mould patches in culture bottles, and for refeeding stock

cultures.

_. The distortions in the ratios that arise from mortality char-

acteristic of given mutants and combinations (1) can not be

eliminated by direct methods. Fortunately, a large proportion

No. 636] GAMETIC RATIOS IN DROSOPHILA 55

of the mutants are little if at all below normal in degree of

viability; that is, when such mutants are compared with the

wild type under identical conditions, the observed ratios show

little or no deviations from expectation beyond those due to

random sampling. As an example may be mentioned white-

ocelli, which is known to have maintained itself with practically

undiminished frequency through 175 generations of competi-

tion, under unfavorable conditions of culture,® with the wild-

type. In the main, the mutant races that show normal viability

are those whose somatic effects are ‘‘slight’’. Thus, white-ocelli

affects the color of the tiny group of three ocelli on the top of

the head. The character, though involving so small an area, is

perfectly sharp and definite, and under proper conditions of

illumination and magnification is fairly easy to classify.

same is true of many other ‘‘slight’’ mutations, such, for ex-

ample, as speck, cross-veinless, and hairy, which are among the

most valuable Drosophila mutations. On the other hand,

mutants that involve more extensive or manifold changes, such

as club, notch, rudimentary, and delta, are also among those

poorest in viability. Some of these changes in themselves inter-

fere with the success of the individuals possessing them: flies

with ‘‘spread’’ wings or ‘‘dachs’’ legs are liable to become

caught in the culture media and die. These changes are also

sometimes obviously accompanied by serious internal derange-

ments. In the case of streak, for example, it can be seen that

the internal muscles of the thorax are largely replaced by bub-

bles and blood sinuses. The correlation between inviability and

the extent of the visible change is high, but is lessened by the

cases in which the internal accompanying changes are of slight

disadvantage. Thus, the mutant ‘‘pads’’ resembles ‘‘elub’’

very much, and appears to be a greater change in the same

direction, but is nevertheless far freer from inviability. Con-

versely, certain mutants that are usually lethal occasionally do

produce offspring, which are then not as strikingly different

from the wild-type as some other mutants that have good

viability. Lethal-10 very occasionally survives, and the indi-

viduals are scarcely to be distinguished from dwarfs of a cer-

tain mutant race (dwarfoid) that is little inferior to the wild;

type in viability.

5 Biol. Bull., 1920, pp. 231-236,

56 THE AMERICAN NATURALIST [Vou. LV

The connection between observed character-change and in-

viability is even more indirect than suggested above. In the

Drosophila work it is not the comparative viability of adults

possessing given character differences that is of the most im-

portance. Even though many of the characters are of such a

= nature that their possessors would be under a serious handicap

in competition, in relatively few cases does this fact lead to

alterations in the observed ratios, since the classifications are

made usually soon after the flies hatch, t.e., every 24 to 48 hours.

It is true that certain mutant forms such as ‘‘divergent’’ and

“gull” and ‘‘bifid’’ wings, also ‘‘dachs’’ and ‘‘reduplicated’’

legs tend to become entangled in the culture media and drowned

immediately after emergence, so that in these cases the observed

ratios are somewhat different from the hatching ratios. There

are also a few mutants—mostly semi-lethals—in which the adult

is unable to live very long even under the most favorable condi-

tions. Among these may be mentioned ‘‘lemon,’’ ‘‘apterous,’’

and especially ‘‘decrepit.’? The ‘‘decrepit’’ flies die a few

hours after hatching in spite of all care in helping them emerge

from the pupa case, in keeping them in quarters not too dry

or wet, and in suplying them with suitable food. It would

seem that the death of such flies as are obviously weak on hatch-

ing is to be referred to difficulties encountered in the pupa stage.

Even inviability arising in the pupal stage, like that in the

adult stage, is less general and significant than that in the larval

stage. Most of the inviability that affects the ratios of adults

is to be referred to differences acting in the larval stage, as is

evident from comparative studies of the results of pair and mass

cultures and of changes in culture methods that affect only the

larval period. The difference between mass and pair cultures

is essentially a difference in the number of larve that are in

competition, the food conditions and the character of the larve

being at first identical in the compared cultures. It is found

that the distortion to the ratios among the adults is roughly

proportional to the number of larve in competition. How ex-

treme such competition may be is evident from the fact that a

point is soon reached after which further increase in the number

of mothers brings no increase in the number of progeny and

may even result in a decrease. So predominant is the larval

stage in its influence upon viability that the chief field of

No. 636] GAMETIC RATIOS IN DROSOPHILA 57

improvement of culture conditions has been that of the char-

acter and methods of use of the food for the larve. There are

specific viability differences among the larve of the different

_ mutant types and combinations. Such viability differences

must depend upon differences in the characters of the larve,

and these, because of the intervening metamorphosis, have little

direct relation to the characters of the adult, but are products

of the action of the same mutant gene. The high correlation

observed between extensive change in adult characters and high

degree of inviability must, then, mean that such genes generally

cause changes which interfere directly with the success of the

larve.

Three larval characters are known—the tumor responsible for

the death of lethal-7 larve, the much shortened larve of

the mutation ‘‘chubby,’’ and a marking on the posterior end,

viz., ‘‘barette.’’ It is supposed that a high proportion of the

larval characters that lead to inviability are differences in in-

ternal structures, but some of these might be detected. How-

ever, no systematic search for larval characters has been made

even in the case of the inviable mutants where such differences

are probably present.

As we have stated, the distortion in ratios that arises from

inviability, especially inviability originating in the larval stage,

can be very materially reduced by improvements in culture

media and in methods. Many poorly viable mutants can be

made quite generally usable, as, for example, dachs. But when

a mutant such as dachs is to be used in linkage determinations,

the experiments should be so planned as not to include more

than one of these characters. The presence of a single

poorly viable character in an experiment does not prevent the

calculation of correct crossover values. The complementary

classes that do not include the inviable character should be in

the same proportions as in the gametic series. Even in cases of

mutants completely lethal, the linkage relations of the lethal

gene can be calculated accurately from the ratios shown by the

other characters of the cross. The classes that include the

inviable character are often also usable, but with less certainty

that the values are correct. Such values are correct when the

presence of the mutant decreases by the same percentage the

size of every class in which it oceurs. The fewer the mutants

involved in an experiment, the greater the likelihood that this

58 THE AMERICAN NATURALIST [Vou. LV

result will follow. Unexpected irregularities in ratios may

arise where many mutant characters are distributed in different

combinations. These peculiarities of inviability are probably

comparable to ‘‘specific’’? and ‘‘disproportionate’’ modifications

in eye-color, ete.®

If more than one poorly viable mutant is present in a linkage

experiment, there is distortion in the ratios due to linkage, and

such experiments are either entirely worthless, or are only to be

regarded as rough indicators of the real relations. As we saw,

if only one inviable mutant is present in a cross, one of each of

the pairs of complementary classes remains undisturbed; and

correct values can be calculated from them. The presence of a

second inviable mutant leaves undisturbed only that class in

which neither mutant occurs. The calculation of crossover

values under this circumstance is somewhat comparable to solv-

ing for two unknowns with a single equation. Solutions can be

obtained only by assuming some relationship between the two

disturbances. Thus, we may assume that the disturbances are

independent; that is, that there is no specific interaction of the

kind mentioned above, and the class in which both mutants

oceur is accordingly of the size that would be expected from the

amount of the disturbance present in those classes in which each

occurs by itself. On this basis, the crossover values are calcu-

lated from the square root of the product of the two comple-

mentary non-crossovers, and likewise of the crossover classes,

instead of from the sums of such complementary classes.’ The

assumption of independence would be approximately correct in

perhaps a majority of crosses in which only two or a few loci

are involved. If the disturbances are related—if they tend to

neutralize or to exaggerate each other—a correction can still

be made by raising an equal number of individuals in the com-

plementary cross. , In the two complementary back-crosses, a X b

(repulsion) and ab wild-type (coupling), the character com-

binations that are non-crossover classes in the repulsion experi-

ment are crossovers in the coupling experiment, and vice versa.

If the presence of a particular class has given a crossover value

too high in the one cross, then it will give a value correspond-

6 Jour. Exp. Zool., 1919, p. 374.

7 This geometric mean ean method was proposed by Muller, who gave an ex-

cellent discussion of the difficulties involved in aiiai viability. (AM.

NAT., Vol. L, p. 351 ff.)

No. 636] GAMETIC RATIOS IN DROSOPHILA 59

ingly too low in the complementary cross, and the mean value

will be correct.’

Since the effects of inviability are likely to be more pro-

nounced, even disproportionately so, as the number of mutant

characters present simultaneously is increased, it is advisable to

plan any linkage experiment in which several characters are to

be involved, in such a way that the characters are distributed

as evenly as possible. The type of back-cross that gives the

evenest possible distribution, as well as the smallest proportion

of individuals in which the higher combinations occur, is that in

which half the mutants have entered the cross from one parent

and the other half from the other parent, and in which the

mutants are ‘‘alternated’’ as regards their positions along the

chromosome. Thus, for example, let us consider a back-cross in

the third chromosome in which the seven mutants to be used are:

roughoid at 0.0; hairy at 25.8; scarlet at 35.1; dichxte at 38.5;

pink at 44.6; spineless at 54.2; and ebony-2 at 66.9. The two

parents should be roughoid scarlet pink ebony-2, and hairly

dichæte SANER: and the formula for the F, multiple heterozy-

gote would |

ru st p pr

h D ss

The production of an individual possessing all seven characters

would require an hexuple crossover, which almost certainly

would not occur.?

A method that: overcomes inviability effects to the greatest

extent where many mutants are involved, but which unfortu-

nately requires too great labor for general use, was devised by

Muller..° F, females heterozygous for any number of mutants

are crossed, not to the multiple recessive as in the ordinary back-

cross, but to the wild-type. Except for the dominant mutants,

none of the characters involved appears in the resulting indi-

viduals, and hence do not exert injurious effects. These indi-

8 This ‘‘ balancing of the inviability’’ has been discussed at greater length

by Bridges, Jour. Exper. Zool., 1915, p. 3 ff.; Jour. Exper. Zool., 1920, pp.

281, 288; by Morgan and Bridges, Carnegie Pub. No. 237, p. 19, 43; and

by Muller, Am, Nar., 1916, p. 353.

"9 See Bridges, Four. Exper. Zool., 1920, p. 295, for discussion and other

examples of the ‘‘alternated back-cross.’’

10 AM, Nart., 1916, p. 354

60 _ THE AMERICAN NATURALIST [Vou. LV

viduals have then to be tested singly to determine what recessive

characters.they carry and hence to what crossover category they

belong.

Thus, by improvements (1) in the type of experiment planned,

(2) in the culture media and methods used, and (3) in the

method of calculation, disturbances in the ratios can usually be

held within negligible amounts.

To these indirect methods of obtaining accurate values is to

be added one still more important—namely, the discovery of

new mutants in which viability is practically normal, and which

ean be substituted for mutants less satisfactory in that regard.

Many of the loci are represented by several mutant allelo-

morphs, which often are different in viability as in other char-

acteristics. Thus, of the eight cut allelomorphs, or appearances,

eut-6 is distinctly the most nearly normal in viability. ees wie

of the five or six allelomorphs of the truncate locus, ‘‘dumpty”’

is the most satisfactory.

In most of the more complex linkage problems, especially

those involving linkage-variations or coincidence, the behavior

‘of particular regions of a chromosome is being examined, an

the particular loci utilized are only indices of the behavior.

What is most essential, therefore, is that there be workable

mutant loci distributed rather evenly over all regions of the

chromosomes. As the number of mutants in a particular region

increases, there is a greater range of choice and greater prob-

ability that one or more of the mutants of that region will have

normal viability. Thus, bifid (at 7.3) was long the only work-

able mutant at a favorable distance from the left end of the

X-chromosomes. More recently ruby (at 7.5) because of its

better viability has displaced bifid from general use, and this in

spite of the fact that ruby interferes with the classification of

several other eye-colors (especially prune and garnet) while

bifid is workable with nearly all other mutants.

In regions less well represented in numbers of mutant loci, a

mutant with excellent characteristics may be used rather than

one whose position is more favorable but whose other character-

istics are poorer. Thus, ‘‘humpty’’ is very favorably located in

the second chromosome in the middle of the long region from

eurved (73.5) to plexus (98.5), but its viability is so poor that

in most experiments it is better to leave this section unfollowed

than to introduce humpty. There are at present few regions

No. 636] GAMETIC RATIOS IN DROSOPHILA 61

that are not satisfactorily represented. By far the longest of

these in the 25.8 unit interval from roughoid to sepia in the left

end-region of the third chromosome. Because of uncertainties

in classification, it is not ordinarily possible to use more than two

eye-color mutants together in an experiment, and such masking

of one character by others affecting the same organ leads to a

continued search for combinations of characters that can be

handled simultaneously. In general, the slight mutant char-

acters mask each other less than do extreme ones, and these are,

usually the least inviable. There is a continual improvement

of the working material by the substitution of better mutants.

A PROBABLE EXPLANATION OF POLYEMBRYONY IN

THE ARMADILLO

PROFESSOR CHARLES R. STOCKARD

CORNELL UNIVERSITY MEDICAL COLLEGE, New York CITY

By arresting the development of the fish’s egg during early

stages double individuals and twins are frequently induced.

The interruption or arrest makes it possible for more than one

potential growth point along the germ-ring to give rise to an

embryonic shield. In other words, accessory invaginations or

blastopore formations occur as the initial structural step in

doubleness. The interruption in the development of the fish

embryo must be introduced during the cleavage stages and be-

fore gastrulation in order to produce such phenomena. Among

hundreds of eggs arrested during later developmental stages no

double monsters or twins ever occurred. A complete account of

these experiments is soon to be published but for our present

purpose two facts are important: First, accessory embryo forma-

tions result from arrests in the developmental process; and

second, the arrest must occur before gastrulation has taken

place.

In the light of these experiments it has seemed possible to

interpret somewhat more clearly than has formerly been done

the remarkable phenomenon of multiple embryo formation in

the armadillo.

On examining the uterus in two pregnant specimens of a

South American armadillo von Jhering, in 1885, discovered that

each contained eight fetuses enclosed within a single chorion.

He correctly concluded that all of the fetuses in each mother

had been derived from a single egg by some process of division

into separate embryonic rudiments. After this valuable dis-

covery and interpretation, the study of the armadillo’s develop-

ment lapsed and nothing of importance was added for almost

twenty-five years. Two series of investigations were then be-

gun simultaneously one on the Texas armadillo by Newman and

Patterson,* and the other on the South American species by Fer-

1H. H. Newman and J. T. Patterson, Jour. Morph., Vol. 21, p. 359, 1910.

No. 636] POLYEMBRYONY IN THE ARMADILLO 63

nandez.? The growth and expansion of these twin studies has

brought our understanding of the phenomena of polyembryony

in the armadillo to a considerable state of maturity.

These authors readily agreed that in most species of armadillo

the individual members of a litter, usually four in the Texas

species and eight in the common South American form, are all

derived from a single egg. It required considerable. effort, how-

ever, to obtain the material that would furnish the morphologi-

cal stages of the process by which the polyembryonie development

was accomplished. We are finally indebted to Patterson,’ for

the very thorough and satisfactory manner in which he has col-

lected and studied the early embryonic conditions; and. particu-

larly for having shown the first stages of the budding process

through which the single blastocyst gives rise to four distinct

embryonic areas, each exhibiting a typical primitive streak

region.

In connection with the fish experiment it now becomes im-

portant to ascertain exactly what degree of development has

been attained by the armadillo blastocyst at the time the bud-

ding process begins. And since, according to my interpretation,

these buds should arise at the time of gastrulation or blastopore

formation, it becomes necessary to consider very briefly the

germ-layers and gastrulation in mammals. The decidedly pre-

cocious and highly modified method of forming the primary

germ-layers in the mammalian blastocyst is not strictly com-

` parable to gastrulation or the method of germ-layer formation

found among the other vertebrates. On the other hand, the em-

bryonic line or primitive streak of the mammalian egg is ex-

actly comparable to the blastopore and head process formation

in the simpler forms. :

The blastocyst of the armadillo has already, by a process of

cell migration and delamination, separated off the primary ento-

derm from the ectoderm and further modified these layers be-

fore the budding which forms the embryonic primordia has

begun. The primordia are first formed by a thickening of the

ectodermal layer of the blastocyst. The primary entoderm then

invaginates into the primordia to form the secondary entoderm

of the gut. The precocious cell migration and splitting into

layers in the mammal’s egg is associated with the early implanta-

2 M. Fernandez, Morph. Jahrb., Ba. 39, p. 302, 1909.

3 J. T. Patterson, Jour. Morph., Vol. 24, p. 559, 1913.

64 THE AMERICAN NATURALIST [yor UV

tion of the embryo upon the uterime-wall of the mother, and the

later primitive streak formation may be interpreted as related

to the actual gastrulation or blastopore formation ain from

which the line of the embryo always develops.

Whether the validity of the above briefly outlined interpreta-

tion of the germ-layer formation is admitted or not, we have

in the armadillo a process of budding taking place from the

blastoderm and associated with accessory or extra blastopore

formation in much the same way as are the accessory embryos

along the germ-ring in the egg of the bony-fish. These buds

also accord with Kopsch’s description of a double gastrular

condition with two blastopores in a blastoderm of Lacerta agilis,

from which he concluded that twin formation as well as anterior

duplication arises from a double Einstiilpungen. And further,

Assheton has described a similar condition in a blastodermic

vesicle of the sheep. He, however, imagined the condition to

have been due to a splitting during the morula stage.

The double primitive streaks in the hen’s egg and other

forms all lend themselves to strengthen the interpretation that

double embryo formation first asserts itself by a double gastrula-

tion or blastopore formation, which is initially a process of

double instead of single bud formation. Patterson’s description

of the origin of the quadruplet buds in the Texas armadillo

furnishes the most striking case in the study of these conditions.

And we may conclude that the budding or accessory embryo

formation in the egg of the armadillo is exactly the same develop-

mental process as that which gives rise to twins and double indi-

viduals in other vertebrate eggs.

However, the very important question yet remains to be an-

swered. Why does this accessory bud formation occur so con-

stantly in the Texas armadillo in contrast to the single embryo

formation of mammalian eggs in general? Patterson failed to

answer this question, but he supplied some very significant data

which Newman,* has appreciated as being intimately connected

with the occurrence of polyembryony.

In connection with the collection of material Patterson? dis-

covered a ‘‘period of quiescence’’ of the embryonic blastocyst.

Regarding this he states:

The fact was first made apparent in 1911, when, after I had started

collecting two weeks earlier than in the preceding year, I failed to

4H. H. Newman, ‘‘The Biology of Twins,’’ Univ. of Chicago Press, 1917.

No. 636] POLYEMBRYONY IN THE ARMADILLO 65

obtain the cleavage stages, although judging from the condition of devel-

opment in the vesicles collected in previous years, one would naturally

expect to find these early stages during the period of my first collec-

tion in 1911.

The following year he began collecting still two weeks earlier

and again had a similar experience.

Practically all of these vesicles lie free within the uterine cavity,

either in the horizontal groove or in the region of the attachment zon

(placental area).

It is evident from these data that the embryonic vesicle remains for

some time lying free within the uterine cavity. Just how long this

period lasts; I am unable to state; for practically every old female

taken at the earliest date (October 15) at which I have collected, pos-

sesses a free blastocyst. . . . Taking all the facts into consideration, I

estimate the “ period of quiescence” to last about three weeks; that is,

from about the middle of October to the third or fourth of November.

In a study of sections no mitotie divisions were found to occur

in the blastocysts during the ‘‘ quiescent period.”’

The only point of interest cited by Patterson in connection

with this peculiar phenomenon of interruption in development,

was the fact that in no other mammal, except the deer, had

such a condition been found. Bischoff had long ago, 1854, re-

ported a ‘‘period of quiescence’’ lasting for some weeks during a

so-called morula stage of the deer embryo.

Newman‘ has recognized the importance of Patterson’ s dis-

covery of a ‘‘quiescent period’’ during the early development of

the armadillo, and states in a discussion of twin formation that

this ‘‘period of quiescence’’ probably, ‘‘holds the clue to the

physiological explanation of polyembryony.’’ In this position

‘ewman is, in my opinion, largely right, but this i is as far as the

data led him, and he finally remarks:

The problem is to locate the factors responsible for the slowing down

of the developmental rhythm. Whatever these factors may be, and

we have no definite knowledge of them, the result of retardation is

polyembryony.

Newman thus fails to appreciate the second point in Patter-

son’s discovery, and that is that the blastocysts always lie free

in the uterus during the ‘‘period of quiescence.’’ This fact en-

ables us to go one step further since the lack of attachment and,

therefore, lack of oxygen supply are very probably ‘‘the factors

responsible for the slowing down of the developmental rhythm.”

66 THE AMERICAN NATURALIST [Vou. LV

The armadillo egg like that of most mammals undergoes its

early development in the fallopian tube and is, therefore, cap-

able of reaching the blastocyst stage on its initial oxygen supply.

After this time, however, it must become attached to the uterine

wall for a further source of oxygen. For some reason in the

armadillo the reaction between the blastocyst and the uterine

wall is postponed, and the blastocyst is incapable of further de-

velopmental progress until this reaction is established and the

necessary supply of oxygen becomes available. In exactly the

same way the development of the blastoderm in the fish’s egg

is experimentally retarded or stopped by reducing the available

oxygen supply and is again made to resume its development by

supplying oxygen. In the case of the fish egg, the supply of

ordinary nutriment is certainly not involved, and reactions simi-

lar to those of the armadillo egg are only obtained as responses to

changes in temperature and rate of oxidation.

In the armadillo egg I also do not believe the retardation is

of the nature of a starvation phenomenon, since we see nothing

- of the kind in other forms. Temperature changes are ruled

out, since the temperature of the uterus is more or less constant.

The absence of oxygen necessary for the energetic process of

cell division, is, therefore, in all probability the arresting cause,

and the retardation results in polyembryony.

Thus Patterson has found the developmental interruption to

exist, and he has also shown the blastocyst to be disconnected

from the uterine wall and its necessary oxygen supply during

this time. However, he has furnished no data bearing on the

reason for the delay in uterine reaction and the consequent fail-

ure of immediate implantation of the blastocyst such as normally

occurs in other mammals. However, from what is known of the

dependence of uterine reactions on conditions in the ovary (Leo

Loeb,” Stockard and Papanicolaou® and others) it may very

probably be that some peculiarity in corpora lutea formation

is primarily responsible for the entire series of reactions leading

to polyembryony in the armadillo.

The consideration of the armadillo egg up to thas point has

taken account only of the external factors influencing its mode

of development. It must now be remembered as a fact of serious

5 Leo Loeb, Jour, Morph., Vol. 22, 1911,

6C. R. Stockard and G. N. Papanieoiaon, Am, ae, of Anat., Vol. 22,

1917.

No. 636] POLYEMBRYONY IN THE ARMADILLO 67

importance that the production of quadruplets from the single

egg of the Texas armadillo is an almost constant occurrence,

while the experimental attempts to produce twins and double

individuals in fish eggs and other forms have given at best only

small percentages of such individuals among the large groups

of eggs treated. It is also a fact that all eggs do not furnish

equally favorable material for artificial twin production. The

eggs of the trout seem unquestionably more disposed to give rise

to twin formations than do the eggs of Fundulus. Thus some

eggs would seem to have a hereditary or truly innate pre-

disposition towards polyembryonic formations. There is much

reason. to believe that aside from the external factors discussed.

the armadillo egg itself is highly disposed toward the formation

of accessory embryonic buds.

There is the possibility, of course, that this natural experiment

with the armadillo egg has become so exactly regulated as to

influence the developmental processes precisely the same way

each time, yet this is highly improbable. The armadillo egg is

not a case of simple twin growths from the blastoderm, but as

Patterson finds, there are primarily two buds, and then very

promptly two secondary ones arise making the four and after

this the budding process ceases. In the South American species,

however, it would appear as though a tertiary budding occurred

giving the usual eight embryos; and in rare cases still another

budding occurs from a few of the existing buds giving a total of

as many as twelve. It would certainly seem as though the

blastoderm in these species passes through a stage of agametic

reproduction or budding of a nature unknown among other

higher vertebrates. But the possibility for such expression might

only exist on account of the delay in implantation of the blasto-

eyst and consequent shortage of the oxygen supply necessary for

the rapid formation and growth of the single embryo.

It is important to keep in mind that there are species of the

armadillo which produce only a single offspring from one egg.

It is not known whether their embryos have a ‘‘period of quies-

cence” but if they have, the period either occurs at a different

developmental stage or the eggs do not possess the inherent.

budding tendency of the other species.

We have further to acknowledge the fact that although the egg

of the deer has a ‘‘period of quiescence’’ during its development

it does not give rise with any degree of frequency to twin indi-

68 THE AMERICAN NATURALIST [Vor. LV

viduals. In the first place it is entirely uncertain from the

scanty accounts as to what time in development the quiescent

period occurs. Assuming that such a period does exist, it might

occur at some indifferent stage when no peculiar result would be

expected, for example after gastrulation, as it does in the bird

with no subsequent effect. In the light of the experimental

production of double individuals it is readily understood that

even though the egg of the deer is interrupted in its development

at an early stage, it might still be capable, on resuming develop-

ment, of giving a normal single embryo. A study of the experi-

mental production of twin and double individuals among fish

leads one to be surprised at the case of the armadillo, and to

expect the reaction found in the deer. The constant interrup-

tion occurring in the development of the birds and other animals

at indifferent developmental moments with no subsequent ill

effects, renders commonplace the fact that the deer successfully

withstands an interruption during its development without

noticeable modifications in structural response. A full considera-

tion of the different results following interruptions at critical

and indifferent developmental moments will be published in a

forthcoming number of the American Journal of Anatomy.

In conclusion we may summarize the cases as follows: The

development of the armadillo is interrupted on account of a

failure to become promptly implanted on the uterus and a con-

sequent exhaustion of the available oxygen supply. The inter-

ruption occurs at a critical period just preceding the primitive

streak and embryonic line formation. This egg appears to have

a decided tendency under conditions of arrest to form accessory

embryonic buds. As a result.of the interaction of these external

and internal forces polyembryony is produced.

In the case of the deer only one probable fact is known, and

that is that a ‘‘period of quiescence’’ occurs. It is uncertain at

what stage the arrest takes place but it is probably due as in the

armadillo to a delayed implantation of the blastocyst. Either

on account of the stage of arrest, or a lack of tendency to form

accessory embryo-buds a typically single individual arises from

this egg. The external factors may be the same as in the case

of the armadillo, but they interact with different internal factors

or different developmental moments to give a very pear rie

result.

NOTES AND LITERATURE

North American Early Tertiary Bryozoa. By FERDINAND

Canu AND Ray S. BASSLER. Smithsonian Institution. United

States National Museum. Bulletin 106. 1920, 879 pages,

279 text figures and 162 plates.

Students of both fossil and recent bryozoa will greet with

interest and pleasure this monumental work long anticipated

and recently issued, for while treating primarily of fossil

bryozoa this monograph contains much of interest to students

of living forms. This work appears in two volumes, one con-

taining the text and text figures, the other consisting of photo-

graphic plates alone. A cursory inspection reveals the fact

that these volumes possess the excellence of copious illustration,

a most satisfactory virtue in the eye of those who will use them.

The text figures are abundant, the number as stated above (279)

by no means giving a true idea of the actual number, since each

figure consists of from two to ten or more illustrations, repre-

senting portions or organs of the species under discussion, and

often besides figures of nearly related species for comparison.

From this point of view the number given should be multiplied

many times, and by actual count the first fifteen text figures

contain more than one hundred separate drawings or prepara-

tions. These are all either original with the authors or are

taken from the illustrations of other bryozoologists. Each pho-

tographic plate likewise contains from twelve to twenty-five

separate photographs. These are distinguished by a remarkable

clearness and definiteness of outline, even of minute details,

revealing an unusually skilful management of light and shade

and producing an excellent and expert piece of work which will

not fail to call forth the gratitude as well as the admiration of

their fellow workers.

Over 700 species belonging chiefly to the two orders, Cyclosto-

mata and Cheilostomata, are treated in this monograph. In the

seventy or more pages of introduction the authors present many

topics of interest involving new points of view which will doubt-

less stimulate further research. Of these topics but three will

be touched upon.

1. It is gratifying to find clear definition and illustration of

70 THE AMERICAN NATURALIST [ Vou. LV

the terms of the more recent nomenclature used in description

and classification. The study of the bryozoa has developed to

such a degree in recent years and so many new terms have been

introduced which are found only in the scattered writings of

numerous authors, that a new compilation and exposition similar

to the classic work of Hincks in his ‘‘British Marine’ Polyzoa’’

would be useful. As far as it was compatible with the limits of

this treatise such a compilation and exposition have been accom-

plished here, and both the beginner in the study of the bryozoa

and the advanced worker will find great assistance in the dis-

eriminating use of the newer anatomical terms.

f the general functions of the bryozoa, of the Cheilosto-

mata especially, the discussion of the hydrostatic function is

perhaps the most interesting because it presents the newest and

latest views on this puzzling subject. The extrusion and re-

traction of the polypide, the action of the operculum and of the

zocecial muscles in these activities, and the relation of these to

the ingress and egress of water was long a puzzle. Jullien in

1888 first discovered the so-called compensating sack or com-

pensatrix under the dorsal surface of the zocecium. Since then

scattered studies have been made on this organ which was soon

found to be present in many species. The present authors have

continued this study and following Levinsen (1909) have made

the presence or absence of a compensatrix the basis of division

of the Cheilostomata into two sub-orders Anasca, without such

a compensating sack; Ascophora, possessing such a sack.

In addition to the zowcial hydrostatic system discovered by

Jullien, the senior author, in 1915 discovered a zoarial hydro-

static system in the Anasea. This investigator found that the

space under the ectocyst, in certain species lacking a compen-

satrix, was continuous from zoccium to zocecium throughout the

colony. Into this space water is introduced or expelled thus com-

pensating for the egress and ingress of the polypides. By such

means minute creeping zoaria as the Lunulites, e.g., are enabled

to maintain themselves on the alge on which live.

3. In line with their insistance on the value of function it is

not surprising that these authors classify the Tertiary bryozoa

on a physiologic rather than a morphologic basis as is the

method followed by the older investigators. Believing as they

do that (p. 70) ‘‘In the bryozoa, as in other living beings, the

form is only the result of functions; therefore in the study of

No. 636] NOTES AND LITERATURE Ii.

morphological variations of the organs, we now substitute that

of the physiologic functions. Our studies are therefore always

directed toward the discovery of functions which modify the

skeletal form.’’ With this frank statement of the primacy of

the Lamarckian principle, family, genus and species are thus

briefly described:

The family is characterized by the larval form.

One genus differs from another in possessing a different func-

. tion. The three essential functions of all bryozoa are:

1. Passage of egg and escape of larve (— rapport of operculum

and ovicell).

2. Hydrostatic system and extrusion of the polypide (= form

of the aperture and rapport of operculum with com-

pensatrix).

3. Calcification and chitinization (—nature of the skeletal

part and of the frontal considered as deposits of the

endocyst).

Specific characters include all morphological variations and

all of the characters of adaption.

Whatever philosophic views one may hold in regard to the

relation between form and function, it is apparent that the char-

acters chosen for family, genus and species present a uniform,

logical system and constitute a good workable plan or hypothesis

which it seems more than worth while for all workers in this

field to attempt to apply. It must be remembered, however, that

while in the class bryozoa, larval characters may afford valid

data indicating relationship such characters have failed to afford

satisfactory data of relationship among some other class of ani-

mals. Granting the validity of the assumption, however, the re-

search necessary to establish this statement can be conducted

only on living species. This, then, is a matter of immediate and

pressing interest. The older bryozoologists considered the oper-

culum a family character. In the present system the operculum

becomes a generic character, changes in it being induced by

changes in the essential zoccial functions 1 and 2, that is, de-

pending in part upon the relation existing between the oper-

culum and ovicell, and in part upon the relation existing between

operculum and compensatrix together with the form of the

aperture. Here again, although much investigation has been

conducted to verify these generic principles among the fossil

Tertiary species, work on living forms should be undertaken to

12 . THE AMERICAN NATURALIST [Von LV

discover how far recent species actually conform to the plan

proposed.

Identification and description of the species of this collection

constitute the main portion of this monograph. A superficial

examination of the text reveals the fact that in order to apply

these principles to present day bryozoa, momentous and wide

spread changes will be necessary. Not only will genera and

families be broken up, but many heretofore considered widely

separate will be regarded as closely related and vice versa.

While one is startled by the number and significance of the

changes involved, yet the present morphologic method is so un-

satisfactory that this attempt to apply a unitary principle which

promises so to simplify classification and to lift it out of

chaos, should be heartily welcome. Too high praise cannot be

accorded the authors of this monograph for the excellence of

this work so full of new and stimulating ideas. ©

ALICE ROBERTSON

SEATTLE, WASHINGTON

SHORTER ARTICLES AND DISCUSSION

THE EFFECT OF YEAST ON THE UTILIZATION OF

FOOD BY WHITE MICE.:

In the present paper the question of the effect of the so-called

vitamines on basal metabolism is considered and a procedure is

indicated whereby it is believed more conclusive data may be

obtained on the question. Preliminary experiments are described

illustrating the method.

Several years ago Hopkins? in a carefully carried out experi-

ment investigated the effect of a small addendum of milk to a

diet of purified food stuffs. In brief his method was to feed in

pairs two sets of young rats of the same origin, weight, ete., on

a basal vitamine-free food and to one set give a small addendum

of milk and determine the food intake and growth increment.

By comparing the energy consumption and growth increments

of the two sets of animals and by comparing these factors on the

same set of animals after reversing the diets he was able to

show ‘‘that the small milk addendum reduced the food con-

sumption for a given weight increment to one half or less.’

In other words the vitamine increased greatly the animal’s

power to utilize its food in the production of growth. And as

he showed that this was not due to difference in absorption

from the intestines, the vitamine must have an effect on some

factor involved in basal metabolism

A number of the criticisms that could be made of Hopkins’

method have been discussed frankly by him and there is left

little or no doubt concerning his conclusions. They have, how-

ever, such an important bearing on problems in nutrition that

it seems very desirable that the question be investigated from all

angles® and particularly with other animals,‘ and with vitamines

1 Contribution from the Department of Pharmacology, Harvard Medical

School.

2 J. Physiol., 1912, 44, 425; see also Biochem, J., 1913, 7, 97.

3 Recently Eddy (J. "Biol. Chem., 1920, XXXIV; see also Am, J. Dis.

Children, 1917, 14, 189), has reported cases of marasmic stg which

showed an increase in the utilization of food, when given vitamin

prepared from beans.

4In another connection the writer has found that mice and rats react dif-

ferently with certain diets. The question is still open whether or not this

difference is qualitative or . quantitative.

74 THE AMERICAN NATURALIST [Vou LV

from different sources. In this connection, it is believed that the

` procedure to be described later by the writer gives more direct

evidence and hence is more free from criticism.

In order to eliminate the question of variability of individual

animals from the point of view of efficiency as energy trans-

formers, it appeared to the writer that a better procedure than

that used by Hopkins would be to feed a number of animals a

basal diet plus such an amount of vitamine-containing material

as to keep the animal in weight equilibrium over a period of

time, The amount of vitamine would vary with the individual,

and would need to be determined in each case. When the animals

had been maintained in weight equilibrium over a number of

days and the food consumption noted, they could then be fed the

same daily ration of a diet containing the same number of

ealories and having the same composition with the exception of

a larger vitamine content and the weight noted. As the greater

vitamine content would stimulate the appetite (directly or in-

directly), there would be no difficulty about the animals eating

the same amount of food as they had eaten in the first stage of

the experiment and they would receive the same number of

calories and a food with the same gross composition within very

narrow limits.

Due to lack of time available for this work, it was ftps bial

in the preliminary experiment to be described, to bring each

animal as near weight equilibrium as was wished, and as it is

felt very certain can be done. The average, however, for twelve

animals is close and if the results are considered from the statisti-

cal point of view, they give further very convincing evidence that

the vitamines increases the efficiency of the body in the utiliza-

tion of the food. The error due to temperature variation it is

believed is not large, but this, of course, should have been elimi-

nated. There appears no very easy way of eliminating the error

due to greater activity of the animals in the second stage of the

experiment. This error may be considerable for there is no

question that the animals were markedly more active when given

the limited diet with. greater vitamine content. The direction of

this error makes the results all the more convincing.

EXPERIMENTAL

Preparation of Diets—The diets were prepared at the be-

ginning of the experiment from the same stock of material and

No. 636] SHORTER ARTICLES AND DISCUSSION 75

placed in stoppered containers in the ice box. The composition

follows:

Diet 401 Diet 403 | Diet 405

MOR PATRE D ee oe ee 17.5% 16.5% 15.5%

Btareh ss eE sh ea 49.5% 48.5% | 47.5%

MBAR bere ee T yt 1.0% 3.0% | 5.0%

sE aa ARNOT ah AR E 18.0% 18.0% | 18.0%

Batters a Pe ore: 9.0% | 9.0% 9.0%

Re ig Oa KOPE E VE E I 5.0% | 5.0% | 5.0%

The yeast contained approximately 0.46 per cent. fat, 46.5

per cent. protein and 38.0 per cent. carbohydrate (32.26 ‘‘carbo-

hydrate” plus 5.8 crude fiber). Omitting the negligible quan-

tity of salts the gross composition of the diets was as follows:

Diet 401 Diet 403 | Diet 405

PEE LY opal 17.96% 17.88% | 17.8 %

Carbeinyate Beers CRU EN Sa 49.885 49.646 | 49.47

RE PE E E TA S Oe 27.05% 27.15% | 27.25%

Salts ee as NEERA EE E te ae ee 5. % 5. 0 ' 5. %

5 Includes 0.06 per cent. crude fiber.

6 Includes 0.12 per cent. crude fiber.

T Includes 0.3 per cent. crude fiber.

The diets richer in vitamine had slightly lower calorific value.

The casein was prepared from the 40-mesh commercial prod-

uct by shaking up for several hours successively with two por-

tions of 50 per cent. aleohol and one of 95 per cent. aleohol and

drying in warm air.

The starch was prepared from commercial cooking starch

(corn) by the same procedure as for casein.

The butter used was the clear fat obtained by decanting the

melted butter through a dry T.

The salt mixture was that described by Osborne and Mendel.’

The mice were kept in wire mesh cages under which were

pieces of glazed paper to catch feces and wasted food. The

food was placed in small salve jars having aluminum covers in

which 1% inch holes had been stamped. The jars were placed

in 5-inch glass crystallizing dishes during the second stage of

the experiment so that any waste food could later be found. by

the mice and eaten. Almost. never in the second stage of the

experiment was any food found under the cage. In all cases

8 J. Biol. Chem., 1919, 37, 572.

76 THE AMERICAN NATURALIST [Vou. LV

the waste was easily separated from the feces and account taken

of the amount.

Sixteen mice were fed ten days on a complete diet and then

nine days on a diet free of vitamines and then transferred to

diet 401 (unlimited amount). After one day on the later intake

records were begun.® Of the 16 mice started two grew nearly

normally on 1 per cent. yeast,’° one became sick and died, an-

other declined very rapidly. These four were not considered

in the experiment. The records of the remaining twelve are

shown in the table. They were of varying size and represented

somewhat different ages so that the results can hardly be ac-

counted for by the action of the normal intermittent growth

impulse.

PERIOD 1. Diet 401 (1 Per CENT. YEAST), UNLIMITED AMOUNT

| A Average

Animal Dan Ba Gia Total Intake | Total Days Daily Intake

| Besa on 15.7 gr 16.62 gr +0.92 27.15 gr 13 2.09 gr

eos 11.2 0.49 —0.71 13.61 10 1.36

Be sac 14.6 15.31 +0.71 24.7 13

pe I Lit 10.45 —0.65 16.53 13 1:27

ERR 12.4 0 +0.60 19.83 13 1.52

Sees 14.0 14.97 +0.97 21.19 13 63

Rasen oe 5.4 —1.72 19.39 13 1.49

TO 14.8 13.71 —1.09 18.25 13

| i eee ee 14.8 14.69 —0.11 21.91 13 1.68

je E ee 8.5 17.05 —1.45 27.85 13 2.14

pee es 10.2 9.12 —1.08 13.77 13 06

r SNe | 16.9 16.58 +0.68 19.97 10 1.997

| 168.6 165.67 —2.93 244.20 150

For 150 mouse days 12 mice ate an average of 1.628 grs. per

mouse per day and lost 2.93 grs. or 1.74 per cent. For mainte-

nance, then, they needed slightly more than 1.628 grs. average

per day.

Figured from Period No. 1 the mice in the 187 mouse days

should have required a little more than 187 X 1.628 grs. and

gained 14.36 grs. or 8.68 per cent.

®It would be well to increase this period somewhat,

10 In pera mice require 5 per cent. of yeast in the diet rs normal

growth; e make substantial gains on 3 per cent. and a very small per

cent. The individual variation for rats is, in our experience,

considerable though not as great as for mice. The average requirement,

too, for mice is markedly more than for rats, the aP egi roughly five

a mice to three for rats expressed in per cent, in the diet.

No. 636] SHORTER ARTICLES AND DISCUSSION 17

Perion No, 2. Dier 403 AND (or) 405 (LIMITED)

Animals | Beginnings End Change Total Intake; Total Days | hed ag:

TOR | 16.62 gr., 18.74gr.;) + 2.12 | 32.64gr. 16 | 2.04

Re ie | 10.49 1: + 1.31 | 10.95 ee eons Fr

ag 15.31 17.03 + 1.72 | 30.40 16 1.9

Sr: | 10.45 11.83 + 1.38 | 18.72 16 1.17

r a AT 13. 14.51 + 3S Ok 16 1.5

RRE A 14.97 14.05 — 0.92 | 24.00 we b

E 13.68 14.45 + 0.77 | 24.00 le oR

eee 13.71 15.25 + 1.54 | 24.00 ees E

ooo 14.69 14.65 — 0.04 | 25.6 w6 l fo

n; 17.05 18.84 + 1.79 | 35.36 w ee

E ROE 9.12 9.69 + ONT 17:00 16 1.06

Yooo | 16.58 19.21 + 2.63 | 38.00 19 2.00

| 165.67 *| 180.05 414.38 | 304.67 187

In order to make more easily comparable the weight changes

with food intake in the two periods, table No. 3 is presented.

Period No. 1 PeriodNo 2

Animal Average Daily Average Daily _

Weight Change | Food nen ec | Days | Weight Change ae zsef Days

deers +0.92 gr 2.09 gr. 13 +2.12 gr 2.04 gr 16

e ET ARE —0.71 1.36 10 +1.31 1.4 8

ee bah a +0.71 1.9 13 +1.72 1.9 16

ee, we —0.65 1,27 13 +1.38 1.17 | 16

EEN N +0.60 1.52 13 +1.51 1.5 16

Bea +0. 1.63 13 | —0.94 1.5 | 16

o SU Re E —1.72 1.49 13 +0.77 1.5 16

bt Me -1. 1.4 13 +1.54 1.5 16

tb ate ae —0.11 1.68 13 —0.04 1.6 16

Boo wes | —1.45 2.14 13 +1.79 2.21 16

Se Ee —1.08 1.06 13 +0.57 1.06 16

N +0.68 OOT 10 +2.63 2.00 19

The energy content of the excreta during the two periods was

not determined, but Hopkins™ and Drummond’? have shown that

the energy content of the excreta of rats when fed adequate

foods is substantially the same as when fed vitamine-free foods.

The writer regrets the inability to repeat the work with a

larger number of animals. It would be desirable to use a

vitamine extract, control the temperature of the animal room,

get the animals more closely in weight equilibrium by using a

vitamine percentage determined by the individual mouse, and to

11 J. Physiol., 1912, 44, 440.

12 Biochem. J., 12, 25.

78 THE AMERICAN NATURALIST [Vou. LV

have periods 1 and 2 run over the same number of days.

Among the animals there should be included several adults which

had been brought to condition of underweight in the preliminary

period with vitamine-free foods. This would answer any possible

question concerning the effect of the intermittent growth im-

pulse. It is hoped that others may be able to carry on this

work. ;

These experiments were performed undér a grant from the

Committee of the Permanent Charity Fund, Incorporated.

: R. R. RENSHAW

HARVARD MEDICAL SCHOOL, BOSTON

INTERFERENCE IN PRIMULA SINENSIS

OUTSIDE of Drosophila, the only data bearing on the question

of interference of crossing over are those which I reported in a

paper on linkage in Primula sinensis. The phenomenon of in-

terference—our knowledge of which in Drosophila dates from

crosses made by Sturtevant and analyses made by Muller in

1912—consists of the fact that the occurrence of a crossing over

in one region of a chromosome reduces the chances for the oc-

currence, in that cell, of another crossing over in a different

region of the same chromosome; thus there is a smaller number

of double crossovers than would otherwise be expected. The

amount of interference is expressed by Muller’s index ealled

‘*eoincidence,’’ which is the ratio of the proportion of double

crossovers actually observed in the experiment to the proportion

of double crossovers which would have been expected to occur on

the assumption that crossings-over in the two regions were inde-

pendent of each other; the latter, or ‘‘expected’’ proportion of

double crossovers is obtained by simply multiplying together

the proportion of crossovers in one region by the proportion of

crossovers in the other region. As I stated in my paper on

Primula, a calculation based upon my total results could not be

sufficiently reliable to decide the question of whether or not in-

terference existed in Primula. This was on account of an uncer-

tainty in the classification; I now find, however, that a calcula-

tion based upon. a selected group of the plants, in which the un-

certainty does not exist, is sufficient to decide the question in the

affirmative—contrary to my earlier conclusion.

Three pairs of genes were involved in the Primula crosses—

1 Altenburg, E., 1916, Linkage in Primula sinensis. Genetics, 1: 354-366.

No. 636] SHORTER ARTICLES AND DISCUSSION 19

those for long style (1), red flower (r), and red stigma (s),

allelomorphie respectively to short style (L), magenta flower

R), and green stigma (S). The order of the loci, as based upon

3684 individuals, was 1 r s; the per cent. of crossovers in the

first region (between 1 and r) was 11.62, and that in the second

region (between r and s) was 34.02. These relations, shown in

a map, are as follows: i 11.62 EE The per cent. of

double crossovers observed in the experiment was 2.52. Accord-

ing to the formula given above, the number of double crossovers

to be expected if crossings over were independent would be

11.62 per cent. X 34.02 per cent., or 4.0 per cent., which, as I

noted in the account of the case, exceeds the observed proportion

of 2.52 per cent. This difference, then, between the observed

and ‘‘expected’’ values in Primula would indicate that inter-

ference existed here, but, as I further stated, the difference was

not significant because of the uncertainty which had attended

the classification of flower color in the plants with the gene for

green stigma. This gene caused the flower color to be lighter

and obliterated somewhat the distinction between red and

magenta.

In the plants with red stigma, however, the flower color was

dark enough to render entirely certain the classification in regard

to red and magenta; these plants, considered alone, would there-

fore furnish reliable data for determining the interference. I

stated that, when these reliable plants alone were taken into

account, no evidence of interference was to be found; but this `

conclusion was due to a numerical error in the calculation of the

‘*expected’’? number of double crossovers, for I now find, in

going over the figures, that the ‘‘expected’’ number is consid-

erably higher than the number observed. Among the 1876

plants with red stigmas, there were 210 crossovers, or 11.2 per

cent., in the first region, and 688, or 36.7 per cent., in the second

region. The ‘‘expected’’ number of double crossovers is there-

fore 11.2 per cent. 36.6 per cent., or 4.1 per cent. There were

54 double crossovers observed, or 2.9 per cent., giving a coin-

cidence ratio of 2.9:4.1, or .7, instead of 1.00, which would be

the ratio in the absence of interference. The difference between

the ‘‘expeeted’’ and observed numbers is beyond the limits of

random sampling, and it must therefore be concluded that in-

terference exists in Primula. Although no reliance can be

8U THE AMERICAN NATURALIST [Vor. LV

placed upon the precise value of the coincidence, it may be noted

that this amount, .7, is just what would be expected for a sim-

ilar distance in the X chromosome of Drosophila melanogaster

(ampelophila).

It should be noted here that Haldane, in referring to my re-

sults in a recent article, called attention to the fact that the

Primula data (using the counts of all classes of plants) fit his

formula for expressing the relations between linkage values in

Drosophila. Inasmuch as any formula expressing the linkage

relations in Drosophila is necessarily the mathematical resultant

of the operation of interference (interference of a type which

diminishes with increasing distance), Haldane’s statement that

the Primula data fit the same formula as Drosophila is equivalent

to saying that interference exists here, as in Drosophila; it is in

this sense a restatement of my observation that the number of

double crossovers found in the total count of the plants is smaller

than the number ‘‘expected’’ in a case of a random occur-

rence of crossing over. It must further be noted that Haldane’s

formula for expressing the linkage relations in Primula is open

to the same objection of unreliability as noted above, since his

calculation is based upon all classes of plants, instead of upon

just those classes which I showed must be used in any reliable

determination.

The finding of interference in another organism, so widely

separated from Drosophila, is of interest because of the bearing

of interference on the general theory of linkage. Interference

- is not accounted for on Trow’s form of the reduplication theory,

although it was the earlier experiments upon Primula itself

which largely supplied the data upon which this theory was

ounded.

: EDGAR ALTENBURG

THE Rice INSTITUTE,

Houston, TEXAS.

ON INTERSEXES IN FIDDLER CRABS.

Nor long ago a few specimens of small female fiddler crabs,

Uca pugnas, were submitted to me by Professor T. H. Morgan

for determination. They appeared to be normal, immature in-

dividuals and I so stated in my reply. In order to demonstrate

the correctness of this view, a loan was made to Professor Mor-

gan from the National Museum collection of a series of imma-

No. 636] SHORTER ARTICLES AND DISCUSSION 81

ture female fiddlers showing different widths of abdomen normal

to the growing female in passing from the juvenile or immature,

` to the sexually mature, condition. However, this exhibit ap-

parently had the opposite effect from that intended, as Pro-

fessor Morgan pronounced them all ‘‘intersexes’’ and thereby

seemingly robbed the female fiddler of its period of adolescence.

In his paper ‘‘ Variations in the Secondary Sexual Characters

of the Fiddler Crab’ we find Professor Morgan’s exposition

of the subject. It is not easy to follow the author owing (1) to

erroneous or incomplete references to figures and (2) to ab-

sence of measurements; for example, ‘‘Fig. 4B”,’’ cited on p.

225, line 12, does not exist, ‘‘Fig. 4B’,’’ p. 225, line 3, is cited

as a female abdomen when it is really a male, and one can not

tell if the two unequal chele of Fig. 4A belong to one individual,

which is probable, and if the two unequal chele of Fig. 4B be-

long to one individual, which is probable but not possible, as the

text says that they are ‘‘of the same size.’ There are no meas-

urements nor indication of enlargement of figures.

The case under discussion belongs in Professor Morgan’s

second category of intersexes. He tells us that the specimens

are always small, that they are female in character except for

the abdomen being narrower than in the mature female, and

the abdominal appendages being different from those of the

mature female but not at all malelike. Why, one naturally asks,

are they not juvenile? His reply is, ‘‘because normal individuals

of the same size have the Re full width.’’ This argument

unsupported is fallacious.

Many species of crabs are known to attain sexual maturity

at a much smaller size than their maximum and to exhibit con-

siderable range in the size at which they attain that maturity.

As an example, two jars full of the common shore-crab of the

Pacific coast, Hemigrapsus nudus, show egg-bearing crabs rang-

ing in width of dorsum from 10.4 mm. to 32 mm., and among

the immature females with narrow abdomens, six individuals

which range from 12.5 to 15.7 mm. in width.

Professor Morgan goes on to say that some of the smallest

interacial” have the narrowest abdomen, that there is no

obvious relation between the size of the crab and the relative

width of the abdomen, but that there is some correlation between

the character of the abdominal appendages and the width of the

1 Amer. NAT., Vol. LIV, No. 632, May-June, 1920, pp, 220-246.

82 THE AMERICAN NATURALIST Vou. LV

abdomen. All these point to normal development as the rational

explanation.

He figures, p. 226, the abdomens of five female ‘‘intersexes,”’

including, I think, but am not sure, two abdomens of successive,

or near successive molts in the aquarium. As no two of these

abdomens are of the same width, the illustrations would indicate

a change in size of body, that is, growth and surely age, with

the molt or molts. But Professor Morgan says, p. 225, lines

13-14, ‘‘that the condition of the abdomen and claws had not

changed.’

The fact of the matter is, that neither Professor Morgan nor

any one else, so far as I know, is aware of the exact growth

changes of our fiddler crabs beyond the first few crab stages.

Hyman, in ‘‘The Development of Gelasimus after Hatching,’”

earries his painstaking researches only as far as a a 4 mm. wide

crab stage.

We-can at present reason only by analogy from the study of

work done on other species of crabs, of which there is altogether

too little compared to the facilities offered by the laboratories

of our coasts; and such analogy seems to indicate that the crabs

upon which Professor Morgan bases his arguments are normal

females which had not, in their particular cases, attained sexual

maturity. Churchill’s ‘‘Life History of the Blue Crab’? may

be cited, and also Pearson’s ‘‘Cancer. (The Edible Crab.)’’*

Both of these give tables which demonstrate the great variability

in the ratio of increase at each act of ecdysis.

It is important, as I have stated elsewhere, that the develop-

ment of each of our common crabs be carried through from the

egg to maturity, that accurate records be made, and properly

labeled material upon which such studies are based be deposited

in an enduring: collection accessible to all who may be interested.

Mary J. RATHBUN

VARIATION IN JUVENILE FIDDLER CRABS ©

Ir is too bad that Miss Rathbun’s ‘kindness in sending me

specimens from the National Museum ‘‘had the opposite effect

from that intended.’’ While regretting this, I can only call

attention to the fact, stated in my paper, that out of more than

2 Jour. Morphol., Vol. 33, No, 2, March, 1920,

3 Bull. Bur. Fisheries, XXXVI, Norba 11, 1919.

4 Proc. and Trans. Liverpool Biol. Soc., Vol. XXII, 1908.

No. 636] SHORTER ARTICLES AND DISCUSSION 83

three thousand individuals that were collected only a few showed

the narrow abdomen concerning which Miss Rathbun raises an

interesting question. These rare individuals, I ventured to sug-

gest, with some hesitation and with considerable reservations,

might be called intersexes, because the variation in question was

in the direction of a character peculiar to the opposite sex.

Miss Rathbun states that my sk deamon that ‘‘they are not

juvenile . . . unsupported, is fallacious.” Again I can only

repeat what was said in my paper, that I examined a very large

number of individuals, many of which were of about the same

size as the variations in question, some of which were even

smaller, and others somewhat larger, and in none of the young

females (except in those recorded as exceptions) did I find

the abdomen narrow.

May I also recall that I specifically referred to the case of the

blue crab in which the abdomen of the juvenile female is narrow,

so this condition was known to me, both from the literature and

from personal examination, although the reader might gain the

opposite opinion from Miss Rathbun’s comments. -

It is rather strange also that Miss Rathbun neglects to point

out that these small crabs with narrow abdomen were stated

in my paper to show either a change towards maleness or pos-

sibly a retention of the juvenile condition. It is quite possible,

of course, that the narrowness of the abdomen of the exceptional

individuals might be interpreted as a variation in the direction

of the juvenile stage found in other species; but it is certainly

not a common stage through which crabs of this size pass.

Whether the juvenile interpretation has any advantage over the

alternative one that I provisionally suggested can only be

settled when we have found out to what this exceptional condi-

tion in the fiddler crab is due. My paper was written more

with the intention of calling attention to a new and very in-

teresting set of variations in these crabs (as the title indicates)

than with the intention of trying to determine what the definition

of intersexes shall include; for, as I said, ‘‘It seems to me not

worth while at present to attempt to classify such material until

we have learned more about it.”

T. H. Morgan .

1 Fig. 48’, page 225, line 3, should be Fig. 5B’. Fig. 4B”, page 225, line

13, should be Fig. 50”. Fig. A, B, C, D, page 225, line next to bottom,

should be Fig. 5 A, B, C, D.

84 THE AMERICAN NATURALIST [Vou. LV

THE TURKEY AS A SUBJECT FOR EXPERIMENT

EXPERIMENTS with our native vertebrates offer many diffi-

culties not encountered when dealing with domestic animals. In

the field of genetics especially, while domestic animals continue

to furnish enticing problems, it is not strange, therefore, that

they practically monopolize the attention of students. No one

can foresee how far work of this kind will proceed but it seems

probable that some important phases of the subject of variation

never can be elucidated by the study of domestic animals alone.

Hence it would be very desirable to work with wild forms

wherever this is practicable. This would be especially interest-

ing for study of the significance of the intergrading subspecies

r ‘‘geographic race’’ which is found so widely in nature but

which appears to have no recognizable counterpart in the ordi-

nary variations of domestic animals.

The so-called subspecies perhaps needs no introduction even

to biologists who do not have first-hand acquaintance with it,

but the extent to which it features in the fauna of the world

‘seems scarcely realized even among those who are quite familiar

with it. Within the memory of the present generation, the ulti-

mate division of classification was the species and attempts to

divide this into races or varieties were often looked at askance as

probably indicating an over-weening desire to multiply names

and magnify differences of no phylogenetic significance. In

wrestling with the question ‘‘What is a species?’? many were

led to eschew classification entirely and contented themselves

with the knowledge that no two individuals were alike and the

belief that efforts to associate them were futile. Meanwhile, in

spite of headshaking in various quarters, ‘‘hair-splitting’’ has

continued until at present nothing is clearer than that the sub-

species is a reality constituting a widespread and obvious evi-

dence of active contemporary change in organisms, not only in

single individuals but in groups of individuals.

In ornithology and in mammalogy, at least, the old-fashioned

species in the vast majority of cases is found to be a composite

or a mosaic definitely divisible into units connected by graded

series and having a plain relation to geographic distribution.

A species of continental distribution in North America, for ex-

ample, may have one subspecies in the east, one in the north and

several in the south and west each occupying a limitable area

and each characterized throughout its range by certain features

No. 636] SHORTER ,ARTICLES ‘AND DISCUSSION 85

not possessed by the others. Along the geographic borders of

each subspecies will be found specimens showing varying de-

grees of intergradation so that each form merges with an ad-

joining one, or, in some cases, one in central position may merge

into several others in different directions. If, for any reason,

the ‘‘areas of intergradation’’ were rendered uninhabitable, the

various subspecies would stand as distinct well-characterized

forms presumably until they themselves began to differentiate

and separate into parts. Sometimes the difference between rec-

ognizable subspecies is slight, or sometimes it is very marked,

but when gradations through several forms are followed, char-

acters are almost always found to change to a degree far be-

yond any probability of an ontogenetic explanation. It may be

emphasized that subspecies of this sort are not the exception, but

the rule. It might almost be said that the existence of diverse

inosculating units correlated with geography is characteristic of

terrestrial vertebrates. Continued study with improved facili-

ties and increasingly comprehensive collections from all parts

of the world constantly reduces the number of forms which

are not known to break up into subspecies. To a very great

extent, the presence of an undivided ‘full species’’ in our check-

lists signifies either that it is a senescent type of limited distri-

bution or that, for lack of material or opportunity, it has not

been studied intensively. The intergrading subspecies has not

been recognized so widely among invertebrates nor in plants, but

neither entomologists nor botanists have collected and studied

their material from the geographic standpoint to such. an extent

as the ornithologists and mammalogists, so it cannot be said that

the process of change illustrated by the subspecies is not even

more widespread than appears from the study of birds and-

` mammals.

The process of formation of these subspecies, Gussie: is

going on before our very eyes in wholesale fashion and it is diffi-

-cult to believe that it is, as someone has said, merely a ‘‘shuffling

of the eards’’ which in the long run means nothing to evolution-

ary progress. Rather does it seem that it must have a physio-

logical basis, a relation to germinal change, and a large poten-

tiality for affecting the general course of evolution. Despite its.

evident importance, the intergrading subspecies is receiving but.

scant attention from experimental zoologists. With the con~

spicious exception of the very significant work being done with

86 THE AMERICAN NATURALIST [Vou. LV

white-footed mice by Dr. F. B. Sumner of the Scripps Institu-

tion,’ there seems to be little or no work under way which can

be correlated logically with the results of speciation and sub-

speciation as the field naturalist and taxonomist find them in

nature. Doubtless one of the principal reasons for this is the

difficulty of finding convenient subjects and suitable conditions

for such work. Perhaps another is the independence of workers

in the respective fields of taxonomy and experimental zoology.

As promising subjects for experiment, it seems worth while

to call attention to the American wild and domestic turkeys.

The common turkey has an exceedingly desirable distinction

from other domestic animals in that there is no important ques-

tion as to its history and lineage. Moreover, the wild stock

from which it was derived represents one of several intergrading

subspecies the natural characters and relationships of which can

be determined with a great degree of accuracy. Hence our

Thanksgiving bird, as a subject for experimental breeding,

might furnish a combination with which naturally. and arti-

ficially induced characters could be studied comparatively. As

at present recognized and understood, the native American

turkey is divisible into six races or subspecies, as follows: One

from the southeastern United States (Meleagris gallopavo syl-

vestris); one from southern Florida (M. g. osceola) ; one from

central Texas and northeastern Mexico (M. g. intermedia) ; one

from Arizona, New Mexico and Chihuahua (M. g. merriami) ;

one from the Sierra Madre of Jalisco and west central Mexico

(M. g. mexicana) ; and one from the eastern cordillera of Vera

Cruz, Mexico (M. g. gallopavo). The range of the turkey group

is thus from the southeastern Atlantic seaboard westward to

the Rocky Mountains and thence south to Vera Cruz. Complete

intergradation between the various subspecies may not be dem-"

onstrable with absolute nicety in all cases because the birds were

exterminated in certain parts of the range before any specimens

were preserved. That intergradation between all the races was

as uninterrupted as it can be shown to be between some of them,

however, is beyond reasonable doubt. The extremes of differ.

PEPOT as usual in such cases, are represented approximately

phical extremes. The characters distinguishing

the et ‘turkey of the eastern United States from that of south-

1See especially Am. NATURALIST, XLIX, pp. 688-701; ibid., LII, pp.

177-454, 1918,

No. 636] SHORTER ARTICLES AND DISCUSSION 87

ern Mexico, therefore, are clear cut and readily recognizable

without the application of any greatly refined methods. The

obvious distinction is found in the feathers of the tail and upper

tail coverts which in the United States bird are broadly tipped

with rich chestnut whereas in the Mexican subspecies these parts

are white or nearly white. Such characters, in animals of un-

known history, might easily be looked upon as produced by

mutation; but with complete gradation from one to the other

known to exist in nature, it is hard, at least for some of us, to

believe that the difference was not accomplished by gradual

rather than sudden change. If it could be shown that char-

acters of this kind behave as hereditary units without any such

blending as requires ‘‘dialectic gymnastics’’ to explain, it would

be a long step forward in the correlation of natural and man-

made experiments. Such characters are in fact heritable, as

has been shown by Sumner in his breeding and transference ex-

periments with Peromyscus. This is illustrated also by an un-

directed experiment to call attention to which is one of the ob-

jects of this communication, namely, the test of subspecific char-

acters which has been carried out in the domestication of the

turkey.

As is widely known to sportsmen, breeders, and many others,

our domestic, so-called bronze, turkey is readily distinguished

from the wild bird of the eastern United States by the coloration

of the upper tail coverts and tail. The reason for this, which is

not so generally known, is not that the domestic bird has changed

in these respects under man’s influence, but because it is the

direct descendant of the Mexican wild race which differs from

the northeastern race by these selfsame characters. Carried

from Mexico to Europe in the early sixteenth century and

thence brought to the United States, it has continued for more

three hundred generations in a new environment maintain-

ing its old established subspecific characters. To-day it may

differ from the Mexican wild race in some details, but its general

coloration is the same and especially does it retain its taxonomic-

-ally diagnostic features. These, therefore, are heritable and

doubtless related to germinal conditions which became fixed in

the wild bird. Since the characters themselves are of the kind

that appear to be produced by insensible gradations and of the

kind that frequently bear an obvious relation to environment.

an easy deduction would be that the germ plasm also has

88 THE AMERICAN NATURALIST [Vou. LV

changed gradually and, at least as a working hypothesis, one

might suppose that the germ plasm had been affected directly or

indirectly by the environment. Such an explanation is, of

course, far too simple and old-fashioned for present-day stu-

dents of evolution. It does not explain a multitude of undeni-

ably important and fascinating results of experimental work.

But neither do theories of mutation and the maze of modern

genetics explain the intergrading subspecies and perhaps there is

room for at least a little experimental work which does not deny

such an hypothesis at the outset.

In the case of the turkey, while the relatively inconspicuous

subspecific character has proved itself stable, violent saltatory

changes have been established easily. These are of the sort

common among close bred domesticated animals but so rare

among wild vertebrates that no one has yet found a case in

which they can be shown to have been perpetuated by natural

process. Thus we now have self-colored breeds of turkeys re-

spectively black, white, buff, and blue gray as well as the breed

called Narragansett in which the feathers are tipped with steel

gray. Hence it seems that the turkey may offer an opportunity

for comparative study of the hereditary behavior of characters

which have developed naturally by what seems to be continuous

variation and those which have appeared discontinuously and

been perpetuated artificially. Sumner (l. ¢., 1918) has found

with Peromyscus that hybridization of different subspecies pro-

duces in the F, and F, generations a blending of the subspecific

characters comparable to the gradations found in nature,

whereas mutant characters (partial albinism, ete.) act as simple

Mendelian units. In other words, natural subspecifie characters

act in hybridization experiments as they would be expected to

do on the assumption that they were produced by continuous

variation. Whether or not the same results would follow with

the turkey and other forms would seem to be well worth deter-

mining. In a general way, it is known to breeders that hybrids

between wild and domestic turkeys are of intermediate type,

but so far as I know careful well-controlled work has not been

done.

WILFRED H. Oscoop

FIELD MUSEUM OF NATURAL HISTORY

No. 636] SHORTER ARTICLES AND DISCUSSION 89

THE ‘‘ONE-LETTER’’ RULE FOR GENERIC NAMES IN

ZOOLOGY

RULES of nomenclature as they affect scientific names in zool-

ogy would no doubt serve their real purpose best and give more

general satisfaction if used, not in an absolute sense, but with

discretion. But there are those to whom a rule is a rule to be

rigidly applied, and the results are such that the question is

raised whether if we must abide by rules, we cannot have better

ones. A nomenclatorial rule used by the American Ornithol-

ogists’ Union has raised this question in the mind of the writer,

and attention is here called to it, not in a controversial way, but

merely to insure that the other side of the case is presented

If such questions are ever taken up again by an International

Commission it is desirable that data and opinions on the vexed

points be available for consideration. In the recently published’

‘‘ Seventeenth Supplement to the American Ornithologists’ Union

Check List of North American Birds,’’ prepared by the Com-

mittee on Nomenclature, we find the following statements relat-

ing to certain generic names:

Oxyura Bonaparte 1828 is considered preoccupied by Oxyurus Swain-

son 1827 (p. 446).

Nyctala Brehm, from whatever date k: is preoccupied by Nyctalus

Bowditch 1825, and Ægolius Kaup, 1829, is preoccupied byÆ golia

Billberg, 1820 (p. 447).

ta tag Baird 1858 is preoccupied by Bucephalus Baer 1827

(p. 446).

Dendrocopos Koch, July, 1876, is preoccupied by Dendrocopus Vieil-

lot, April, 1876 (p. 448).

But

Heteroscelus Baird 1858 is not invalidated by Heteroscelis Latreille

1825 (p. 443).

Pas Billberg (1828) is not preoccupied by Tyta Billberg (1820)

. 447).

and

Moris Leach . . . adopted because considered neither a nomen nudum

nor preoccupied by Morum Bolten, although Morus Vieillot . . . having

a termination differing merely in grammatical gender from Morum

Bolten is thereby invalidated (p. 441)

1 The Auk, Vol, 37, No. 3, July, 1920, pp. 439-449.

90 THE AMERICAN NATURALIST [Vor LV

Even the experienced taxonomist might be greatly puzzled by

this collection of apparently inconsistent assertions, did he not

turn to the Code of Nomenclature of the American Ornithol-

ogists’ Union (1908 Edition) and find the following explanatory

remark under Canon

Generic and specific names ... are to be considered identical .. .

whether the ending is masculine, feminine or neuter or in Greek or

Latin form.

In the principal codes of zoological nomenclature the practise

called for by this rule is sanctioned only by that of the Ameri-

ean Ornithologists’ Union. The parent (we may say) of the

A. O. U. Code, namely the Stricklandian Code, in so far as it

touches on the point, would seem to accept very similar generic

names, even those differing by only one letter. Section 10? says

A name should be changed which has before been proposed for

some other genus in zoology or botany.

This section is further elaborated as follows:

By Rule 10 it was laid down, that when a name is introduced which

is identical with one previously used, the latter one should be changed.

wholly coinciding with the earlier. We do not, however, think it ad-

visable to make this law imperative, first, because of the vast extent

of our nomenclature, which renders it highly difficult to find a name

which shall not bear more or less resemblance in sound to some other ;*

and, secondly, because of the impossibility of fixing a limit to the

degree of "gp EEEN beyond which such a law should cease to

con ourselves, therefore, a putting forth this

Pio poitii gear as a recommendation to naturalists, in selecting

generic names, to avoid such as too closely misura words already

adopted (p. 118).

These provisions were adopted (with a reservation as to botani-

cal names) by the British Association for the Advancement of

Science in 1865 as part of a code which more than any other

guided the course of subsequent nomenclature practice.

2 Rep. British A. A. S., 1842 (1843), p. 113.

3 If this was true in 1842, how much more difficult the situation must be

now after 80 additional years of taxonomic activity,

No. 636] SHORTER ARTICLES AND DISCUSSION 91

In Dall’s ‘‘Discussion of the Subject of. Nomenclature” of

1877 which was based on a circular responded to by 45 American

naturalists in addition to previous codes and other publicatiéns

on the subject, the point under consideration receives the foilow-

ing attention in Section 65, Paragraph 10,

When a name is identical, when properly spelled according to a

derivation given by its nubhot with a prior valid name in the same

kingdom it must be rejected.*

In other words, if names are not identical they stand.

The Entomological Code (1912, Paragraph 82) has this to

say on the subject:

A generic or ARONS name is a homonym and subject to replace-

ment when it is spelled exactly like a previous valid generig or sub-

generic name, letter for letter. However, I and J, and Eu and Ev at

the beginning of a name are considered the same, and other words that

are equivalent in established Latin usage.

In extracts from a code of Nomenclature in Ichthyology (Jor-

dan, Evermann and Gilbert) published in the Condor in 1905,

Canon XVII (Second paragraph), is quoted as follows:

As a name is a word without necessary meaning, and as the names

are identified by their orthography, a generic name (typographical

errors corrected) is distinct from all others not spelled in exactly the

same way. Questions of etymology are not pertinent in case of adop-

tion or rejection of names deemed preoccupied. (Note.) This canon

prohibits change of names because prior names of similar sound or

etymology exist. It permits the nse of generic names of like origin

but of different genders or termination to remain tenable.

The International Code which, so far as it goes, is adhered to

by a majority of zoologists, alludes to this subject in a recom-

mendation under article 36. The language follows:

It is well to avoid the introduction of new generic names which differ

from generic names already in use only in termination or in a slight

variation in spelling which might lead to confusion. But when once

introduced, such names are not to be rejected on this account.

4 Nomenclature in Zoology and Botany, Salem, Mass., December, 1877,

p. 49.

92 THE AMERICAN NATURALIST [Vou. LV

Opinions 25 and 34 of the International Commission support

the wording of the foregoing recommendation which is referred

to in the opinions as an effective part of the code.

Thus zoological codes in general support the so-called ‘‘one-

letter rule.’’ The point in this connection that appeals to the

present writer with special force is that there would seem to be

no good defense for the practise of rejecting names differing

in terminations expressing gender and at the same time accept-

ing other names differing by no greater margin (often by only

one letter).

Thus under A. O. U. practise Otostomus, Otostoma and Oto

stomum are treated as identical, while Odostoma and Otostoma,

Tcteria and Icterias, Pica and Picus are considered distinct. The

fact that the latter words had different terminations, or different

meanings in classical usage has nothing to do with the case.

Nomenclature is not the Latin language; it is a mass of in-

vented, adopted, derived and compounded words, some of which

are in Latin form, others not, but all of which, nevertheless, have

equal standing in the scientific world. Principle V of the A.

O. U. Code, itself, asserts that

A name is only a name, having no meaning until invested with one

by being used as the handle of a fact; and the meaning of a name so

used in zoological nomenclature, does not depend upon its signification

in any other connection.

Literally construed this principle is fully in accord with the

definition of scientific names as arbitrary combinations of letters,

and it would seem unnecessary even to state with respect to

arbitrary combinations, that we can only regard each different

one (even if by only one letter) as a distinct name. It would

seem clear, therefore, that in scientific nomenclature names are

merely labels for conceptions; that their use demands precision,

and with precision all names appreciably different can be used

without confusion.

Small (even one-letter) differences in scientific names are by

no means confined to terminations; they occur in all points in

words. Consider: Neothripa, Neothrips; Felicea, Felicia, Dona-

tia, Donacia; Isotoma, Isosoma; Leptopora, Leptoprora; aah ea

Merciera; Teliocrinus, Teleiocrinus; Sciurus, Seiurus; Sus, M

Consider also such a series of names as Monocerus, M e

No. 636] SHORTER ARTICLES AND DISCUSSION 93

Monocercus, Monocercis. These names all stand under the A. O.

U. Code, as do also words like the following: Rolanda, Rolandra;

Oga, Ogoas; Orophia, Orophila; Menida, Menidia; Lyria,

Lyrcia; Passerina, Passerita; all of which differ only in the last

few letters as do those with terminations denoting gender, and

are equally liable to confusion by typographical errors.

What justification is there for accepting names so nearly alike

as many of these but considering as homonyms such terms as

Nyctala and Nyctalus; Nettion and Nettium? The aim of codes

of nomenclature is to conserve names, not to make opportunities

for the creation of new ones. But the A. O. U. custom of con-

sidering homonyms, names differing in terminations indicating

gender is a breeder of new names. This is clearly shown by two

notes’ published in a recent number of The Auk, in which it is

asserted that Phwochroa Gould 1861 is preoccupied by Pheochr-

ous Laporte, 1840, and Elminia Bonaparte 1854 by Elminius

King, 1831, and a new name is proposed in each case. The

same criticism applies to certain other suggestions in connection

with Canon XXX of the A. O. U. Code, namely those that would

homonymize such words as Athene and Athena; Contopus and

Contipus. Those who look with favor on homonymizing words

whether they differ only by endings denoting gender, whether

the root is taken from the Attic or other dialect, whether the

connecting vowel of compound words be a, i, or o, or for other

philological reasons should remember that there is no more

reason for stopping at one point than another in the path of

purism. Always there will be more and more advanced purists,

who would sink generic names differing far more widely. For

instance, consider the following pairs of names for which it has

actually been proposed that the second name in each couplet be

regarded as homonym: Callitriche, Calothrix; Myosuros, My-

urus; Galarhoeus, Galactorheus; Korycarpus, Corythrocarpus;

Tonactis, Iactis ; Genyscoelus, Coelogenus.

Philology .is an interesting and important science, but what

has classical purism to do with a hodge-podge of names such as

zoological nomenclature now is, with names coined, with names

classical, with those borrowed from nearly every language ancient

and modern? What would be the fate of nomenclature if the

purist were allowed to work his will with such namés as: Abudef-

5 The Auk, Vol. 37, No. 2, April, 1920, p. 295, and p. 302.

6 A very few reject words of similar sound—phononyms,

94 THE AMERICAN NATURALIST [Vou. LV

duf, Avahi, Aye-aye, Bagre, Cachalot, Djabub, Grysbock, Jafar,

Jukaruka, Kahavalu, Louti, Mabuya, Maki, Ompok, Potto,

Sandat, Sheltopusik, Tlja, Susu, Wallago, Zingel and the like?

Or with such personal and local derivatives as: Amiskwia,

Ernestokokenia, Ischikauia, Mitsukurina, Mordwilkoja, Schlag-

inhaufenia, Takakkawia, Wankowiczium, Wlassicsia and Zschok-

keella?

The writer does not defend the choice of such names, but once

on record they are an integral part of nomenclature and an out-

burst of purism sufficient to do away with them will not occur.

Whether we will or no, we are dealing with essentially arbitrary

combinations of letters arbitrarily selected. The conglomeration

of generic names in zoology, may be, nay is, subject to criticism,

but it exists, is in use. It is part and parcel of the language of

Science and classical purism can no more be applied to it than

to any other modern language which is constantly growing, ever

adding to itself terms from a multitude of sources.” A condi-

tion not a theory confronts us; practicability must reign and

pedantry be forgotten.

Practically all rules relating to the validity and priority of

generic names ho some saving clause as ‘‘typographical ee

corrected,” or ‘‘except for obvious typographical errors.” A

common-sense E Soci of such clauses would do away with

the most vexatious cases of emendation, cases often cited to

show the necessity of homonymizing similar generic names,

namely those in which an author mis-spells names of his own

establishing when using them subsequently to the original cita-

tion. In such cases why can we not take an author at his word;

he intended to treat of the same group as before, and his emended

name, whether intentional or not should be regarded as a

synonym of the original. We do not recognize an author’s

efforts to change a published name, except to correct typo-

graphical errors. Why should we give any weight to emenda-

tions which themselves, in many if not most cases, are almost

certainly typographical errors. The same rule should apply to

names mis-speélled by others than the original author when it is

clear they intended to refer to the same genus. The fact that

the species included under such names are now considered to

belong to different genera is of no consequence; these genera

7 Thus we adopt into English but do not agai such words as hangar,

machete, fez, mufti, a host of which could be cite

No. 636] SHORTER ARTICLES AND DISCUSSION 95

should date from the time formally recognized and should bear

the name then given. It is a travesty on priority to credit an

author with conceptions he never entertained, and to use for

them mis-spelled names for which he no doubt often had occa-

‘sion.to regret his inadvertence. In brief, regard all emendations

as typographical errors unless there is definite evidence to the

contrary. With the treatment suggested, such cases as Pogonius,

Pogonias, Pogonia (a name spelled three ways in the same publi-

` cation), and similar instances lose their troublesome aspect, and

suggestions for homonymizing them, much of their force. The

chief cause for anxiety in connection with the one-letter rule

seems to be that numerous emendations may be revived, but it

can confidently be asserted that, from a practical viewpoint, most

emendations are clear synonyms from the beginning and their

status would not be changed under the one-letter rule.

Moreover changes under this rule need be feared only in

branches of zoology in which the practice advised by the A. O. U.

Code has been followed, that is the study of birds and mammals,

Certainly the one-letter rule has been used, since the adoption

of the International Code, if not before, by most American stu-

dents of animal parasites,’ echinoderms, crustaceans, insects and

fishes? and as shown in preceding paragraphs their practice’?

in this respect is overwhelmingly supported by the various zoolog-

ical codes.

ee discussions by Ch. Wardell Stiles (Zool. Jahrb., 15, 1902, pp. he

His ‘t The difference of a single letter, entirely regardless o of the e

ogy, excludes the possibility of the words being identical, hence sead os

ee oth their being homonyms’’ (PP 172-1

in Jordan and Evermann, ‘‘The Fishes of North and Middle

hice Vol I, 1896, P. vy, o con regard all generic names as different

unless originally spelled alik

10 An attempt to develop gre usage, in this respect, is followed in a

larger number of zoological specialties, was made by mailing a brief ques-

tionnaire to 30 systematic zoologists. The questions asked were:

1. In your specialty have one-letter differences been regarded in recent

years (at least since adoption of International ae as sufficient to estab-

lish the distinctness and validity of generic nam

2. Or has the ruling of American Oraii" Union Code relating to

homonymizing terms differing only in endings indicating gender, ete., been

followed?

Only 17 replies were received, of which 2 were noncommital, 9 reported

no established usage and those which indicated adherence to one or the

other of the opposed practise numbered 3 in each case. The result of this

mail test at least supports the seca ’s contention that the subject is one

ripe for publie discussion.,

96 THE AMERICAN NATURALIST [Vor. LV

Since one-letter differences in generic names are sufficient in

many cases as shown by citations in this article and the practise

of nomenclators, why are they not in all? The one-letter rule is

practicable, while one based on grounds of classical purism is

not, and as the framers of the Ichthyological Code properly

remark: 7

If all names are regarded as different unless spelled alike, these

matters offer no difficulty. Any other view gives no assurance of

stability.

Finally, discarding names of independent origin and distinct

application, that are not spelled identically, overthrows the law

of priority and like all practises of that tendency (so long as

the priority system is followed) is not for the lasting good of

scientific nomenclature.

W. L. McATEE

U. S. BIOLOGICAL BURVEY

THE

AMERICAN NATURALIST

VoL. LV. March-April, 1921 No. 637

IMMUNE SERA AND CERTAIN BIOLOGICAL

PROBLEMS!

PROFESSOR M. F. GUYER

DEPARTMENT OF ZOOLOGY, UNIVERSITY OF WISCONSIN

Mr. President, Members of the Academy of Medicine

of Cincinnati: I am deeply appreciative of the honor con-

ferred upon me by the invitation to address this medical

society. Although my own researches have lain outside

the conventional limits of medicine, it happens that sev-

eral of them have crossed the border lines of this science

and have thereby quickened the naturally keen interest

in the scientifie aspects of medicine that I have always

. entertained. In my present researches, indeed, I have

borrowed some of my most important tools and ideas

from the field of immunological studies and discoveries,

made in the main by medical researchers. The luxuriant

growth of literature in recent years on immunity, anti-

toxins, cytotoxins, bacteriolysins, hemolysins, opsonins,

precipitins, agglutinins, anaphylaxis, and what not, is

known to you all. Naturally, the brilliant series of prac-

tical applications of this new knowledge in diagnosis,

prophylaxis and therapeutics, stimulated every medical

investigator to redoubled effort until the field has become

almost exclusively the domain of the bacteriologist and

the pathologist.

It may seem presumptuous of me, a biologist, to step

outside the traditional bounds of my science and to come,

carrying coals to Newcastle as it were, in recounting to

you various facts long since learned by physicians—

1 An address delivered before the Academy of Medicine of Cincinnati,

98 THE AMERICAN NATURALIST [Vou. LV

facts which lie at the very foundation of modern medical

theory and practice. I offer my apology in advance for

the lack of novelty in much of what I shall say. My only

justification is that a reconsideration of such familiar

knowledge gives one a good running start, so to speak,

for a leap into less known realms; realms of great inter-

est to the embryologist, the cytologist, the student of

heredity and of evolution; regions in which lie hidden

the secrets of all life and form, of hereditary transmis-

sion, and of its converse, variation.

It is clear that the phenomena which constitute the

field of immunology, although to-day viewed mainly from .

the standpoint of infection and immunity, all have

broader biological aspects. They must in last analysis

be but heightened or specialized reactions of the funda-

mental processes which underlie all life phenomena.

They are but one of the many expressions of that deli-

cately balanced stereochemical system we call proto-

plasm, and they are inextricably interwoven in the ebb

and flow of metabolism, with such fundamental biologic

processes as growth, reproduction, irritability and adap-

tation.

The physiologically minded biologist also inevitably

suspects close relationship between the reactions de-

scribed by the serologist and those manifested normally

in a living animal by that wonderful system of chemical

messengers or internal secretions, the hormones and

chalones, which, independently of the nervous reflex, can

stimulate or inhibit the activity of some organ in a part

of the body far distant from the source of the secretion

itself, and which undoubtedly play an important part in

development. There seems no reason to doubt that both

hormones and antibodies, for example, represent com-

plexes of atoms which were originally parts of body-

cells concerned in the normal metabolic processes. One

is extruded into the body fluids under the influence of a

usual and therefore normal stimulus, the other is the

product of an accidental stimulus resulting from disease

or other unusual condition.

No. 637] IMMUNE SERA 99

In any event this whole field of endocrinology and

serology stands as a perpetual challenge to the experi-

mental biologist. Some sixteen years ago Nuttall pub-

lished his remarkable series of studies on ‘‘Blood Im-

munity and Blood Relationship’’ in which he reported

the results of his examination of some nine hundred dif-

ferent samples of blood from various kinds of animals.

He demonstrated that by the precipitin test a differential

scale of actual blood relationships among animals can be

established. As you doubtless recall, when an animal of

one species is injected parenterally with successive doses

of blood-serum of another species over a period of a few

weeks, the blood-serum of the injected animal acquires

the ability to form a precipitate with that of the alien

species when the two sera are mixed. When the reaction

is carried on in vitro, even in dilute solutions, the cloudi-

ness and ultimate flocculation which results are easily

seen. If, for example, a rabbit is thus repeatedly in-

jected with human blood its blood-serum when mixed

with slightly diluted human blood-serum in vitro will

almost instantly yield a noticeable precipitate, though a

control mixture of human blood-serum and the blood-

serum of an untreated rabbit will remain clear. The in-

gredient which has been engendered in the serum of the

rabbit is termed a precipitin, and the foreign serum which

was injected—human blood-serum in this case—is called

the antigen, or more specifically, the precipitinogen. It

is known that not only blood-serum, but also milk, glob-

ulins, various albumins and bacterial products—in fact

probably any foreign protein—may serve as antigen for

the formation of precipitins. The reaction is not abso-

lutely specific in low dilutions since species of animals

related to the one from which the antigen was taken will

also, though in less degree, give the effect. Closeness of

relationship is determined by finding the dilution in which

the serum to be tested will react. For instance, Nuttall

found that when rabbit serum which has been sensitized

against human serum is mixed with the moderately di-

100 THE AMERICAN NATURALIST [ Vou. LV

luted sera of man, apes and monkeys, respectively, it re-

acts to all, though in a varying degree. When mixed with

more highly diluted sera from such animals, it forms a

precipitate only with the serum of man and the manlike

apes (chimpanzee, orang-outang, gorilla), the chimpanzee

standing nearest to man. Absolute specificity may be

obtained if the antigen is sufficiently diluted. On the

basis of extensive experience, Uhlenhuth sets a dilution

of antigen of 1 to 1,000 as a standard beyond which no

precipitation will occur except with the specific antigen

employed in the sensitization. |

Thus the precipitin test became useful to the zoologist

in discriminating between different species, and it may

prove of importance in establishing the taxonomic posi-

tion of new forms, or in confirming or changing the classi-

fication of groups already known. The delicacy of the

test is remarkable. A properly sensitized serum may

give a reaction with blood diluted 20,000 or even 50,000

times. Sera have been obtained, indeed, in which specific

antigen could be detected in a dilution of 100,000. When

one recalls that ordinary chemical tests cease to give

detectable reactions in blood diluted 1,000 times, he can

appreciate the value of these physiological methods of

measurement to the biologist. They apprise him of

species differences between the proteins of various ani-.

mals which can not be determined by any known chem-

ical methods.

The value of the precipitin test in forensic medicine,

in determining the nature of blood stains on clothing,

weapons or other objects, is well known to all of you, as is

doubtless their utilization in meat inspection, such as for

the detection of horse-flesh or dog-flesh in sausages or

other chopped meats, and in various other types of

adulteration.

One thing that interests the biologists greatly in the

precipitin reactions is the fact of so-called ‘‘species spe-

cificity’’—the fact that blood sensitized against one tis-

sue of a given foreign species will react with extracts of

No. 637] IMMUNE SERA 101

_ the other tissues of that species. Thus the blood-serum

of a rabbit which has been treated with sheep blood-serum

will form a precipitate not only with the sheep serum, but

with the extracts of sheep muscle, sheep liver, sheep

spleen, and other organs of the sheep. This clearly im-

plies that each species of animal possesses something in

common throughout all its tissue proteins, something

peculiar to that particular species which in last analysis

must be resolved into a problem of its general metabolism

and stereochemistry. This does not mean that organs

may not also have protein complexes peculiar to them-

selves. Indeed, it is an established fact that they do.

And what is more, some of these organal peculiarities

may be common to various species. For example, the

fact of ‘‘organ specificity’’ has been established for the

crystalline lens. According to Uhlenhuth, immunization

with crystalline lens of a given species of animal yields a

precipitin which reacts with the lens proteins of many

different species of animals. Von Dungern and others

have secured similar results with proteins derived from

the testis. Confirmatory evidence of this fact that a

type of specificity attaches to the nature of the organ

itself, irrespective of species, has also been established

by means of the reaction of anaphylaxis.

The precipitin reactions, then, teach the biologist that

in the chemistry of the general proteins of a given ani-

mal, there are certain fundamental similarities, also that

there are constant species differences between the homol-

ogous proteins of different species of animals, and lastly,

that some proteins, in certain highly specialized organs

at least, though existing in different species, possess

similar chemical characteristics.

These and related facts when considered in conjunction

with such as those of Reichert and Brown regarding the

stereochemical correspondences in the living matter of

allied species as demonstrated in the crystallography of

their hemoglobins, or the studies of Reichert on the rela-

tions of the starches and tissues of parent-stocks to those

102 THE AMERICAN NATURALIST [Vou. LV

of hybrid-stocks in plants—such facts taken all together

are gradually constructing for the biologist a rational

biochemic basis for the study of the fundamental proc-

esses operative in metabolism, heredity and evolution.

- But let us now turn our attention to another type of

serological reactions, those concerned with the cytotoxins

or cytolysins. You doubtless all recall the well-known

experiments of Bordet, in 1898, in which he found that

the blood of guinea-pigs which have been repeatedly in-

jected with the red blood corpuscles of the*rabbit, acquire

the property of rapidly dissolving rabbit corpuscles.

This is the familiar phenomenon of hemolysis, and the

substance in the blood-serum of the guinea-pig which

brings about solution of the red corpuscles of the rabbit

is termed a hemolysin. Bordet showed further that this

enhanced solvent action of the serum of animals treated

with the red blood cells of a different species exists only

for the kind of red corpuscles used as antigen, not for

those of other species of animals. Exceptions occur,

though in the main the reaction is specific. The similar

facts regarding bacteriolysins, which are now, common-

places of every-day medicine, had been established a

year earlier.

It was soon discovered that other materials such as

leucocytes, nervous tissue, spermatozoa and crystalline

lens, when injected into the blood of a foreign species

will form lytic or toxie substances more or less specific

for the antigen used in the immunizing process. While

it is probable that none of such cytotoxins or cytolysins

acts only upon its own antigen—all studied so far have

been found to be somewhat hemolytic—the important

fact, for our present purposes, is that although a par-

ticular cytolytic serum may affect some other tissues, it

vigorously attacks the special tissue used as antigen.

This fact, when fully grasped, suggests inevitably to

the biologist, or at least to the investigator interested in

the mechanism of heredity and variation, queries such as

the following: if a special serum can thus be constructed

No. 637] IMMUNE SERA 103

which will single out and destroy a certain element of an

adult organism, is it not possible that there is sufficient

constitutional identity between the mature substance of

that element and its representatives in the germ-cell that

they too will be influenced? Is this not a way of getting

at the old yet ever new problem of the inheritance of

body acquirements, or at least of breaking in on the

germ? Is it not possible to secure selective action on

certain parts of the developing embryo and thus shed

some light on the genesis of congenital abnormalities?

And by using the cytolytic and other immunologic

methods may not additional knowledge be gained con-

cerning the relations of mother and fetus?

Of this series of problems the one which tantalizes the

biologist most of all, perhaps, is that concerned with the

possible hereditary transmission of characters acquired

directly by the body of a parent. As you know, this has

been a bone of contention for many years. The so-called

Neo-Lamarckians follow, at least in a modified form, the

teachings of Lamarck to the effect that such ‘‘acquired

characters’’ are or may be inherited; the other school,

often called Neo-Darwinians, strenuously deny such in-

heritance, and assert that the sole font of specific change

lies in the germplasm. According to them any new in-

heritable feature which appears first arises in the germ

and only finds somatic expression when this germ devel-

ops into a body.

How important he considered the correct solution of

this problem is shown in the following statement of Her-

bert Spencer. He said: ‘‘Concerning the width and depth

of the effects which the acceptance or non-acceptance of

one or the other of these hypotheses must have on our

views of life, the question, Which of them is true? de-

mands beyond all other questions whatever the attention

of scientific men. A grave responsibility rests on biolo-

gists in respect of the general question, since wrong

answers lead, among other effects, to wrong belief about

social affairs and to disastrous social actions.”

104 THE AMERICAN NATURALIST [Vou. LV

Lamarckism at the present time, among American

biologists, has all but disappeared. Some paleontolo-

gists, who in reading the records of the past find that

whenever new conditions for existence occurred, new

forms of life admirably adapted to those conditions ap-

peared, are prone to believe that the environment has in

some way directly molded these new inhabitants to its

bounds. Since this performance has occurred again and

- again, they are a bit skeptical of the selectionist tenant

that each occasion has had to await, not only the acci-

dental occurrence of a favorable germinal variation, but

of a host of them, which must in turn be sifted and par-

celed and perfected by natural selection into that adapt-

edness to the surroundings which characterized the or-

ganisms in question. Various students of geographical

distribution also are inclined to regard the direct action

of environment as instrumental in molding the fauna of

a given region. In brief, those who look at the problems

of evolution from wide perspective tend to postulate that

altered function or environment, if long continued, in

some way modifies descendants, but they don’t tell us

how. ‘Those who view the problem from the standpoint

of the few generations intensively studied by the geneti-

cist, or from the germ-cell lineages of the embryologist,

or the chromosomes of the cytologist, almost without ex-

ception reject the Lamarckian interpretation. And it

can not be denied that the latter have an incomparable ad-

vantage in directly testing the matter, since they have

their material in hand for direct observation or experi-

mental control. So it has come about that the believer

in Lamarckism, silenced if not convinced by the formid-

able array of negative evidence amassed against him,

and still more perhaps by his own inability, from the

basis of carefully controlled experiments, to cite specific

examples of inherited somatic acquirements, has sub-.

sided into mute acquiescence or but faint-hearted advo-

eacy of his theory. :

The fertilized egg develops into an adult individual

No. 637] IMMUNE SERA 105

through a series of cell-divisions and specializations of

the new cells thus produced.. During development cer-

tain cells are set apart, often very early in embryogeny,

for reproducing the next generation. Thus the germ-

cells and the body-cells of a given organism develop at

the same time and neither is the product of the other;

each alike has originated by division from the fertilized

ovum. ‘There is no necessity, therefore, for collecting

samples from all parts of the body and concentrating

them in germ-cells, as Darwin supposed was done, for

the samples are already there, derived from the same

supply that produced the parental body. They exist not

in the form of such parts of an organism as are visible

to us, but simply as certain ingredients which when com-

bined in certain ways and developed in certain directions

give rise to the parts in question. Sooner or later the

body dies, but in the meantime one or more of the germ-

cells have passed on to become expressed as new bodies

and new germs. Thus a child does not inherit its char-

acteristics from corresponding characters in the parent-

body, but parent and child are alike because they are

products of the same fundamental materials.

How, indeed, can a change in a brain-cell or a muscle-

cell find expression in a germ which is itself a cell that

possesses neither brain nor muscle? How can an influ-

ence at a distant part of the body even reach a germ-cell?

How can immature young, even larve in some instances,

produce young which ultimately come to manifest the

characteristics of the adults of the species? How can

recessive Mendelian unit-characters disappear, perhaps

for generations, to reappear at last apparently with

qualities undimmed? How, on the Lamarckian basis of

use-inheritance, can the highly specialized characters of

the worker-bee have originated and become perfected

when the individual itself is sterile? How account for

adaptive characters based on passivity, or for mutual

adaptations such as may exist between plants and certain

animals? ‘These and a host of questions like them con-

106 THE AMERICAN NATURALIST [Vou. LV

front the Lamarckian when he strives to resuscitate the

faith that is in him.

_The opponent of Lamarckism certainly shines as a dis-

coneerting questioner. Moreover, he is clearly correct

in his contention that the idea of germinal continuity is

the simpler one, and probably the only tenable one, as

regards the inheritance of characters, once they have

been engendered. But the crux of the whole problem

lies in the question, where do new characters come from?

According to the followers of the great biological theo-

rist, Weismann, not only do new heritable characters

originate in the germ, but a change which first appears

in the body can not in any way become incorporated in

the germplasm. Unquestionably, constitutional changes

in a germ-cell at any time may find expression as a new

or modified character in the subsequent organism which

comes from this germ. But while this is an obvious fact,

it gives no real explanation of the origin of the character

in question, since it tells us nothing about what induced

the constitutional change. Weismann regarded sexual

reproduction, the intermingling of two lines of germ-

plasm, as an important cause of germinal variation, but

our modern genetical studies indicate that this is prob-

ably not true. Dual ancestry, of course, makes possible

new arrangements of germinal constituents which reveal

themselves in new combinations of characters, but the

germinal antecedents of such combinations are unitary

in nature, and there is no evidence that sexual mixture

originates any new units. So the Neo-Darwinian, al-

though highly successful in pointing out the shortcomings

of Lamarck, has been little if any more successful in ex-

plaining satisfactorily how changes are initiated in the

germ-cell. Yet it is this very item of change, of varia-

tion, that is the real basis of evolution.

Some selectionists glibly assert that new characters

arise as the result of spontaneous changes i in the germ.

What is meant by this? Just what is a spontaneous

change? No one has ever succeeded in telling us. And

No. 637] IMMUNE SERA 107

we may suspect, though perhaps it is heresy to do so,

that it is a well-sounding phrase that is the equivalent of

the three words, I don’t know. Unwilling to admit of

the modifying influence of external agencies on the germ,

such theorists resort to the fiction of a spontaneous

change. Coleridge somewhere has said ‘‘What’s gray

with age becomes religion.’? We have toyed so long with

this idea of germinal continuity and the invulnerability

of the germ, that it has become for some of us well nigh

sacrosanct. Living matter is living matter wherever it

may be found, but when it happens to be in the germ-

cells, verily, ‘‘this corruptible has put on incorruption

and this mortal immortality’’!

Now, no one to-day, qualified by his knowledge of em-

bryology and genetics to the right of an opinion, would,

I think, deny that the new organism is in the main the

expression of what was in the germ-line, rather than of

what it got directly from the body of its parents, but does

this fact necessarily carry with it the implication that

the germ is insusceptible to modification from without?

Is not the serum of organisms with blood or lymph an

excellent medium through which external influences may

operate upon it? Is it not more reasonable to postulate

the origination of germinal changes through some such

mechanism as this than to attribute it to mysterious

‘*spontaneous changes’’?

With such thoughts in mind I and my research asso-

ciate, Dr. E. A. Smith, set about making various tests.*

Without attempting to tell you of our as yet unsuccess-

ful attempts to secure cytolysins which will operate in

the developmental stages of such periodically renewed

structures as feathers, or to weary you with the history

of our various other failures—of which there are an

abundance—I wish to speak briefly about certain ante-

natal effects we secured in rabbits by means of fowl-

serum sensitized against rabbit crystalline lens, and of

the fact that such induced defects may become heritable.

2 Jour, Exp. Zool, XXXI, 2, Aug., 1920.

108 THE AMERICAN NATURALIST [ Vor. LV

The crystalline lens of the rabbit was selected as

antigen, and fowls as. the source of the antibodies. The

lenses of newly killed rabbits were. pulped thoroughly

in a mortar and diluted with normal saline solution.

About four cubic centimeters of this emulsion was then

injected intraperitoneally or intravenously into each

of several fowls. Four or five weekly treatments with

` such lens-emulsions were given. Then a week or ten

days after the last injection the blood-serum of one or

more of the fowls was used for injection into pregnant

rabbits. The rabbits had been so bred as to have the

young advanced to about the tenth day of pregnancy,

since from the tenth to the thirteenth day seems to be a

particularly important period in the development of the

lens. It is then growing rapidly and becomes surrounded

by a rich vascular network that later disappears. From

four to seven cubic centimeters of the sensitized fowl-

serum were injected intravenously into the pregnant

rabbits at intervals of two or three days for from ten

days to two weeks. Several rabbits died from the treat-

ment and many young were killed in utero. Of sixty-

one surviving young from mothers thus treated, four

had one or both eyes conspicuously defective and five

others had eyes which were clearly abnormal. It is pos-

sible that still others were more or less affected, since

we judged only by obvious, visible effects. We found

later in some of the descendants of these individuals that

rabbits which passed for normal during their earlier

months subsequently manifested traces of defects in their

lenses or in other parts of the eye.

The commonest abnormality seen in both the original

subjects and in their descendants was partial or com-

plete opacity of the lens, usually accompanied by reduc-

tion in size. Other defects were cleft iris, persistent

hyaloid artery, bluish or silvery color instead of the

characteristic red of the albino eye, microphthalmia and

even almost complete disappearance of the eyeball.

. Taking into account the method of embryological devel-

No. 637] IMMUNE SERA 109

opment, however—the relation of lens, optic cup and

choroid fissure—the defects are probably all attributable

_ to the early injury of the lens. In some cases, both

among originals and descendants, an eye microphthalmic

at birth may undergo further degeneration such as col-

lapse of the ball and what appears to be a resorption as

if some solvent were operating upon it. The eyes of the

mothers apparently remained unaffected. This is prob-

ably due to the fact that the lens tissue of the adult rab-

bit is largely avascular and therefore did not come into

contact with the injected antibodies.

That the changes in the eyes of the fetuses resulted

from the action of lens antibodies is indicated by the fact

that in not one of the forty-eight controls obtained from

mothers which had been treated with unsensitized fowl-

serum or with fowl-serum sensitized to rabbit tissue

other than lens, was there evidence of eye-defects, and I

may add, that among the hundred or more young ob-

tained later from mothers which were being experimented

upon with various types of sera or protein extracts, for

other purposes, not a single case of eye-defect has ap-

peared.

As already stated, once the anomaly is secured it may

be transmitted to subsequent generations through breed-

ing. So far we have succeeded in passing it to the eighth

generation without any other than the original treat-

ment. The imperfection, indeed, tends to become worse

in succeeding generations and also to occur in a propor-

tionately greater number of young. ‘Though not anal-

yzed completely as to its exact mode of inheritance, it

has in general the characteristics of a Mendelian reces-

sive. Like such anomalies as brachydactyly or poly-

dactyly in man, the transmission is not infrequently of

an irregular, unilateral type, sometimes only the right,

at others only the left eye showing the defect. In the

later generations, probably in some measure as the result

of selective breeding, there is an increasing number of

young which have both eyes affected.

110 THE AMERICAN NATURALIST [Vou. LV

To determine whether the reappearance of the defect

was due merely to the passing on of antibodies or kindred

substances from the blood stream of the mother, or to

true inheritance, we mated defective-eyed males to nor-

mal females from strains of rabbits unrelated to our

defective-eyed stock. The first generations produced in

this way were invariably normal-eyed, but when females

of this generation were mated to defective-eyed males

again, we secured defective-eyed young after the manner

of an extracted Mendelian recessive. It is obvious that

in such cases the abnormality could only have been con-

veyed through the germ-cells of the male, and that it is,

therefore, an example of true inheritance. Subsequent

matings have shown that these young transmit the eye-

anomalies as effectively as do individuals of the original

lines. A new strain of defective-eyed young, established

about the time our original paper went to press, is also

flourishing and, as regards transmission of the defect,

seems to differ in no way from the earlier stock.

But now, let us inquire as to where all this leads.

Without entering into a discussion of just what, sero-

logically, is taking place in the body or in the germ of

fetuses borne by the lens-treated mothers, the point I

wish to emphasize is that a certain specific effect has

been produced; and, what is of greater moment, once the

condition is established it may be not merely transmitted,

but inherited. Whether the lens of the uterine young is

first changed and then in turn induces a change in the

lens-producing antecedents in the germ-cells of these

young, or whether the specific antibody simultaneously

affects the eyes and the germ-cells of the young is not

clear. In any event it is evident that there is some con-

stitutional identity between the substance of the mature

organ in question and the material antecedents of such

an organ as it exists in the germ.

Biologically considered, the most significant fact is

that specific antibodies can induce specific modifications

in the germ-cell. Whether these antibodies are trans-

No. 637] IMMUNE SERA 111

mitted from the mother’s blood or engendered in that of

the young would seem to be of secondary importance.

It stands to reason that antibodies originated in an ani-

mal’s own body will modify germinal factors if corre-

sponding antibodies introduced from without can accom-

plish this.

The whole question as to how important such a fact

may be in contributing to an understanding of the causes

of the germinal changes in organisms in general, which

lead to variation and evolution, hinges on the question

of whether changes in an animal’s tissues will induce the

formation of antibodies or kindred active substances in

its own body. We have steadily accumulating evidence

that such reactions do occur.

In our own laboratory, for example, after many at-

tempts we have succeeded in securing a defective-eyed

young rabbit from a mother of normal stock by injecting

her repeatedly with pulped rabbit lens before and during

pregnancy. Since the young rabbit in question has both

eyes badly affected there can be no question that a rabbit

can build antibodies against rabbit-tissue which are as

effective as those engendered in a foreign species such as

the fowl. We have likewise found it relatively easy to

secure spermatoxins by directly injecting rabbits, both

male and female, with rabbit spermatozoa. Moreover,

a given male will develop antibodies against his own

spermatozoa if he is injected intravenously with the

latter.

We are also securing evidence that serologic reactions

induced in the fetus through operations on the mother

are not mere passive transmissions, but may become

actively participated in by the tissues of the fetus. For

example, female rabbits sensitized with typhoid vaccine

followed by living typhoid germs may transmit to their

young and even to their grand descendants the ability to

agglutinate typhoid bacilli in serum diluted from 60 to

160 times. From the standpoint of heredity we have no

reason so far for maintaining that this is anything but

112 THE AMERICAN NATURALIST [Vou. LV

placental transmission, though we are going to practice

immunization generation after generation for a number

of generations to determine if a truly hereditary im-

munity will be established. However, facts have come to

light which show that there is more concerned in the

operation than a mere transfer of antibodies from

mother to fetus. For instance, the blood of young shortly

after birth may show a higher titer than that of the

mother. Again, after two or three months of develop-.

ment the young of certain of the sensitized mothers have

shown a rather sudden rise in titer, much above that of

the mothers. In such cases it would seem that some

mechanism in the young rabbit itself is constructing

antibodies which supplement those passively derived

from the mother. Possibly in the process of develop-

ment some organ important in such reactions just came

into functioning. If this is true further experiments

may throw some light on the perplexing question of

the source or sources of the antibodies in an animal.

After a few weeks, in such cases, the titer drops back

again. In still another set of experiments we found

that young from a sensitized mother, when nursed by

a normal untreated mother, retained a fairly high titer

for several months and even showed the rise of titer

mentioned. On the other hand, young of an untreated

mother when nursed by a sensitized mother acquired

a fairly high titer from the milk of the foster mother

but lost it rapidly after weaning time. Thus there are

evidently constitutional factors operative in the young

which have acquired their immunity through the placenta

which are absent in the young whose antibodies were

conveyed through food.

That changes in the blood serum may be caused by

changed conditions in the tissues is further attested by

many facts. For example, in pregnancy, the newly form-

ing placenta may set free cells or cell-products which,

sometimes at least, cause changes in the blood-serum of

the mother, though the exact nature of these changes is

No. 637] IMMUNE SERA eae

in dispute. Römer, using the complement-fixation tech-

nique, found that the serum of adult human beings may

possess antibodies for their own lens proteins. Bradley

and Sansum, employing anaphylactic reactions, found

that guinea-pigs injected with guinea-pig tissue-proteins

(liver, heart, muscle, testicle, kidney) develop immunity

reactions. Again during the late war, the type of toxic

action to which anaphylactic shock conforms was found

to exist after extensive injury of the soft tissues. It

resulted apparently from the absorption of poisonous

substances of tissue origin into the circulation. In fact,

various cells and tissues when injured liberate such

poisons, and even blood in clotting is known to acquire

a transient toxicity of this type.

With facts such as these before us, is it not a rational

hypothesis to assume that changes in various parts of a

body may on occasion influence the representatives of

such parts in the germ-cells borne by that body? This

appears all the more probable when we recall the facts

learned from the study of precipitins and of anaphylaxis

that each species of animal has a thread of fundamental

similarity underlying the proteins of all its tissues.

There is no reason to suppose that germinal tissue forms

an exception. The further fact that homologous tissues,

though existing in different species of animals, possess

similar chemical characteristics, shows that to get an

effect there need not be absolute identity between the |

protein with which the result is obtained and the original

antigen. Since this is so, in order to have a lens anti-

body affect the germ, there need not be absolute chem-

ical identity between the substance of such a tissue as the

lens and the germinal constituents of which it is the ex-

pression. And if this is true for lens, _ not for other

tissues?

The blood-serum of any organism with blood thus af-

fords a means of conveying the effects of changes in a

parental organ to the germ-cell which contains the ante-

cedent of such an organ. As long as there is little

114 THE AMERICAN NATURALIST [Vou. LV

change in the somatic element its germinal correlative

would presumably remain constant, but any alterations

of the soma which give rise to the formation of anti-

bodies or other active agents, particularly if long con-

tinued, might induce changes in the germ. Such a hy- -

pothesis would seem to be plausible at least in account-

ing for degenerative changes such as the deterioration

of eyes in such forms as the mole, or in fact, in the for-

mation of vestigial organs in general. ;

On the other hand, there is no reason to infer that

changes induced in the blood-serum may not also be in-

strumental in leading to progressive as well as regressive

evolution. If we may have germinally destructive con-

stituents engendered in the blood there is no valid reason

for supposing that we may not also have constructive

ones. When we learn more about what initiates and pro-

motes growth in a part through exercise, or what causes

hypertrophy of an organ, we may likewise find how cor-

responding germinal antecedents of that part may be

‘enhanced. Until such time we shall probably remain in

the dark regarding the mechanism of progressive germ-

inal changes. As already indicated, in the hormones and

chalones we have a wonderful series of secretions nor-

mally circulating in the blood and maintaining general

physiological equilibrium. That reciprocal stimulations

of various organs occur by this means is a well-estab-

lished fact. Hypertrophy or atrophy of an endocrine

gland may produce pronounced effects in the further-

most reaches of the body. Again we may inquire, is it

reasonable to suppose that the germinal tissues will be

inviolate to all this ebb and flow of chemical influence?

Should we not expect specific reactions or selections here

no less surely than in other tissues? Destruction of the

pars buccalis of the hypophysis in the frog-tadpole will

cause profound alteration in other endocrine organs such

as the adrenals and thyroids, will retard the growth rate,

render the entire organism albinous, and produce in the

individual pigment cells a condition of sustained con-

No. 637] IMMUNE SERA 115

traction. Shall we conclude that such a far-reaching in-

fluence as this, particularly in a developing organism, will

pass the germ-cells by unscathed?

Similarly, growth in man is known to be controlled by

a pituitary secretion that is carried by the blood to the

various organs. The normal development of secondary

sexual characters is determined by products from the

testes or ovaries, and the activities of the generative

organs themselves are intimately associated with the

functioning of the adrenal and other glands. The periods

of ovulation are inhibited by secretions from the corpus

luteum; lactation is incited by products of the corpus

luteum, the involuting uterus and the placenta; the car-

bohydrate metabolism in the liver and even in the most

distant muscles is profoundly influenced by substances

formed in the pancreas; the pancreas, liver, and intes-

tinal glands are set to secreting through the stimulus of

a product formed in the duodenal and jejunal mucose.

And still others of such. remarkable interrelations can be

cited.

Truly one may pronounce that social complex of recip-

rocating individuals termed cells which make up an

organism, ‘‘members one of another.” And with all

of these cooperative activities of the various parts of the

body it is inconceivable, to me, at least, that the germ-

cells, bathed in the same fluid, nourished with the same

food, stand wholly apart.

May we not surmise then that as regards inheritance

and evolution, Lamarck was not wholly in error when he

stressed the importance of use and disuse of a part, or

of modifications due to environmental change, in altering

the course of the hereditary stream, particularly if we

conceive of these influences as being prolonged, possibly

over many generations? Have we not in the serological

mechanism of the body of animals an adequate means for

the incitement of the germinal changes which underly

certain aspects of evolution?

DOMINANCE AND THE VIGOR OF FIRST

GENERATION HYBRIDS

G. N. COLLINS

Bureau OF PLANT INDUSTRY, U. S. DEPARTMENT OF AGRICULTURE

A STIMULATION of growth has come to be recognized as

one of the results of hybridization. The phenomenon is

of so much importance, practical as well as theoretical,

that it has been given a special designation, heterosis.

(Shull, 1914.)

New interest has been given to the study of the causes

of this increased vigor by the work of Dr. Donald F.

Jones! (1917 and 1918). Briefly stated the theory ac-

cepted by Jones is that growth is affected by a number

of different characters or factors, the dominant members

of each character pair being favorable and the recessive

unfavorable to growth. Each strain or variety possesses

some dominant and some recessive characters. When

two strains are crossed the first generation. of the hybrid

exhibits the dominant characters of both parents and is

in consequence more vigorous than either parent. In sub-

sequent generations the number of dominant characters

in any individual can not be greater than in the first gen-

eration and in a large majority of instances will be less,

hence the average vigor of the second generation, al-

though still above that of the parents, will be below that

of the first.

The theory is not new, but has not been generally ac-

cepted because of outstanding objections. Dr. Jones has

reviewed the earlier work and has advanced a very in-

genious and entirely novel explanation of the objections.

This explanation will be discussed later.

Bruce (1910) from purely mathematical considera-

1 The theory has been further elucidated in the monograph on ‘‘ Inbreed-

ing and Outbreeding’’ by East and Jones (1920).

No. 637] FIRST GENERATION HYBRIDS 117

tions, showed that if dominance is correlated with vigor, -

crossing would produce ‘‘a mean vigor greater than the

collective mean vigor of the breeds.” Only a few days

later appeared the paper of Keeble and Pellew (1910)

with a concrete illustration and the suggestion that the

‘‘greater height and vigor which F, generations com-

monly exhibit may be due to the meeting in the zygote of

dominant growth factors of more than one allelomorphic

pair.’’ It appears to me unfortunate that in elaborating

this theory, Jones has retained the form of statement

used by Keeble and Pellew, and describes the phenome-

non of heterosis as due to the accumulation of dominant

growth factors instead of placing the emphasis on the

suppression of deleterious recessive characters. It may

seem that the difference is only verbal since a dominant

growth factor presupposes a recessive allelomorph.

There is, however, a difference in the point of view, espe-

cially if the evolutionary significance of the phenomenon

is considered. In speaking of dominant growth factors

we seem to assume as a starting point, strains of low

vigor subsequently improved by the appearance of domi-

nant mutations. It is known that advantageous varia-

tions, whether dominant or recessive, are of extremely

rare occurrence and while evolutionary progress as a

whole must be dependent on such rare progressive changes

‘the effect of these is negligible as a factor explaining

heterosis.

HETEROSIS IN MAIZE

In all varieties of maize there are to be found plants

that are abnormal in some particular and these abnormal

individuals are almost always deficient in vigor and yield.

When varieties are self-pollinated for a series of years at

least a large part of the degeneration that follows is

caused by these abnormalities.

The bearing of abnormalities on heterosis will be more

easily understood if the behavior of two or three ex-

amples is described.

118 THE AMERICAN NATURALIST [Vou. LV

A very common abnormality consists of small yellow

spots thickly distributed on all the leaves which develop

later than the seedling stage. While undoubtedly inter-

fering with the proper functioning of the chlorophyl, the

effect of this abnormality is not serious. Even in breed-

ing experiments seed may be saved from a spotted plant

and in a population containing these partly chlorotic in-

dividuals many of the ovules on the most vigorous plants

will be fertilized by pollen from affected plants. It is

easy to see how characters of this kind persist.

A more serious and less common abnormality is one

that prevents the leaves from unrolling properly, with the

result that the plant is bent and contorted and in extreme

cases never reaches maturity. Seed would seldom be

saved from plants affected with this disorder, but they

frequently produce pollen in normal quantities and the

character in consequence is widespread and difficult to

eliminate.

Albino seedlings may be taken as an example of a still

more serious type of abnormality. In this case all indi-

viduals that show the character die in the seedling stage.

It might appear at first that disorders of this type would

be self-eliminating. The character is recessive, however,

and in many strains there are plants which are hybrid

for the albino character. These hybrid plants show no

trace of the character, yet one half of the pollen grains

and one half of the ovules carry albinism. If either of

these unite with one of their kind an albino plant results,

while if they unite with a normal gamete another hybrid

plant like the parent is produced. Such characters may

be carried along in this manner for any number of gen-

erations in a completely latent form, coming into expres-

sion only when pollen grains and ovules both bearing the

character chance to unite.

Breeding experiments have shown that the more con-

spicuous of these abnormalities are recessive Mendelian

characters which have come into expression through the

chance meeting of male and female gametes both bearing

No. 637] FIRST GENERATION HYBRIDS 119

the character. Only a few of the more obvious of these

abnormalities have been studied, but there is no line of

demarcation between these conspicuous changes and

those that are less evident down to variations that can

not be distinguished visually from environmentally in-

duced fluctuations.

Different varieties possess different assortments of

deleterious characters and in a cross between two unre-

lated strains all of the recessive lethal or semi-lethal

characters? not common to both parents are kept from

expression, since the recessive characters of each parent

are brought into combination with, and suppressed by,

their dominant allelomorphs in the other parent. Free

from the depressing effects of these recessive characters,

the first generation of a hybrid is usually more vigorous

than either parent. In subsequent generations the old

recessive characters again come into expression in some

of the plants, thus reducing the general vigor below that

of the first generation.

If the above explanation of heterosis is to be accepted

it should follow that a majority of the departures from

the normal must be deleterious and recessive while those

which are advantageous are dominant.

The existence, on the other hand, of advantageous re-

cessive, or deleterious dominant, variations would operate

to make F, populations less vigorous than the average

of their parents and conversely inbreeding would tend to

increase vigor.

Variations IN MAIZE CHIEFLY DELETERIOUS AND RECESSIVE

None of the recorded Mendelian variations of maize

is of a nature that would be advantageous to a wild

2 The term character has been used in many places in this paper where

it would have been more in conformity with current usage to employ the

term factor.

Since heterosis results from the combined action of independent units it

might seem proper to call these units factors. Taken individually, hovewer,

each of the units is assumed to produce a tangible effect and is, properly

speaking, a character.

120 THE AMERICAN NATURALIST [Vou. LV

plant and most of them are obviously detrimental. More-

over, if variations occur at random the chances are almost

infinitesimal that any particular variation would consti-

tute a favorable addition to the complex mechanism of a

highly specialized plant or animal. A chance alteration

in the parts of a machine would seldom improve its

efficiency.

Of the recorded heritable variations in maize the de-

parture from the normal condition is recessive in a great

majority of the cases. Aside from a number of wide-

spread characters where neither member of the allelo-

morphic pair may be considered more normal than its

mate, the only dominant variations in maize that come to

mind are pod corn and fasciated or bear’s foot ears.

On the other hand, the recessive variations already

described number more than 20 and it would be safe to

say that hundreds of others are known to maize breeders. .

In a complex organism we may expect that deleterious

variations will occur more frequently than beneficial

variations, but that such a large proportion of the char-

acters should be recessive calls for comment. East and

Jones hazard the suggestion that natural selection has

suppressed the tendency to produce dominant unfavor-

able variations while the tendency to produce unfavor-

able recessive variations has been tolerated.

It should be kept in mind that the observed preponder-

ance of recessive characters does not necessarily imply

that a corresponding preponderance of mutations or

germinal changes are recessive. Dominant disadvan-

tageous variations are eliminated much more promptly

than recessive and the gradual accumulation of recessive

characters soon would place them in the majority in any

cross-bred species, It seems not improbable that the

great preponderance of recessive over dominant char-

acters is a measure of the extent to which dominant char-

acters are eliminated. In a cross-bred form even varia-

tions that result in sterility or death may persist indefi-

nitely if recessive. It may well be that the rate at which

No. 637] FIRST GENERATION HYBRIDS 121

dominant characters appear represents roughly one half

of the germinal changes that are taking place, new reces-

sive characters originating at approximately the same

rate. The preponderance of recessive characters would

then be explained, as the result of their preservation in

a hybrid condition. |

MINOR VARIATIONS Occur WITH GREATER FREQUENCY THAN

AJOR VARIATIONS

This assumption is made necessary by the fact that

the abnormalities which are sufficiently conspicuous to

be identified and isolated will account for a part only of

the reduction of vigor that follows inbreeding. A part

must be due to the combined effect of minor unfavorable

variations, the effect of individual variations being in-

sufficient to produce changes that can be distinguished

from environmental fluctuations.

That minor variations are more numerous than major

is almost self-evident if large and small variations form

a continuous series, as they seem to do, since there is a

limit to the largeness of variations but none to their

smallness. If further proof is needed it follows from the

fact that most major variations can be resolved into less

comprehensive variations and these subsidiary or minor

variations must be more numerous than the major varia-

tions of which they form parts.

As East has noted, our classification of variations into

large and small may have only a remote relation to the

‘importance of characters in the plant’s economy. But

whether judged by the change in appearance or by their

importance to the organism, it is certain that larger or

more fundamental changes must occur less frequently

than smaller or less important variations.

THE Nature oF VARIATIONS IN MAIZE

The appearance of deleterious characters when maize

is inbred and their disappearance when crosses are made,

would follow whether the characters were the result of

recombination or of mutation.

122 THE AMERICAN NATURALIST [Vou. LV

There seems to be no sure way of distinguishing be-

tween the behavior of characters that appear as the re-

sult of recombination and those that result directly from

a germinal change. Changes that occur in homozygous

strains must be mutations. It is, however, theoretically

impossible to be certain that a strain is homozygous re-

gardless of the number of generations that it has been

selfed and, practically, the criterion of homozygosity is

fixed by the accuracy with which comparisons can be

made. With quantitative characters in maize it is diff-

cult to detect with certainty differences of less than 10

per cent., yet sister progenies of strains that have been

selfed for as many as 8 or 9 generations usually show

differences too large to be ascribed to chance. This diff-

culty of obtaining uniformity may be due'to the large

number of factors involved, but also may be due to the

frequency of minor mutations. If a new character ap-

pears in a relatively uniform strain that has been selfed

for a number of generations and the character behaves

as a simple Mendelian unit, it usually is ascribed to a

mutation. Even in such cases, however, the character

may be due to recombination. If the two factors of a

dihybrid recessive character arose independently in

nearly the same position on homologous chromosomes,

the close linkage of the dominant allelomorph of one fac-

tor with the recessive allelomorph of the other would

long postpone the appearance of an individual with both

recessive factors and. when it did appear the departure

of its behavior from that of a simple character would be

difficult to detect.

Although the nature of variations does not affect the

bearing which the preponderance of recessive characters

has on the explanation of heterosis, there are practical

as well as theoretical reasons for wishing to know

whether new characters that appear in breeding stocks

are mutations or result from the combination of factors

already present in the germ plasm. If the undesirable

characters that appear from time to time, even in well-

No. 637] FIRST GENERATION HYBRIDS 123

bred’ varieties, are the result of recombination, the

breeder will be encouraged to expend the time and labor

necessary to eliminate them. If, on the other hand, these

new characters are the result of an unstable germ plasm

other means must be sought.

Already the importance of deleterious recessive varia-

tions has found application in the breeding of maize. It

soon was realized that to successfully eliminate recessive

characters it is necessary to bring the characters into

expression by inbreeding. Once a strain has been freed

of undesirable characters, vigor may be restored by com-

bining the inbred lines or the full advantage of domi-

nance may be realized by growing first-generation hy-

brids of the better strains.

This method of breeding will be relatively unsuccess-

ful if unfavorable mutations are of frequent occurrence.

It is perhaps too early to be sure that this is not the case,

but it is encouraging that in strains self-pollinated for

13 generations Jones finds no conspicuous variations

after about the 7th generation. The next step is to

demonstrate that no unfavorable variations appear when

the selfed lines are crossed. ‘This has been shown to be

the case in the first-generation, but a certain percentage

of multiple factor recessives are to be expected in sub-

sequent generations.

If linkage is operative these recessive characters would

come to light slowly, much as they appear in successive

generations of an open-bred variety. As already pointed.

out, there is as yet no way of distinguishing between

mutations and multiple factor characters, when the fac-

tors are linked.

NATURE or DEGENERATION Tmar FoLLOWS INBREEDING.

In discussing the nature and causes of the reduction

of vigor that follows inbreeding, it is necessary to choose

words with great care. To state the questions at issue

in such a way as to distinguish between differences of

fact and differences in the use and meaning of words

will tax the possibilities of the language.

124 THE AMERICAN NATURALIST [Vou. LV

Many of the older writers on heredity have held that

inbreeding is a cause of degeneration. In avoiding am-

biguous words ‘‘cause’’isone of the first that must go. If

forced to define their position this school would probably

be content with the statement that degeneration is a nec-

essary consequence of inbreeding, the intermediate steps

or nature of the process being unknown. Is this concep-

tion really at variance with the idea that degeneration

results from the increased number of unfavorable re-

cessive characters brought into expression by increasing

homozygosity? Does not this conception rather amplify

the older, general and indefinite position by explaining

how the degeneration may be brought about?

It excites unnecessary opposition, and is not entirely

fair, to read into the early writings the idea that inbreed-

ing was held to be the immediate and direct cause of the

subsequent degeneration. Such words as ‘‘ecause’’ and

‘‘per se” have perhaps been used, but is there not suff-

cient latitude in their meaning to allow the later dis-

coveries to be looked upon as explaining rather than re-

futing the old doctrine?

In the attempt to bring the two views into sharp con-

trast the newer explanation is sometimes stated in terms

which likewise must be interpreted with latitude if the

explanation is to be accepted. Thus Hast and Jones (p.

123) state one of the results of inbreeding maize as fol-

lows: ‘‘There is a reduction in size of plant and produc-

tiveness which continues only to a certain point and is in

no sense an actual degeneration.” It is difficult to

imagine a degeneration more ‘‘actual’’ than that usually

following the inbreeding of maize.

In another place (p. 139) the same authors say: ‘‘The

only injury proceeding from inbreeding comes from the

inheritance received.’’ Such statements have an unfor-

tunate air of finality that probably was not intended.

The relation between inbreeding and degeneration has

been greatly clarified by the work of these authors, but

the above statement taken literally places them in a posi-

No. 637] FIRST GENERATION HYBRIDS 125

tion similar to that of the older writers who stated that

inbreeding was the ‘‘cause’’ of degeneration. -There

may well be other and important ways in which inbreed-

ing is associated with degeneration.

There is for example definite evidence that vigor is

reduced by continued asexual reproduction (Shull, 1912),

and although it may be urged that this result is asso-

ciated with the phenomenon of senescence, so may be the

decline of vigor that follows inbreeding.

Furthermore, it has been shown by Calkins (1919)

that in the ciliate, Ursoleptus, conjugation between sis-

ter cells of an asexually propagated line increases vigor.

OBJECTIONS TO THE EXPLANATION

Two objections have stood in the way of accepting

dominance as an explanation of heterosis. The first of

these is that if this explanation is the correct one it

should be possible to obtain an occasional F, individual,

homozygous for all dominant allelomorphs, the progeny

of which should be uniformly as vigorous as the F,. It

is held that no such F, individual has been found. The

second objection is that the distribution in F, should be

skew with the mode above the mean while in fact F,

populations show a symmetrical distribution.

Jones has proposed a novel and ingenious explanation

of both objections. He has pointed out that it is only

necessary to assume that the phenomenon of linkage,

which plays such an important rôle in the inheritance of

Drosophila, is operative also in maize.

If groups of characters are inherited as units with

little or no crossing over, both dominant and recessive

characters being represented in any particular unit, the

first generation would still exhibit all the dominant char-

acters of both parents, but when segregated into pure

lines each pure line would again exhibit recessive char-

acters, with a consequent decline in vigor.

This assumption of group inheritance or linkage would

also meet the objection. that F, populations exhibit a

normal and not a skew distribution.

126 THE AMERICAN NATURALIST [ Vou. LV

There ‘are a number of instances of coherence or link- —

age known in maize, but if the characters studied to date

are a fair sample the rôle of linkage must be of minor

importance. Linkage relations have not been studied in

sufficient detail even to state with assurance that the

characters are arranged in linear series corresponding

to the number of chromosomes, although this conclusion

is indicated. The linkages of most of the Mendelian

characters are very loose and it would seem necessary to

conclude that, if the characters of maize are arranged in

a linear series,-the chromosomes must be either very

long or very flexible.

While admitting that linkage would meet the objec-

tions urged against the simple hypothesis that the sup-

pression of recessive characters explains heterosis, it

may be well first to make sure that any such assumption

is necessary. An examination of the maize literature

indicates that the difficulty of securing uniform strains

with the vigor of the first generation has been assumed

rather than demonstrated. No case was found where

selection following hybridization had been continued long

enough to approximate homozygosity. There are also

very few cases where the more vigorous F, individuals

have been chosen as parents of the F,. The most exten-

sive series of experiments are those of Emerson and

East (1913). 7

Height is probably the most satisfactory character to

use as a measure of heterosis. ‘There are 23 compari-

sons of F, and F, populations in the work of Emerson

and East. To these six can be added from our own ex-

periments.

In these 29 cases the mean of the F, was below that

of the F, in every instance but in ten of the 29 cases the

largest of the F, plants equalled or exceeded the largest

of the F, individuals and in every case where a progeny

was grown from a plant near the upper limit of the

range of the F, its mean exceeded that of the F,.

Other characters reported by Emerson and East for

No. 637] FIRST GENERATION HYBRIDS 127

which the F, was measurably larger than the mid-

parental value are length and diameter of ear and length

of internode. With respect to length of ear (Tables XIII-

XV), there are 13 F, progenies that may be compared

with the F,. In 10 the mean was higher than the mean of

_ F,. Four F, progenies were grown from F, individuals

above the mean of F; and in 3 of these the mean exceeded

the mean of F}. .

With respect to diameter of ear (Tables XVIII and

XIX), 8 F, progenies may be compared with’the F,. In

2 of the 8 instances it would appear that the mean was

above that of F,. Seven of the 18 F, progenies were

grown from F, individuals above the mean of F, and in

every case the F, mean exceeded the mean of F..

Length of internode is the character showing the most

decided increase in F, over the mid-parental value. In

the two crosses reported (Tables XXXIII and XXXIV)

this increase was 33 and 27 per cent. None of the F,

progenies grown the same season as the F, equalled the

F,. The mean of the F,, however, was exceeded by the

mean of 5 of the F, progenies grown the following season,

although the parents were not selected for internode

length. The results with these characters give little or

no evidence of non-heritable vigor in the F,, neither is

there any proof that it is difficult to select progenies

with the vigor of F..

EXPECTATION oF OBTAINING F, PRoGENIES WITH THE

IGOR OF F,

In the absence of experimental data, to secure which

will require very extensive experiments extending over

many years, it may be instructive to consider what re-

sults may be expected if suppression of semi-lethal char-

acters is the true explanation of heterosis.

The diffculty of obtaining individuals homozygous for

all or even a limited number of characters has been fre-

quently pointed out, but the bearing of this on the diffi-

culty of retaining the vigor of the first generation seems

not to have been appreciated.

128 THE AMERICAN NATURALIST [Vou. LV

With a sufficiently large number of characters influen-

cing vigor it would be impossible in practice, even without

the assumption of linkage, to obtain homozygous indi-

viduals having the vigor of the first generation.

Thus with 10 pairs of characters over 700,000 indi-

viduals would have to be grown before there would be an

even chance of obtaining an individual homozygous for

all of them.

More than ten separately inherited Mendelian char-

acter differences affecting growth have been identified

and there is no reason for believing that more than a

small proportion have been isolated or that more than a

small proportion produce conspicuous morphological

changes that would be readily detected.

A near approach to the vigor of the F, might be ex-

pected, of course, without complete homozygosity. :

Some idea of the chances of isolating strains that are

practically homozygous may be obtained by calculating

the size of the populations that must be grown to insure

a reasonable chance of finding an individual homozygous

for say 70 per cent. of the characters.

Table I indicates the size of the populations necessary

to fulfil these conditions with the number of character

pairs ranging from 10 to 30.

TABLE I

Number of F? Individuals Necessary to Provide an Even Chance of

Number of Pairs Obtaining at Least One Individual Homozygous for 70-90 Per Cent. of

of Characters the Character Pairs

Involved

70 Per Cent. 80 Per Cent. 90 Per Cent.

WO ia Hi ee 199 1,760 23,400

Bhs Sey Gaels ae 6,000 56,000 16,150,000

BUG US eevee 23,600 1,800,000 429,000,000

yds ER OS APE TE E a, 457,000 39,100,000 28,200,000,000

Se a E | 2,470,000 1,710,000,000 7,070,000,000,000

It will be seen that to have a reasonable chance of ob-

taining an individual homozygous for even 70 per cent.

of the character pairs it is necessary to limit the char-

_ acter pairs to 15 or less. Another way of gaining a

quantitative idea of the degree of homozygosity that

FIRST GENERATION HYBRIDS

‘SOINDY JUVOYIUSIS sory} juo 0} ported əre SUOTyRUIMAJOp OY, t

GODKUONEL fo ee ee ee eee ee re ere ee oe =

coposror = | Goop t ae o Toe E ee o a e a a a +1

0000¢2 OOQUESS< (0000090 o T o I a a ei er

00099 000STZ OMNDOTI osovo = T a ee a e a a e kn =

0009 OOFLT 00929: {O00GTE | oodo] To o To a a a, It

028 0£0z 06'S OOFST 00798 oooh: o et a l eer ee Or

COL Gee 002 OLLI 00¢¢ oorgee J oogst io Teo o ee ee as 6

IF 89 EZI OF P89 0921 ozo. Ioe BE cg T o o g oaeo 8

£I 61 6% 6h ¥6 661 91g OT ON a cd Pee L

g L 6 eT oz Se OL Sol (91 | OF8z ae fie aes 9

€ € p g 9 6 v1 9 «+9 WP ete Pe aes g

3 4 Z Z g £ ? 9 OL | Bt | SY. | Bet ee F

T T Z Z z Z Z Z g F Z i yn a eoor <

T T: T T I I T I I z z e G yp eertain Z

ST ial £r gI TI or 6 8 4 9 g A € g | Sopuwyo snog

; -AZOWOH JUVU

PƏAJOAUI SI9JOVIVYO JO Jaquinn

uoa Jo Jaquinn

TANO NWATOQ NI ŒALVOIANI SAALOVAVHO LNVNIWOQ AO UAaWOAN AHL

Uod ‘SnODAEOROH ‘IVOGIAIGNT ONO Lsvay Ly DNINIVLAQ dO DONVHO NAW NV AdIAOUT OL AYVSSAOUN S'TVOGIAIGNT * JO WadWoAN

Il WTA vb

130 THE AMERICAN NATURALIST [Vorn. LV

may be expected is presented in Table II, which gives the

number of individuals necessary to provide an even

chance of obtaining at least one individual homozygous

for different numbers of characters in crosses when from

two to 15 character pairs are involved. It should be

kept in mind that even though one should obtain an indi-

vidual homozygous for a sufficiently large percentage of-

the characters involved to approximate closely the F, in

vigor, it would be necessary to grow a progeny from this

individual before its inherent vigor could be demon-

strated. The numbers given in the tables might thus be

taken to represent F, progeny rows instead of F, in-

dividuals.

The conclusion is that perjugate progenies equalling

or even closely approximating F, in vigor are hardly to

be expected in breeding experiments and consequently

no assumptions are necessary to account for their non-

appearance.

Skew Distripution Dur to DOMINANCE

The second objection, that of the failure of F, pro-

genies to show a skew distribution, may now be consid-

ered. There can be no question that a series of inde-

pendent, dominant characters influencing size would

bring about a skew distribution. Assuming the charac-

ters to have equal effect, two characters would give a dis-

tribution of 1, 6, 9, three, a distribution of 1, 9, 27, 27,

and with four characters the distribution would be 1, 12,

54, 108, 81. It is apparent that with an increase in the

number of characters the skewness becomes less pro-

nounced. It may be of interest to determine whether,

with a reasonably large number of characters, the skew-

ness would be detected in populations of the size usually

grown in experiments.

With 20 pairs of characters giving 21 classes 1,099,-

514,627,776 individuals would have to be grown to obtain

a representative population. Of this population 99.91

per cent. would fall in the 12 classes with the largest

No. 637] FIRST GENERATION HYBRIDS 131

number of dominant characters. That is, populations of .

over 700 would have to be grown before there would be

an even chance of getting any individuals smaller than

those represented in these 12 classes. With ordinary

sized populations then the distribution would be fairly

represented by the distribution of the 12 largest classes.

The distribution among these 12 classes would be as

follows:

No. Dominant Allelomorphs Proportion of Individuals Expected

Disk he S E E T E see kee 3

Penre as a oan a ee es 1.0

i D Bhar ae Fatal aie RR eK Ray A AE Pa SN rose 2.7

A io E Epa ed A gg Rg EE ome GAL Sera 6.1

DDoS hE) DEOR ated Ghee eS FS ee eee ee wie 11.2

Ae E E E A E E 16.9

e E T A E E A a ONE 20.2

tO oeira a LA EA 19.0

A a E E E E E ENA E AE E 13.4

s E A See A a A a 6.7

Le E E E E CT ETA TA ET 2.1

UF ee Na ae Sad es eel gta atte are 3

A distribution of this nature, with populations of ap-

proximately 500 individuals, conforms to the normal fre-

quency curve as closely as would be expected. The mode

departing from the mean by only 3/100 of a class.

The theoretical distribution of an F, population in-

volving 20 pairs of characters of equal weight with com-

plete dominance is shown in the accompanying diagram.

It will be noted that although the curve as a whole is

skew, the portion to the right of the class with 9 domi-

nant characters, which comprises 99.91 per cent. of the

area, is practically symmetrical.

With 10 character pairs there would be 11 classes and

99.65 per cent. would fall in the 7 largest classes and in

this portion of the theoretical population the mode would

be separated from the mean by only 3/10 of a grade.

Even should it be possible to grow F, populations suffi-

ciently large to detect departures from a normal fre-

quency distribution there is yet another reason for ques-

tioning that a skew distribution should be expected, when

THE AMERICAN NATURALIST [ Vou. LV

ow)

SFSR

;

ee...

= ase

“J

NUMBER OF INDIVIDUALS 9 CIPHERS OMITTED

Va \

423A 5678 Mt 1213 1P15 16 1718 1920

NUMBER OF DOMINANT CHARACTERS

Distribution of individuals in an Fe population involving 20 pairs of characters

of equal weight and showing complete dominance.

SOSESSLIIVBss

plotted in the customary way. It has come to be accepted

that the effect produced by a given growth factor is de-

pendent on the size of the organism. For example, if a

growth factor increases the length of the internode by a

given amount, it is clear that the height of a plant with

30. internodes will be increased more than that of a plant

with only 15 internodes. In other words, the effects are

factorial instead of additive. A convenient method of

classifying a population on a factorial basis has been

proposed by Zeleny (1920), who takes the range of each

class as a constant percentage of the value of the mid

No. 637] FIRST GENERATION HYRRIDS 183

point of the class. The result of this change in plotting

is to increase the range of sizes included in the higher

classes and consequently to raise the mode.

In conclusion it would seem, therefore, that the as-_

sumption of linkage, while perhaps not improbable, is

superfluous so far as the explanation of heterosis is con-

cerned, since neither of the objections which it was

framed to meet have foundation in fact.

It is, perhaps, too much to assert that the suppression

of deleterious recessive characters completely explains

heterosis or that the reappearance of these characters is

the only factor in the decline in vigor that follows in-

breeding, but the behavior of maize is in full accord with

this explanation.

LITERATURE CITED

Bruce, A. B.

1910. The Mendelian Theory of Heredity and the Augmentation of

Vigor. Science, N. S., 32: 627-628, Nov. 4, 1910

Calkins, G. N.

1919. Dircleptas mobilis Eugelm. IL Renewal of Vitality through

onjugation. Jour. Experimental Zoology, 29: 121-156, Oct.

9. -

4 `

East, E. M., and Jones, $

1920. Inbreeding ink Outbreeding.

Emerson, R, i d East,

1913. i ES Characters in Maize. Nebr. Exp. Sta. Re-

aoe Bull. No. 2

Jones, D. F.

1917. Dominance of Linked Factors as a Means of Accounting for

Heterosis. Genetics, 2: 466-479, Sept., 1917.

1918. The Effects of Inbreeding and Crossbreeding upon Development.

Conn. Agric. Exp. Sta. Bull. 207, Sept., 1918.

Keeble, F., and Pellew, C.

1910. The Mode of Inheritance of Stature and the Time of Flowering

in Peas. Jour. of Genetics, 1: 47-56, Nov. 18, 1910.

Shull, A. F.

1912. The Influence of Inbreeding on Vigor in Hydatina senta. Bio-

logical Bulletin, 24: 1-13, Dee., 1912.

Shull, G. H.

1914. Duplicate Genes for Capsule Form in Bursa bursa-pastoris

Zeitschrift fiir induktive Abstammungs- und Vicibungddiie,

12: 97-149.

Zelney, C.

1920. The Tabulation of Factorial Values. American Naturalist, 54:

358-362, July-August, 1920

CORRELATION OF TAXONOMIC AFFINITIES

WITH FOOD HABITS IN HYMENOPTERA,

WITH SPECIAL REFERENCE TO.

PARASITISM?

PROFESSOR CHARLES T. BRUES

Bussey Institution, HARVARD UNIVERSITY

Ewtomovoeists can all agree that the attachment of

most phytophagous species belonging to the more highly

specialized orders of insects is very firmly fixed, and the

Hymenoptera form no exception. We can also agree,

although in less definite terms, that many parallels exist

among plant-eating Hymenoptera between taxonomy and

food habits. I do not propose to treat of this series,

-however, partly because I am not sufficiently familiar

with them, but also on account of the great interest which

attaches to the parasitic groups of Hymenoptera.

During the last decade our conception of the process

of nutrition in insects has undergone considerable change,

due to the discovery that various microorganisms form

an important part of the food supply of many forms.

It is quite certain that certain sapropl , sarcoph-

agous and coprophagous ones probably feed directly, not

at all upon decaying and fermenting plant materials,

carrion or excrement, but upon the bacteria, yeasts, ete.,

always abundant in organic material undergoing decom-

position. We must judge of the protein requirements of

such insects not by the gross substances or substratum,

but on the basis of the microorganisms present (Baum-

berger, 719).

This aspect does not appear to enter into the economy

of the Hymenoptera, although there may be'a relation

between fungi and nutrition in some Cynipide, as the

1 Contribution from the Entomological Laboratory of the Bussey Institu-

tion, Harvard University, No. 181.

134

No. 637] PARASITISM 135

galls of these insects are commonly invaded by fungi and

yeast-like organisms. On the other hand, these same

gall-wasps exhibit remarkable correlations between struc-

ture and habits as shown by many workers, most elabo-

rately and clearly by the recent studies of Kinsey (’20).

The Cynipide are at present restricted to a very limited

series of plants on which they induce the formation of

galls. With the exception of some undoubtedly primi-

tive forms, they occur almost exclusively on Rosacee of

the genera Rosa and Rubus, and on the unrelated genus

Quercus, the latter harboring a very large number and

a far greater variety of forms. In the gall wasps, we

see, therefore, a nearly exclusive association with a very

few genera of plants and what is still more striking is

the fact that the more primitive ones, although few in

number, exhibit a wide range of food-plants. This leads

to the inevitable conclusion that we can trace the evolu-

tion of host relations in this group as now living, from

a very generalized condition to a highly specialized one.

Many other insects, particularly Homoptera, harbor

certain probably symbiotic organisms, and recent studies

(Brues and Glaser, ’21) tend to show that these may be

very important factors in the nutrition of the insects.

Almost nothing is known concerning such organisms in

Hymenoptera, but if they are present in some cases, as

seems probable, they must be reckoned with in any com-

plete studies of food-habits.

The relation of food-habits to taxonomy in the Hymen-

optera becomes particularly interesting in connection

with the appearance of parasitism in several forms,

which is of widespread occurrence in the order. In fact,

a large proportion of the Hymenoptera are parasitic,

and with the development of this mode of existence have

come such elaborate structural modifications, and spe-

cialization in behavior, that species have multiplied at

a very rapid rate. From this apparent chaos, even tax-

onomists have as yet brought only very partial order,

and any discussion of parasitism in the Hymenoptera is

136 THE AMERICAN NATURALIST [Vou. LV

necessarily very incomplete and may be misleading in

some details.

In the Hymenoptera the designation ‘‘parasitic’’ has

been applied to habits of extremely diverse nature, and

this very loose use of the term may easily lead to serious

confusion unless we consider the matter carefully before

proceeding further. It is most commonly applied to

several large and abundant groups whose members live

in the larval condition in other insects which they almost

invariably destroy after attaining full growth. Such

habits are quite similar to those encountered in other

orders of insects, although in no other do they attain

such a high degree of specialization; nor do they involve

such a series of unrelated smaller groups, with the pos-

sible exception of the Diptera. While this is the most

abundant and widespread type of parasitism among in-

sects, we must not lose sight of the fact that it is a rather

unusual condition so far as animals and plants in general

are concerned, in that the parasite and host belong to the

same class and are thus closely related. A similar rela-

tion exists among the Crustacea where certain decapods

are parasitized by other members of the same class, and

even among Protozoa there are ciliates parasitic in the

bodies of other infusorians. Such associations are very

rare, however, and the vast majority of parasitic Crus-

tacea and Protozoa, as well as other parasitic animals,

depend upon animals far removed from themselves for

hosts, although there is very generally a close correla-

tion between the host and parasite in that related para-

sites depend upon related hosts. Another peculiarity of

this type of insect parasitism lies in the prompt death of

the host, which does not usually ensue as a result of other

animal parasites, although it is a common result of pres-

ence of some protozoan parasites of the higher animals.

In others, again, like the well-known nematode, Trichina,

the ultimate death of the host is necessary for the con-

tinued propagation of the parasite, but actual death nor-

mally results from other causes. Still another charac-

No. 637] PARASITISM 137

teristic of this type of insect parasitism is its restriction

to the larval stages, although extending over the entire

growth period. This is by no means unique among ani-

mals, but it is one of the distinguishing characteristics

between this type of parasitism and the one next to be

considered. In its perfected state this relation between

host and parasite is a marvelously balanced association

and one which we might expect to furnish valuable data

on the correlation between taxonomy and habits.

A second type of parasitism encountered in the Hymen-

optera is that exemplified by most of the parasitic bees

and wasps. This has recently been discussed by Wheeler

(719). Here the parasitic larva is really at first preda-

tory so far as food-habits are concerned; devouring the

host larva shortly after hatching. The sequence of events

is initiated by the preparation of the larval food-supply

of the host by the mother bee or wasp. Most solitary

wasps store up, in a nest which they have prepared, one

or more insects which they have paralyzed by means of

the sting, and attach one of their eggs to the body of the

prey. Under normal conditions the larva hatching from

this egg consumes the prey, attaining full growth, and

later completing its metamorphosis. Bees behave in

much the same manner, but the store of food in the nest

consists of honey and pollen. When parasitism inter-

venes, the egg of the parasite is also placed upon the food

supply, and on hatching, the larva of the host and para-

site find themselves in proximity, each ready to appro-

priate the contents of the nest. In numerous cases that

have been studied (Graenicher, 05), the larva of the

parasitic form has more powerful jaws than its rival, and

encounters little difficulty in destroying it. It now pro-

ceeds to consume the food-supply exactly as the host

larva would have done, casting off its enlarged mandibles

at the first molt. Thus the actions of the larva savor

not at all of parasitism, but it is in the fixed habits and

instincts of the adult, which require the nests of particu-

lar wasps and bees, that the parasitic relation holds.

138 THE AMERICAN NATURALIST [Vou. LV

Correlated with such habits, structural modifications of

the body appear, such as the loss of the pollen-collecting

apparatus in parasitic bees.

In certain ants there occurs a third type, social para-

sitism (Wheeler, ’04) whereby the young females of

some ants that do not establish their own colonies insin-

uate themselves into the nests of other species of ants,

do away with the queen, and take on themselves the func-

tion of egg-laying. As the larve from these eggs are

raised to maturity, they produce worker individuals of

the parasitic species which gradually supplant the orig-

inal population. Finally, the colony becomes pure and

maintains itself through its own efforts, giving no evi-

dence of the temporary social parasitism by which it

has originated. In a very few cases social parasitism

may become permanent with the complete elimination of

the worker caste.

The term entomophagous parasite may be applied with

some appropriateness to all of the three types described,

but is most suitable for the first one, since there the par-

asite not only consumes its host, but feeds upon nothing

else during its developmental stages. By far the largest

number of species in the order exhibit this type and it is

the only one which I shall consider in any detail.

There are several ways in which such parasitism may

have originated, but the question of origin is best de-

ferred until its several phases have been discussed at

greater length.

Defining parasitism in its several forms as enumerated

on a previous page, we find that there are parasitic

genera included in nearly allof the larger groups of Hy-

menoptera. Thus, the Ichneumonoide, Serphoidea and

Chalcidoidea, each represented by a number of families,

are almost exclusively entomophagous parasites, while .

in the Aculeata, numerous parasitic genera appear, scat-

tered through a series of families with generally non-

parasitic habits. In addition to these there is the prim-

itive family Orysside, now known definitely (Rohwer,

No. 637] ; PARASITISM 189

17) to be parasitic, and a few other families nearly all of

somewhat doubtful affinities. -Thus of the nearly one

hundred families included in the order, between forty

and fifty are composed either entirely or almost exclu- ;

sively of genera. with parasitic habits, the remainder

being phytophagous or predatory with isolated cases of

parasitism, among both the predatory series, and one of

the phytophagous ones.

Considering these larger groups, the suborders and

superfamilies, more in detail we find that the most prim-

itive of all known Hymenoptera, the suborder Chalasto-

gastra, are phytophagous. Of these, about a dozen

families, comprising the sawflies or superfamily Tenthre- -

dinoidea, are almost exclusively defoliating forms, feed-

ing in their larval stages on the leaves of various flower-

ing plants. Another family, the Siricidæ, feeds inter-

nally on the tissues of woody plants, and, at least so far

as food-habits are concerned, there are two other fami-

lies which form a transition between the sawflies and

wood-wasps. It is in the groups above these that the

parasitic habit appears, and with the possible exception

of one family, the Orysside, to be mentioned later, all

these groups are usually associated as a second suborder,

Clistogastra, contrasted to the more primitive Chalasto-

gastra. Among them several groups of families, con-

veniently classed as superfamilies, are three extensive

parasitic ones: first, the Ichneumonoidea, comparatively

large species comprising about half a dozen families;

second, the Chalcidoidea, represented by small or minute

species comprising fully a dozen families; third, the

Serphoidea another half dozen, mainly very small species.

Together with a part of the Cynipoidea, these form the

enormous complex commonly known as the Hymenoptera

Parasitica. All are quite closely related, but rather easily

grouped and distinguished, in spite of certain annectant

and aberrant families.

The habits of the several series are also very uniform.

140 THE AMERICAN NATURALIST [Vou. LV

The egg is nearly always laid upon the body of the host

or thrust into it, usually the latter, to which purpose the

extrusible, stiletto-like ovipositor of the female is adapted

with great nicety. Oviposition may take place either in

the egg of the host, in the larva, or even in a later stage,

and the parasite may complete its development either in

the stage of the host in which it is laid, or development

may be delayed and not completed till the host has pro-

ceeded to a further stage in its ontogeny. Under such

conditions the larva is to a great extent passive, although

in its earlier minute stages it frequently exhibits (e.g.,

in certain Serphoidea) great modifications in body form,

and develops monstrously specialized jaws or other or-

gans to aid in attacking the massive tissues or yolk-

masses of its host.

When such modifications of the young larva are tran-

sitory and disappear almost completely after one or two

ecdyses, they form a transition to several very clearly

defined cases of hypermetamorphosis which have been

noticed in certain Chalcidoidea by several observers

(Wheeler, 707; Smith, 712, and Brues, 719). In members

of two families, the Eucharide and Perilampide, they

have found an active, free-living, first stage larva known

as a planidium which is quite similar to the triungulin

of the Meloid beetles and the Strepsiptera. Like them,

the planidium becomes helpless once it has become para-

sitic. Great interest attaches to the planidium, but until

its distribution is much better known it can not be con-

sidered of taxonomic value, especially as quite similar

larve are known in several other orders of insects.

Another series of Hymenoptera, certain parasitic bees,

are known through the researches of Graenicher (’05)

and others to possess much larger jaws in the first larval

stage. As we have mentioned previously, the type of

parasitism in this case is very different, for the parasite

simply eats the host larva that it may appropriate its

food-supply, and we have a parallelism in structure, of

No. 637] PARASITISM 141

independent origin, and hence of no classificatory im-

portance.

Comparative anatomy and post-embryonic develop-

ment show very clearly that, with the exception of some

secondarily phytophagous forms, only the primitive

Hymenoptera are phytophagous. As one can not seri-

ously question the monophyletic origin of the order, the

varied food-habits now represented must have been de-

rived from some form of vegetarianism.

In all of the higher Hymenoptera or Clistogastra,

active and aggressive characteristics are very prominent

in the behavior of the adult females, whatever may be

the food-habits of the larve. Thus in the wasps, the

parent captures as prey suitable insects with which to

feed her larve, or to provision her nest, if her young are

to receive no post-natal care. In all cases she prepares

some sort of a cell or nest for her brood, and frequently

this requires marvelous skill in the selection of particular

materials and the collection of specific insects for food.

Where nests are provisioned in advance, the prey is

stung and paralyzed after a manner that requires very

complex instinctive behavior. If, on the other hand, we

look at the activities of the larva of one of the wasps

that stores a paralyzed insect away and places her egg

upon it, we see the larva consuming its food supply much

after the fashion of an externally feeding entomophagous

parasite. In fact, it is difficult to distinguish any really

fundamental differences. In each case the host is stung

and the egg attached to it, always externally by the wasp,

but sometimes externally also by the parasite. The wasp

paralyzes her prey, which the parasite does not do, as

her sting is not so severe, and she does not further bother

with the host insect. The egg of the parasite is deposited

at the time of stinging, and that of the wasp by a later

operation of the same organ, the ovipositor with which

she has previously paralyzed, but not killed, her prey.

Thus, aside from the maternal instincts, the entomophil-

ous wasp is scarcely more different from the ichneumon-

142 THE AMERICAN NATURALIST [ Vou. LV

fly, than some ichneumons from others. Equally varied

habits exist in at least a few cases even in a single species

of ichneumon, for certain Itoplectis may be either para-

sites of caterpillars, hyperparasites, or inhabitants of the

ege-cocoons of spiders where they devour the contained

eggs. From the entomophilous wasp has been developed

the parasitic one and we have alluded to its origin as

traced by Wheeler (719).

From the foregoing, it is seen that we might derive the

habits of the wasp from those of the parasite, or vice

versa, with but little diffieulty, although the more elabo-

rate instincts of the wasp appear more naturally as the

latter development.

If now we return to the free-living phytophagous

Chalastogastra, it appears for morphological reasons

especially that the entomophagous ichneumon flies have

been derived directly from them and I think that the

transition from phytophagy to parasitism is quite clear.

Whether it involves the interpolation of predatism or

sarcophagy is perhaps more a matter of conjecture.

The Siricoidea of the Chalastogastra, on account of

their legless, eruciform, lignivorous larve and reduced

wing venation appear to have been derived from some

sort of ancestor with a caterpillar-like larva having the

more complex wing-venation seen in the saw-flies or Ten-

thredinoidea. So far as is known, no member of either

group is parasitic. Until recently the family Orysside

has been regarded as a degenerate group quite closely

allied to the Siricide. Rohwer has, however, shown that

they are really very different and finally (Rohwer, 717)

regarded them as a distinct suborder of Hymenoptera.

It seems reasonable to suppose that they have Siricid-

like ancestors, and as they are now known definitely to be

parasitic on the larve of wood-boring Coleoptera, it ap-

2 This matter, as well as several others discussed in the present paper,

have been recently dealt with by Picard in a publication (La faune ento-

mologique du figuier, 719) which I unfortunately did not see until too late

to refer to it in the text of this article. Picard’s ‘‘Considérations sur les

parasites,’’ pp. 166-172, are of extreme interest.

No. 637] PARASITISM 143

pears that we have in them the most primitive parasites

in the order Hymenoptera. The hosts of the Orysside

consist partly, although probably not entirely, of Bu-

prestide, which paleontology shows to be an ancient

family. Handlirsch (’08) has even gone so far as to sug-

gest that the parasitic Hymenoptera may have been de-

rived from the Jurassic Pseudosiricide which no longer

laid their eggs in wood, but in the eggs of beetles occur-

ring in the wood. This is entirely speculative and so I

think must be at the present time any suggestions as to

how the Orysside, or the Ichneumonide, which Hand-

lirsch had in mind, became parasitic. That their larve

first found and fed upon their hosts after hatching seems

much more probable. It must be said, however, that

predatory or carnivorous Chalastogastra are not known

among living forms, except certain adult sawflies which.

fed in this way (ef. Mrázek, 09). From this point on-

ward we have little trouble in tracing the probable origin

and relationships of the Ichneumonoid families as I have

attempted to do in a previous paper (Brues, 710). Thus

the Stephanide are structurally primitive and strikingly

like the Orysside in the peculiarly horned head which

had been remarked on before the habits of the Oryssids

were known. On account of the presence of a costal cell

in the wing, the polymorphic family Evaniide is neces-

sarily also more primitive than the Ichneumonide or

Braconide, and, through one subfamily, the Fæninæ, re-

semble the Stephanide as has been already noted by

Bradley (’08). Some Braconide, the Stephaniscine,

Spathiine and Hormiine are much like Stephanids, se

much so that it is difficult to believe that they are

directly derived from them. One other family, the

toniide, recently segregated from the Bracon

pears to be very definitely related to the

ized Evaniide (Aulacine). Omitting in this

sideration several less pertinent families, an ring

other recently segregated ones, we have left na the

Ichneumonide, related possibly through the Alysiide to

144 THE AMERICAN NATURALIST [Vou. LV

the Braconide. Structurally this relation seems plaus-

ible, but as the Alysiids attack almost exclusively the

highly specialized Diptera it is very difficult to regard

them as closely related to the ancestors of the Ichneu-

monide, so highly diversified in habits and structure.

The latter then are not so easily derived and may go

back to Evaniid-like forms.

One extremely interesting fact in connection with the

primitive families of parasitic Hymenoptera is their as-

sociation with wood-boring insects. Thus the Orysside,

the most generalized group of Evaniide, the more primi-

tive Braconide, many of the structurally primitive Ich-

neumonide, and the Capitoniide are restricted to hosts

having such habits. This shows undoubtedly that such

habits have not easily been changed and that similarity

of host-habits is an important factor in determining what

insects may be attacked. This supports strongly our

thesis of the interrelation of taxonomy and habits.

In connection with the parasitism of certain chalecis-

flies, the French entomologist Marchal (’98) discovered,

some years ago, a most anomalous method of precocious

multiplication which he designated as polyembryony or

germinogony. In species exhibiting this phenomenon,

the embryo becomes dissociated into a large number of

parts, and from the numerous germs thus produced there

is formed a veritable swarm of minute parasites, the

extent of which is limited only by the available food

supply in the host. Marchal’s first observations have

been much extended since by himself and numerous other

workers, and the same condition has been found to exist

ny other Chalcidoidea and also in the Serphoidea

shal, 03). It has recently been recognized in an-

‘idely different family, the Dryinide by Korn-

719) and probably occurs sporadically in several

rasitic families, although I believe no other cases

havi ‘been absolutely substantiated. From the regular

association of numerous individuals in single hosts in

the case of Microgaster, allied genera of the Braconidae,

No. 637] PARASITISM 145

and in Spheropyx (Cushman, 713), in a few Ichneumonide

(e.g., species of Cryptus) and in a few Bethylida, it would

seem likely that they also are polyembryonie.

The widespread occurrence of germinogony and its

apparently erratic distribution show that it can be of no

general taxonomic interest at least as an aid to classifi-

cation. It is indeed quite the opposite, for the develop-

ment of the egg in the process of fragmentation is so

similar in the Chalcidoid and Serphoid that we might be

led to believe it of common origin. As their ancestors

were undoubtedly not polyembryonie, such can not be the

case and the process must have originated independently,

just as it has in several totally unrelated animals like

certain annelids (e.g., Helodrilus) which exhibit it in an

imperfect condition (Weber, 717) and in the armadillo

(Newman and Patterson, 710) among mammals where it

has become completely established. A quite similar

modification of development is seen in the formation of

the rediæ in the sporocysts of Distomes. Still similar,

but delayed until the larval stage, is the process of pxdo-

genetic multiplication in the Cecidomiid fly Miastor

(cf. Felt, 711), well known to all entomologists.

It appears from any general survey of the habits of

the parasitic Hymenoptera that we find certain taxonomic

groups of host commonly attacked by discrete groups of

closely related parasites. It is natural that such combi-

nations should impel our attention, as they may be fitted

with the least effort into a classified scheme, and further-

more their mere recurrence is sufficient to indicate that

they are not due merely to chance.

The following list includes a few striking instances of

this sort drawn at random from widely separate sections

of the order:

Parasites Hosts to Which They are Restricted

Families

_ Alysiidee Dipterous larve.

Trigonalide Vespide.

146 THE AMERICAN NATURALIST [Von LV

Subfamilies

Evaniine Cockroaches and their odthece.

Ichneutine Saw-fly larve.

Genera

Polygnotus Cecidomiid larve.

Coccophagus Soft scales.

If we should reverse the order of the above list and

attempt to tabulate groups of related hosts that are af-

fected only by certain groups of parasites we should have

great difficulty in finding examples. This, of course, is

to be expected on account of the passive condition of the

host and the active rôle of the parasite, whereby it first

came to infest some certain kind of host. Inheritance of

such specific instincts over long periods of time, during

which groups were becoming differentiated, will lead

naturally to the evolution of groups of parasites attached

to groups of hosts which have meanwhile been developed.

Such reasoning appears to be sound and may explain

some of the conditions tabulated above.

I think, however, that there is a deeper basis than this,

and that we can not fully understand such combinations.

without inquiring into the actual physiological relations

between host and parasite.

It has been customary among entomologists to place

great emphasis upon the maternal instinct of invariable

selection as determining and restricting the range of

hosts affected by specific parasites. Among zoologists

who deal with other parasites, particularly Protozoa and

lower invertebrates, no such idea has ever been enter-

tained, as the parasite plays a passive role in attaining

its host. The malarial parasite is ingested by all insects

that suck human blood, but is able to continue its para-

sitie life only in certain particular mosquitoes. Similarly,

a certain Cestode worm parasitic in birds has as interme-

diate host, the garden slug, from which the definitive host

obtains it by eating the slug. That this Cestode does not

occur in other hosts that may eat infected slugs is a phys-

No. 637] PARASITISM 147

iological matter and is always regarded as such by hel-

minthologists who encounter many instances of this kind.

On account of the definite requirements of such parasites,

Cobb (’04) suggested some years ago that they might

give valuable clues to the taxonomic affinities and physi-

ological peculiarities of their hosts, the latter particu-

larly in cases where there is a wide range of hosts.

In insects, and, quite fortunately for the present dis-

cussion, in the parasitic Hymenoptera, there are avail-

able some extremely pertinent observations made by

Timberlake (’12) relating to the fate of eggs in the

bodies of host insects in which they do not normally

develop. His experiments were made with an Ichneu-

monid, Linnerium validum, commonly parasitic in cater-

pillars of the fall web-worm. This parasite will also ovi-

posit in larve of various other moths, when persuaded to

do so in captivity, by depriving it of its normal host; but

it can not complete its development in the experimental

hosts. This is due to the death of the young larve, which

succumb to the reactions of the host soon after hatching,

or possibly in some cases even before hatching. The an-

tagonistic action of the tissues of the host is manifest by

the accumulation of amcebocytes about the unwelcome

objects. In one other abnormal host, the tent-caterpil-

lar, this Linnerium may survive and complete its trans-

formations, but there is a high mortality among the par-

asites, for many are destroyed by the host.

These experiments show very clearly why this parasite

is restricted to certain hosts and, from the nature of the

reaction, which is so similar to that exhibited by animals

in general toward microorganisms and other foreign ma-

terials, there is little reason to doubt that insects usually

react in this fashion.. This also furnishes an explanation

for the continued restriction of parasites to specific hosts,

based upon natural selection, since individuals choosing

unsuitable hosts will suffer a very material reduction in

the number of their immediate progeny. This is, I think,

especially important, as it takes much of the burden from

148 THE AMERICAN NATURALIST [Vou LV

the already greatly strained principle of the fixity of in-

stinct in the imaginal insect.

It also aids greatly in understanding the relation pre-

viously referred to, where extensive groups of parasites

attack discrete groups of host. Adaptation to one host

means ordinarily greater physiological suitability for

another closely related host than for a widely different

one. This, no doubt, applies to cases like the Alysiid

parasites, for here the series of hosts, while quite uni-

form, is so extensive that it can not be explained on the

slowly acting basis of concomitant differentiation of the

hosts and parasites.

Instances, like one cited by Pierce (’08) where several

species of parasites suddenly became abundant enemies

of the boll-weevil due to the scarcity of their more

favored hosts, must depend upon selection, as suggested

above, leading to the rapid improvement of partial adap-

tations.

We have already referred to the fact that the parasitic

Hymenoptera, and quite generally also most parasitic

insects, attack other insects, and pointed to this as a

characteristic more or less peculiar to insect parasitism

or at least to its most prevalent types.

The attachment to closely related animals as hosts is

shown still more clearly in Hymenoptera that are sec-

ondary parasites on parasitic species of the same order,

of the same family, or even of related genera. This phe-

nomenon is not restricted to Hymenoptera, but is most

extensively exhibited by them. Thus certain genera of

Ichneumonide, Braconide, and Chalcidoidea develop

regularly in the larve of primary parasites which become

established in a free living host.

Secondary parasites are not absolutely distinct from

primary ones in some individual cases, for this relation

is known to be facultative in a few species of Hymen-

optera which develop in either way. In 1903, Fiske (’03)

showed from careful breeding experiments that certain

Ichneumonide of the genus Itoplectis may be either pri-

No. 637] PARASITISM 149

mary or secondary parasites of the tent-caterpillar, at-

tacking a member of their own family in the latter case.

Since then other examples have come to light, but they

are by no means common. Another fact which is sig-

nificant in connection with secondary parasites is that

they are very generally much less particular than pri-

mary ones in restricting themselves to a small series of

hosts.

In searching for the origin of secondary parasitism,

it is certain that it must be derived from the primary

form, since it is naturally dependent upon the latter for

its mere existence. The only other possibility appears `

to be the assumption that the primary parasites were

free-living forms when first parasitized, and that they

have since developed parasitic habits of their own. As

the secondaries are frequently structurally reduced such

a supposition appears still more improbable.

If, then, secondary parasites are derived from primary

ones, what can have caused them to desert their free-

living hosts? We have already seen how the restriction

of hosts among primary parasites seems to have a physi-

ological basis, in that the reaction of the tissues of the

host has been shown (Timberlake, ’12) to eliminate para-

sites not adapted to it. In attacking insects very closely

related to themselves parasites should stand a much

better chance for successful growth, as the physiological

antagonism of all animals toward closely related forms

is much less than that toward very different ones. Young

larve of parasitic species should therefore meet with less

difficulty in developing in the bodies of related forms,

and secondary parasitism might arise with little diffi-

culty when eggs were placed in another parasite rather

than in the body cavity of the free-living host. This ex-

planation may account for the prevalent type of hyper-

parasitism, but not for cases like that of the Chalcidid

Dibrachys which attacks Hymenoptera and Diptera alike.

This may simply be a case of great adaptability in cer-

tain species like some mentioned in connection with pri-

150 ; THE AMERICAN NATURALIST [ Vou. LV

mary parasitism, although it may depend upon a general

similarity in the tissues of all entomophagous parasites,

or a less aggressive condition of the tissue in parasites

due to their generally secluded and protected environ-

ment. As the latter condition seems not unlikely, it prob-

ably acts regularly to make hyperparasitism an easily

acquired characteristic.

Striking divergencies, like the following, noted by

Swezey (’08), are of interest in this connection. In his

studies of Dryinid leaf-hopper parasites, he found a

Ceraphronid parasitic on a species of the related Dryi-

nidæ, although the group normally and abundantly para-

sitizes entirely different types of insects.

The adaptation of animals and plants in conformity

with the demands of diverse environmental conditions

is now an axiom among biologists. From its manifesta-

tions it is evidently a physiological adjustment which

leads secondarily to structural changes, and many con-

vergences in form and function are traceable to it. On

account of the close interdependence of plants and in-

sects it appears, in some instances at least, to exert an

indirect influence upon phytophagous insects (Brues,

’20), whereby a species may feed rather indiscriminately

on herbs, and another on woody plants, but not upon the

two in combination.

In the case of parasitic Hymenoptera there are many

instances which might be cited where environment ap-

pears to have exerted a direct influence upon the acqui-

sition of host relations and others where we must, I

think, believe the influence to be indirectly related to the

environment through a second insect, the host. This

rather obscure statement may be clarified by a few ex-

amples. From what we have said in connection with

hyperparasitism, it seems quite clear that a species which

may assume the role of either a primary or secondary

parasite, responds quite directly to the environment, in

this case the primary host, which may be either sound

or already infested by a parasite which is in turn at-

No. 637] PARASITISM 151

tacked. This influence seems to be a rather direct one.

On the other hand, I may quote from a previous paper

(Brues, ’08) the following: ‘‘The European Chalcid-fly,

Ormyrus tubulosus, has been minutely studied by Mayr,

who has bred it from no less than 27 species of Cynipid

galls, and I have from Massachusetts what is apparently

the same species, bred from about half as many North

American species by the late Dr. M. T. Thompson. The

galls formed by the various hosts of this species are

many of them entirely dissimilar in form, the only re-

semblance between them, aside from their gross gall-

like form, being their more or less uniform habitat at-

tached to twigs and leaves.’’ Howard (’91) mentions

Enrytoma rose as having over 50 cynipid hosts. A range

of hosts of this sort appears to be due not directly to the

environment of the host, but to the similar physiological

condition of the various Cynpids themselves, which, as

we have already said, are closely confined to a very nar-

row range of food-plants.

The great difficulties occasionally pkd upon para-

sites in attaining their hosts may be purely a matter of

environment, as illustrated by the following considera-

tions.

` An interesting series of parasitic Hymenoptera are

those which prey upon aquatic insects. In several well-

known cases, the behavior of the adult parasites has be-

come so profoundly modified that the females not only

enter the water in search of their hosts, but they may be,

occasionally at least, accompanied by the males. The

first observation of this sort was made nearly a century

ago by Francis Walker (’36) on Agriotypus, and the well-

known observer Sir John Lubbock (’63) later gave an

account of the habits of two aquatic Chalcis-flies in which

he describes the actual process of swimming. One species,

the Mymarid (Cataphractus cinctus) makes use of its

ciliated, paddle-shaped wings for this purpose, while the

other, a Trichogrammid (Prestwichia aquatica) propels

itself by means of the legs. Numerous other ana

152 THE AMERICAN NATURALIST [Vou. LV

tions, notably those of Von Siebold (’58), W. Miiller

(789), Marchal (’00), Rousseau (’08), Heymons (’08),

Schulz (’07, ’10*, 710°), and Matheson and Crosby (’12),

have added much of interest, not only in bringing to light

aquatic members of several families, but in determining

some of the host species upon which they prey. In many

eases the adaptation to aquatic life is not so perfect as

the cases just mentioned, although several other species

are known to swim readily, using either the legs or wings,

which usually show modifications adapted to such be-

havior. 7

In view of the frequent occurrence of aquatic imaginal

forms in other orders of insects such as the Coleoptera

and Hemiptera, it is perhaps not surprising to find cer-

tain parasitic Hymenoptera adopting this habitat.

Viewed more in detail, however, the matter is quite a

different phenomenon. Such Coleoptera as Gyrinide,

Hydrophilide, Dytiscide, etc., are uniformly aquatic in

both preparatory and imaginal stages, and such is also

true of the brachycerous Hemiptera. All of these insects

are highly modified to conform with their aquatic envi-

ronment, particularly in reference to the functions of

locomotion and respiration.

In the aquatic Hymenoptera, a series of families is

represented and only a comparatively small number of

genera are included. The structural modifications are

far less profound, indeed they frequently represent very

slight changes. They are more closely parallel to the

natatorial habit shown in isolated genera such as the rice

water-weevil, Lissorhoptus simplex, a beetle that has be-

come aquatic and oviposits in the roots of the rice plant

(Tucker, 712). It has been shown experimentally by

Szymanski (718) that many terrestrial insects may be

induced to swim if submerged and we may easily suppose

that the truly aquatic habit of the parasitic Hymenoptera

just mentioned may have arisen through the seeking of

their hosts in aquatic plants, first at or above the surface >

of the water, and later through a search for further indi-

No. 637] PARASITISM 153

viduals below the surface. Even memory could easily

play a part here, if the host were submerged during the

development of the parasite, and the latter emerged as

an adult below the water, from which it must escape by

locomotion through the water.

In the case of aquatic Hymenoptera, it is readily seen

that we can not correlate taxonomy with habits accord-

ing to any generalized scheme, although the several

genera show structural characters associated with their

unusual habits. Most striking is the number of Myma-

ride and Trichogrammatide included, minute insects

whose wings are naturally well suited for swimming.

Frequently a secluded habitat acts as a powerful fac-

tor in restricting the kinds of parasites that can attack

certain types of hosts. Thus, wood-boring insects can

be reached only by species provided with long oviposi-

tors. Such restrictions are clearly defined and many

other examples might be cited. Partial inaccessibility

of the host may even occur in the case of parasites other-

wise well suited to their host, as for example in the case

of a éommon parasite of the eggs of the gipsy moth,

which is able to oviposit only in the eggs occupying a

superficial position in the egg-mass of the host. Some-

times difficulties may be overcome by the presence of an

active first-stage larva. This may exhibit most extraor-

dinary behavior as has been described by Smith (717)

in the Chalsis-fly, Perilampus. Here the Perilampus

egg is deposited:and hatches away from the body of the

host as a planidium which later attaches itself to the

host and remains there till the host completes its growth,

after which the planidium begins its parasitic life. A

second species that is a hyperparasite seeks out the pri-

mary parasite in the caterpillar host through which it

bores its way and there awaits the exit of the primary

parasite before proceeding with its development.

Again, the female of some egg-parasites attach them-

selves to individuals of the host species and are thus

carried to the place where the eggs within which they will

154 * THE AMERICAN NATURALIST [Vou. LV

develop are to be deposited. Certain Chalsis-flies and

Serphoids have adopted this curious method of trans-

portation (Brues, 717) which occurs sporadically in di-

verse insects (Banks, ’11). The way in which many

modifications of this kind appear in similar form makes

it impossible to consider them as guides to taxonomic

affinity. The elongated ovipositor, the active first-stage

larva, and many other adaptations for attaining the host

are of course good taxonomic characters, but they reap-

pear independently in more than one group, and can be

used only in combination with characters of less vital

importance to the animal, to characterize completely any .

extensive groups. Nevertheless, the lengthened oviposi-

tor can be used to separate numerous families and

smaller groups in the parasitic Hymenoptera and as it

bears a certain relation to habits, the latter are thus re-

flected in taxonomy on a structural basis. However, the

habits of many such insects do not seem to require such

a long ovipositor and represent not the primitive habit

for the group, but recent modifications which break down

the homogeneous correlation of structure and habits.

Closely connected with the specific association of nat-

ural groups of hosts and parasites is the great variation

shown by different parasites in the number and diversity —

of the species that serve as their hosts. Just as we can

find among phytophagous insects, omnivorous forms,

strictly monophagous ones, and all intergrades between

the two, so there exists among parasites an almost equally

varied series of associations with one, several or many

hosts.

Although parasitic Hymenoptera are so abundant,

both in species and individuals, their food habits are not

so easily observed as those of plant-eating insects and

our knowledge concerning them is far less complete. The

large number of secondary parasites also lead to con-

fusion, as these may not always be distinguished on a

structural basis. oe

If parasitism demands a nice physiological adjust-

No. 637] PARASITISM 155

ment, we might expect to find that egg-parasites affecting

the organism at an earlier and less highly differentiated

stage of ontogeny, are more catholic in their tastes. This

is, however, not borne out by observation to any extent,

and egg parasites are usually as closely restricted to

particular hosts as their relatives who confine their atten-

tion to larval insects.

Small size is a prerequisite of all true, internal egg-

parasites except a few that occur in the oothece of cock-

roaches, where the comparatively large species of Evania

undergo their development. Some parasites oviposit in

the host-egg, but live at the expense of the larva; they

are, except in polyembryonic forms, larger, and not

classed as egg-parasites.

On the basis of size, then, practically all egg ‘parasites

are either Chalcidoidea or Serphoidea and this habit

characterizes a number of families, and smaller taxo-

nomic groups (cf. Girault, ’07, ’11). Among them the

strange, tropicopolitan,genus Podagrion attacks only the

eggs of Mantide. The large cosmopolitan genus Tele-

nomus occurs in the eggs of various insects, mainly Lepi-

doptera while the very similar genera Phanurus and

Trissolcus are restricted to eggs of Tabanide and Penta-

tomide. Again, Scelio and several related genera at-

tack only the eggs of the Orthoptera Saltatoria. Thus,

if used with due caution, egg-parasites are in the main

illustrative of close correlation between the taxonomy of

host and parasite in spite of the fact that we may nat-

urally regard insect eggs as more similar inter se, than

insect larve.

It is true that the ubiquitous little Trichogramma af-

fects eggs of several orders and many families of insects,

but like other less conspicuous examples, it stands quite

apart from its commonplace associates.

With their larger and more variable size, and great

diversity in habits and structure, larval insects present a

correspondingly varied series of opportunities for para-

sites. We find also that practically no genera are known

156 THE AMERICAN NATURALIST [Vou. LV

to parasitize both eggs and larve, although the poly-

morphic and widespread Eupelmus among the Chalcis-

flies appears to be an exception. As the eggs and larve

of many insects frequently occur together at the same

time, this fact is rather surprising and shows that the

parasitic association must depend greatly upon gross

form, as well as upon the factors of environment and

specific physiological reactions, which we have already

mentioned. One case which comes to my mind in this

connection is quite instructive and there are no doubt

others of a similar nature. All the several genera in-

cluded in the Evaniine are, as previously mentioned,

parasites in the egg-cases of cockroaches, with the excep-

tion of a single reliable record (Picard, 713) of the rear

ing of Zeuxevania from the body of the blattid itself.

Quite likely the future may bring forth other similar ob-

servations on Evaniines, but this one shows that parasit-

ism has been transferred to the cockroach from the

oötheca, which is of course carried about by the female

for some time before deposition.

Larval parasites have been more extensively reared

than those living in eggs and their habits are conse-

quently better known. Many observations upon indi-

vidual species of hosts show that the larval stages harbor

a far more extensive series of parasites (e.g., Howard

and Fiske, 712) than the eggs or pupx, while hymen-

opterous parasites of the adult are almost unknown.

Among larval parasites it is easy to recognize two gen-

eral series, so far as the number of hosts utilized. Some

species are very conservative in this respect and others

extremely versatile. These two terms are equally suit-

able for genera and larger groups, and the difference is

more important when it involves all or most of the species

of quite extensive groups. Thus the highly modified

members of the family Dryinide (Perkins, ’05) are re-

stricted to several families of Homoptera. A few which

parasitize Membracids are insects of quite ordinary ap-

pearance, but the remainder affecting Tettigoniellids and

No. 637] -` PARASITISM 157

Fulgorids have the fore tarsi of the females misshapen

to form chelx or pincers, by means of which they cling

to their host. Such structures are elsewhere unknown

among insects. The group has become highly specialized,

apterous in several genera, and has probably reached the

end-stage in its evolution. Like all creatures which have

attained this condition, it shows no further adaptiveness

in habits. This is a clear-cut case of correlation between

habits and taxonomic affinities.

Versatile groups naturally include large numbers of

genera and species with varied habits which enable them

to grasp every opportunity to earn (or, in the case of

parasites, to steal) a livelihood. Numerous species and

varied habits, are as inseparable as form and function.

The former binary involves an added series of factors,

since any group of insect parasites comes into keen com-

petition with the members of other groups as it reaches

out for new hosts. Some have spread widely among

hosts of very similar types, restrained by some insuper-

able obstacle, probably physiological in nature, from at-

taching themselves to strange insects. They show a cor-

relation between habits and structure. Others have broken

their fetters more quickly and completely, and adapta-

tions in habit have far outstripped structural modifica-

tions, resulting in natural taxonomic groups which show

only imperfectly such correlation.

The climax in this direction is reached by certain

groups which have cast aside parasitism entirely and

become phytophagous. This has occurred independently

in several families of Chalcis-flies, a group in which the

struggle for existence must be very severe. One of these

aberrant series, Megastigmus and its allies (Crosby, ’13)

feed within the seeds of plants, mainly those of various

trees, upon a rich protein diet, probably similar to that

of their entomophagous forebears. Another, Isosoma

and its allies (Howard, ’91 and ’96; Phillips and Emery,

19) occur in far less delectable vegetable tissue, such

as the culms of grasses in which they sometimes cause

158 THE AMERICAN NATURALIST [Vou. LV

galls. <A third (Mayr,’’05), including some genera re-

lated to the remarkable parasitic Perilampus, which we

mentioned a few moments ago, produce conspicuous

galls on certain plants.

The production of galls by the phytophagous Chalcids

is quite suggestive, since many forms related to Isosoma

(Harmolita) are parasites of gall-making Cynipids.

Megastigmus also belongs to a group including many

parasites of Cynipids. Since we do not know exactly

how galls are formed, however, the matter can not be

profitably discussed at the present time.

Although they may not aid us greatly in formulating

any general causes leading to divergence in habits among

related forms, I should like to append a few observations

made by various entomologists which suggest a variety

of factors.

The effect upon the parasite of almost complete elimi-

nation of a host through excessive parasitism has often

been commented upon by entomologists. An especially

clear case has been described by Aldrich (712) where an

invasion of the western pine-butterfly was suddenly

checked by Theronia fulvescens. The parasite then found

it necessary to eke out an existence from scattering and

less suitable forest insects and under such stress, selec-

tion must be very keen (cf. p. 147). Complete parasitism —

of 100 per cent. of the related cabbage butterfly by Apan-

teles glomeratus has also been reported by Chittenden

(705).

Errors or aberrations of instinct have also been occa-

sionally observed. Thus Marchal (’07) saw a Chaleidid

parasite of coceinellids (Lygellus) repeatedly oviposit in

the pupal exuvia when living material was not available.

Still more incongruous is the behavior of Trichogramma

observed by Holloway (’12) who found this insect actually

ovipositing in small globules of partly solidified plant

juice on the foliage of okra plants. One of the common

hosts of this egg-parasite, the cotton boll-worm, fre-

quently oviposits on the leaves of okra and the globules

No. 637] PARASITISM : 159

were evidently mistaken for eRe -eggs. Premature ovi-

position is generally attributed to physical necessity in

relieving the pressure in the body, but here at least it is

accompanied by the outward appearance of instinct.

Whether this Chalcid tasted the strange new Łost is not

stated, but it is a common procedure among Chalcids

(Howard, ’10) to tap the host with the ovipositor, and to

lap up the exuding body-juices quite independently of

egg-deposition. What her reasons for this may be are

obscure; possibly it is to test the suitability of the host;

perhaps to secure food, or she may even retain a specific

appetite for the kind of food consumed in her earlier

days.

Marchal, Vayssiére (’07) and Loiselle (’08) have com-

mented upon the retarded development observed in cer-

tain Ichneumon-flies whereby emergence of some indi-

viduals was delayed a year. Such occurrences might

serve to bridge over the gap of a season when host in-

sects were scarce; on the other hand, if the time were not

exactly twelve months it easily might lead to a new

‘trial’? association in the absence of the proper host at

that season.

That these factors might lead to divergence in habits,

I can not. doubt, but hesitate to apply them to any con-

crete cases of aberrant habits.

BIBLIOGRAPHY

Aldrich, J. M.

1912. Note on Theronia fulvescens. Jour. Econ. Entom., Vol. 5, pp.

87-88.

Ashmead, W. H.

1894. Descriptions of Two New Hymenopterous Parasites from Water

Beetles. Canadian Entom., Vol, 26, pp. 24-26.

1900. Some Hymenopterous Parasites ron Dragon-fly Eggs. Entom.

News, Vol. 11, pp. 615-616.

Banks, N.

1911. Cases of Phoresie. Entom. News, Vol. 22, pp. 194-197.

Baumberger, J. P.

1919. A Nutritional Study of Insects, with Special Reference to

Microorganisms and their Substrata, Jour, Exper. Zool., Vol.

28, 1-81, figs. 18,

160 THE AMERICAN NATURALIST [Vou LV

Bradley, J. C.

1908. The Evaniide, Ensign-flies, an Archaic Family of Hymenoptera.

ms, American Entom. Koc., Vol. 34, pp. 101-194, pls, 11.

Brues, C. T.

1908. The Correlation between Habits and Structural Characters

among ane Hymenoptera. Jour. Econ, Entom., Vol

123-12

1910a. The ere Hymenoptera of Ea Tertiary of Florissant, Col-

orado. Bull. Mus. Comp. Zool., Harvard, Vol. 54, pp. 1. -125,

figs. 88, pl. 1.

1910b. Some Notes on the Geological History of the Parasitic Hy-

menoptera. Bull. New York Entom. Soc., Vol. 18, pp. 1-22,

figs. 5.

1917. Adult Hymenopterous Parasites attached > the vated » their

Host. Proc. Nat. Acad, Sci., Vol, 3, pp. 136-140, fig.

1919. A New Chaleid: iy Parasitic. on the "Australian Buli- oe Ant,

Ann. Entom. Soc. America, Vol, 12, pp. 13- ey

1920. The Selection of Food- aiiis by Insects, with Special a a

to oe ere Larve. AMERICAN NATURAL., Vol.

313-3

Brues, C. T., an ak ae: ser.

1921. A Symbiotic Fungus Occurring in the Fat-body of Pulvinaria

innumerabilis. Biol. Bull., in pros:

Chittenden, F. H.

905. An Instance of Complete Parasitism of the Imported Cabbage

Worm. Bull. U. 8. e Agric., Bur. Entom., No. 54, p:

Cobb, N. A.

1904. Parasites as an Aid ih Determining Organic Relationship.

Agric, Gaz. N. S. Wales, Vol. 15, pp. 845-848

Crosby, C. R.

1913. A Revision ya e North American Species of Megastigmus

alman . Entom. Soe. America, Vol. % pp. 155-170,

figs. 10.

Crosby, C. R., and R. Matheson,

191 rE Insect Enemy of the Four-lined Leaf-bug (Pæcilocapsus

lineatus. Fabr.). Canadian Entom., Vol. 47, pp. 181-183,

e 4.

Cushman, R.

1913. e Notes on a few Rare, or Little-known Parasitic Hy-

menoptera. Proc. Ent. Soc. Washington, Vol. 15, pp. 153-160.

Dyar, H. G.

1899. Note on an External-feeding Hymenopterous Parasite. Proc.

Entom. Soc, Washington, Vol. 4, pp. 23

Felt, E. P.

1911. Annual Report of the Entomologist. Bull. New York State

us., No. 147, pp. 1-180, pls. 35.

Ferton, Ch.

1890. L’évolution de 1’instinet chez les Hyménoptéres. Rev. Scient.,

Vol. 45, pp. 496-498 :

No. 637] PARASITISM 161

Fiske, W. F.

1903. A ees of the Parasites of the American Tent Caterpillar.

Tee ull, New Hampshire Expt. Sta., No. 6, pp. 185-230,

a 6.

Giard, A.

1896, Retard dans l’évolution déterminé par anhydrobiose chez un

xe png tees chaleidien. C. R. Soc. Biol. Paris (10), Vol. 3,

pp. 837-389.

Girault, A. A.

1907-1911. Hosts of Insect Egg-parasites in North and South America,

Part I, Psyche, Vol. 14, pp. 27-39. Part II, Psyche, Vol. 18,

pp. 146-153

Graenicher, S.

1905. Some Observations on the Life-history and Habits of Parasitic

Bees. Bull. Wisconsin Nat. Hist. Soc., Vol. 3, pp. 153-167,

ples

= ea A.

1908. Die fossilen Insekten und die Phylogenie der rezenten Formen.

Leipzig

Heymous, R.

1908. Me pe a aus der Umgebung Berlins. Deutsch.

tom. Zeitschr., 1908, pp. 137-150

Halloway, T. E

1912. An Experiment on the Oviposition of a Hymenopterous Egg

Parasite. Entom. News, Vol. 23, pp. 329-330.

Hood, J. D.

1913, Notes on the Life Hstory of Rhopalosoma poeyi Cresson. Proc.

Ent. Soc, Washington, Vol. 15, pp. 145-147, fig. 1

Howard, L. O.

1891, The ee of the Hymenopterous Insects of the Family Chal-

æ. Proc. U. 8. Nat. Mus., Vol. 14, pp. 567-588.

1893. tthe reese of Structure and Host-relation er the En-

eyrtine. Wilder Quarter-century Book, pp. 177-1

1896. The Grain Joint-worm Flies and their Enemies: A Peete

of Some North American Phytophagie Eurytomine. Bull. U.

S. Dept. Agric., Div. Entom., No. 2, pp. 24, figs. 10

1897. A Study in Insect Parasitism: A Consideration of the Parasites

of the White-marked Tussock Moth, with an Account of their

gr ena and with Descriptions of New Species. Bull.

Dept. Agric., Bur. Entom. Tech, Ser. No. 5, pp. 57,

figs

1910. cad Habit with certain Chalcidoidea of Feeding at Puncture

Holes made by the Ovipositor. Journ. Econ. Entom., Vol. 3,

pp. 257-260.

1911. A New Species of Coccophagus with a Table of the Host Rela-

tions of those Species of the Genus known to the Writer.

Journ. Econ. Entom., Vol. 4, pp. 276-277.

162 THE AMERICAN NATURALIST [Vou. LV

Howard, L. O., and Fiske, W. F.

1912. The Pape aE into the United States of the Parasites of the

Gipsy Moth and the Brown-tail Moth. Bull. U. S. Dept.

Agric., Bur Entom. No. 91, pp. 344, illus.

Kinsey, A.

1920. Phylogeny of Cynipid Genera and Biological sas

. American Mus. Nat. Hist., Vol. 62, pp. 357-

Kornhauser, 5 T

1919. Jer Sexual Characteristics of the Membracid, Thelia bimacu-

a (Fabr.). Journ. Morphol., Vol. 32, pp. 531 1-636, figs. 54.

Laloy, L.

1906. Parasitisme et mutualisme dans la nature. Paris, Félix Alcan,

0, pp. fi 82

Loiselle, A.

1908. Sur 1’ éclosion tardive de certains parasites. Bull. Soc. Entom.

France, 1908, pp. 213-214.

Lubbock, J.

1863. On two Aquatic Hymenoptera, one of which uses its wings in

swimming. rans. Linn. Soc., Zéol., Vol. 24, pp. 135-141,

figs. 6.

Marchal, P.

18 Formation d’ une espéce par le parasitisme: Étude sur le

Sphecodes gibbus. Rev. aap: Vol. 45, pp.

1898. La dissociation de 1’ oeuf en un grand nombre d’ individus

distinets et le cycle évolutif si 1? Encyrtus fuscicollis. C.

. Acad. Sci. Paris, Vol. 126, pp. 662-664

1900. ee un aah Hyménopière peara le Limnoðytes gerri-

gus. . Ent. France, Vol. 69, pp. 171-1

1903. la i Guile ri Potoa ai (Lindm.), ja Soe.

Entom. France, 1903, pp. 90-93.

1907. Sur a Lygellus epilachne Giard (parasitisme; erreur de 1’

inet; évolution). Bull, Soc. Entom. France, 1907, pp.

Matheson, R., and Crosby, C. R.

1912. Aqu atic Hymenoptera in America. Ann. Entom. Soc. America,

. 5, pp. 65-71, figs. 3.

Mayr, G.

` 1905. Ueber Perilampiden. Verh. zool.-bot. Gesellsch. Wien, Jahrg.

» Pp- 549-571, figs. 3. -

Mrázek, A.

1909. Fleischfressende Blattwespen. Zeits. wiss. Insektenbiol., Vol.

Müller, W.

1889. Ueber Agriotypus armatus. Zool. Jahrb. Abth. f. Syst., Vol.

Newman, H. H., and Patterson, J. T.

1910, The Die of the Nine-banded . Armadillo. Journ.

Morph., Vol. 21, No. 3 :

No. 637] PARASITISM 163

Perkins, R. O: L.

1905. Leaf-hoppers and their Natural Enemies. Part I, Dryinide.

Bull. Div. Entom., Hawaiian Sugar Planters’ Assoc., No. 1,

pt. 1, pp. 69.

1918. Sur le genre Zeurevania Kieffer et sur les moeurs du Zeuze-

vania splendidula Costa. Bull. Soc. Entom. France, 1913,

pp. 301-304, fig. 1.

1919. La faune entomologique du figuier. Ann. Service ‘Epiphyties,

Vol. 6, pp. 35-174, figs, 34.

Pierce, W. D.

1908. Factors Controlling Parasitism, with Speciat Reference to the

tton Boll-weevil. Journ. Econ. Entom., Vol. 1, pp. 315-323.

Pierce, W. D., et a

1912. The Insect Enemies of the Cotton Boll-weevil. Bull. U. S.

‘Dept. Agric. Bur, Entom. No. 100, pp. 99, figs.

gene W. J., and Emery, W. T.

919. A henii of the Chalcid-flies of the genus Harmolita in

America North of Mexico. Proe. U. 8. Nat. Mus., Vol. 55,

pp. 433-471, pls. 9.

Rohwer, S. A.

1913. peny on Rhopalosoma poeyi.) Proc. Ent. gra Washing-

15, pp. 147-148.

1917. Tasca, a new Suborder of Hymenoptera, with notes on:

the Immature Stages of slew Proc. Ent. Soc. Washing-

ton, Vol. 19, pp. 89- 98.

Rousseau, E.

1967. Les Hyménoptéres aquatiques, avee déscription de deux espèces

nouvelles par W. A. Schulz. Ann. Biol. lacustre, Vol. 2, pp.

888—401.

Schulz, W. A.

1907. R Braconiden. Ann. Soc. Ent. Belgique, Vol. 51,

pp.

1910a. Stsvanser Hymenopteren aus dem See von Overmeire. Ann.

e, Vol. 4, se ea pls. 2, figs. 2.

19106. Naser Beitrag zur Kenntnis der Witiecmann: Ann. Biol,

l e, Vol. 4, pp. eats: figs. 4.

Smith, H. S. ;

1912. The Chalcidoid Genus Perilampus and its Relations to the

blem ml Parasite Introduction. Bull. U. S. Dept. Agrie.

Bur. Entom. Tech. Ser., No. 19, pt: 4, pp. 33, 69, figs. 31.

1917. The Habit pe Leaf-oviposition Among the Parasitic Hymen-

optera. Psyche, Vol. 24, pp. 63-68, figs. 4.

Swezey, O.

1908. On Peculiar. Deviations from Uniformity of Habit Among

Chaleids and Proctotrupids. Proc. Hawaiian Entom. Soc.,

Szymanski, J. S.

1918. Das Verhalten der Landinsekten k eena gegenüber. Biol.

Centralbl., Vòl. 38, pp. 340-344, fi

164 THE AMERICAN NATURALIST [Von. LV

Timberlake, P. H.

1912. Experimental Parasitism: A Study of the Biology of Linnerium

an (Cresson). Bull. U. S. Dept. Agrie. Bur. Entom.,

h. Ser., No. 19, pt. 5, pp. 71-92.

Tower, D a

1916. Comparative Study of the Amount of Food Eaten by Para-

sitized Larvæ of Ae unipuncta. Journ. Agric. Research,

Vol. 6, pp. 455-45

Tucker, E. S.

1912. The Rice Water-weevil and Methods ton its Control. Roz

U. S. Dept. Agric., Bur. Entom., No. 152, pp. 20, figs.

Vayssière, A.

1907. Eeloni d? un Cryptus leucopygus. Bull. Soc. Entom.

France, 1907, pp. 311-312.

Von Siebold, C. T. E.

1858. Ueber Agrio in le picicorne. Ber.

samml. Nae. pe 1858, p. 211.

Walker,

aC Aoir armatus. Entom. Mag., 1836, p. 412.

Ward, H. B.

1907. The Influence a Parasitism on the Host. Science, N. S., Vol.

25 0l-

Weber, R. A.

1917. Observations on the Structure of Double Monsters in the Earth-

worm. Biol. Bull., Vol. 33, pp. 339-348, pls.

Wheeler, W. M.

1904. A New Type of Social inapyncone in Ants. Bull. American

us. Nat. Hist., Vol. 20, pp 5.

1907. The Poirie of co with an Account of Some Singular

ka ormalities Due to Parasitism. Bull. American Mus.

, Vol. 23, pp. 1-93, pls. 6.

1919. Ths i ator’ Katata a Study in Evolution. Proc. American

Philos. Soc., Vol. 58, pp. 1-40, 6 tables

ONCE MORE THE SUCKING-FISH

PROFESSOR LEO WIENER

Harvarp UNIVERSITY

Is 1919, while Professor E. W. Gudger’s excellent

series of articles ‘‘On the Use of the Sucking-fish for

Catching Fish and Turtles” appeared in THE AMERICAN

Natvrauist, I was at work on my first volume of “‘ Africa

and the Discovery of America,’’ where I had to touch on

the remora story in the early voyages to America, in

order to show that they were all a myth, based on the lit-

erary influence of Odoric of Pordenone on Columbus.

As my sources were naturally of a different character

from those of Professor Gudger, who was chiefly inter-

ested in the zoological side of the question, I was able

to supplement his thorough discussion with a number of

new data, which the zoologist will not consider to be

amiss.

The remora was dimly known to all the Arabic voy-

agers. We meet with it in the middle of the ninth cen-

tury, in the very beginning of the ‘‘Chain of Chron-

icles.’’!

In the Indian Ocean there is a fish, twenty cubits in length, in whose

belly there is a fish of the same kind, in whose belly there is similarly

a third fish. All these fishes are alive and moving. This large fish is

called al-wal. In spite of its size it has for its enemy a fish only a

cubit in length, called el-leshek. When the large fish becomes angry

and attacks the other fishes in the sea and maltreats them, the little fish

takes charge of him: it attaches itself to the root of his ear and ‘does

not let go of him until he is dead. The little fish also attaches itself

to boats, and the large fish dares not approach it, because of the fear

with which it inspires him. :

1M. Reinaud, ‘‘ Relation des Voyages faits par les Arabes et les Persans

dans 1’Inde et a la Chine dans le IX° siècle de 1’ére chrétienne,’’ Paris,

1845, Vol. I, p. 2f.

165

166 THE AMERICAN NATURALIST [Vou. LV

This account is obviously an exaggeration of some

story about the shark, but wal was soon identified with

the whale, as appears from the later Arabic sources. A

century later Mas‘iidi wrote:

There is a fish in this sea called el-Owal, which is from four to five

hundred ‘Omari cubits long; these are the cubits in use in this sea.

The usual length of this fish is one hundred perches. Generally the head

of the whale is out of the water; and when it powerfully ejects water,

it gushes into the air more than one bowshot high. The vessels are

afraid of it by day and night, and they beat drums and wooden poles

to drive it away. This fish drives with its tail and fins other fish into

its open mouth, and they pass down its throat with the stream of water.

When the whale sins God sends a fish about one cubit long called esh-

shak (al-leshek, as-sal), it adheres to ihe root of its tail and the whale

has no means to make itself free from it. It goes therefore to the

bottom of the sea and beats itself to death; its dead body floats on the

water and looks like a great mountain. The fish called esh-shak, ad-

heres frequently to the whale. The whales notwithstanding their size,

do not approach vessels; and they take flight when they see this little

fish, for it is their destruction.”

Idrisi merely said that in the Sea of Oman the wali,

which is of white color and one hundred cubits in length,

is usually accompanied by the small leshek, which kills

it. Ad-Damiri definitely identifies the soba fish with

the bal, the whale.

When it begins to tyrannize the other animals of the sea, God sends

a fish about a cubit in length, which attaches itself to its ear, and the

bal seeing no means of freeing itself from it, goes down to the bottom

of the sea and strikes its head on the ground until it dies, after which

it floats on the top of the water like a big mountain, and the men on

the East Coast of Africa are generally on the lookout for it. When

they find it, they plunge harpoons on it and drag it to the shore where

they cut open its belly and take out of it ambergris.*

The important point in all these stories, which ob-

viously emanate from the same original account, is that

2 A. Sprenger, ‘‘El Mastidi’s Historical Encyclopedia, pour ‘ Mead-

ows n. Gold and Mines of Gems,’ ’’ London, 1841, Vol. I, p. 2

P. A. Jaubert, ‘‘ Géographie d’Edrisi,’’ Paris, 1836, Vol ; 3.

4 Sire ’s ‘Hayat al-Hayawan’’ (a zoological lesicon); ivhadinted by

A. 8. G, Jayakar, London, 1906, va I, p. 237.

No. 637] ONCE MORE THE SUCKING-FISH 167

the remora is found off the coast of Zanzibar, where it

is in the same way connected with the catching of large

fish. But we have a circumstantial report of the em-

ployment of the sucking-fish in the catching of sea tur-

tles in João dos Santos’ ‘‘ Ethiopia oriental, ” which was

published in 1609:

The fishermen kill turtles at sea [along the coast of Mozambique]

in a strange manner. First they eatch in certain parts of the sea

among the rocks near the coast a kind of fish two spans in length,

called by the Moors sapi, which is as much the enemy of the turtle

as the ferret is of rabbits. The sapi has a very dark grey skin in-

clining to black, and a long thin head ending in a snout similar to that

of a pig. Its neck is about half a span long, on the back of whieh is

a shell of the same length and three fingers wide, which is formed of

hard and porous furrowed skin with which it clings to the stones as

leeches do, and it has the same faculty of sucking blood. For this

or legs with this shell, and sucks its blood until it is satiated, leaving

the turtle nearly dead, it being unable to noan} or get away, as it is

large and unwieldy and the sapi very nimble.

When the fishermen have caught some of these sapis they put them

in a basin of salt water and take them in the boat with them. They

tie a long line to their tails and then put out to sea in search of turtles,

which usually swim on the surface of the water. When they catch

sight of a turtle they throw out the fish fastened by the tail, as one

lets loose a ferret in a leash after a rabbit, and the fish immediately

attacks the turtle with as great force as if it was free and had received

no harm from the hook with which it was caught, or as if it was not

itself a prisoner. When it reaches the turtle it fastens on it so tightly

that it never looses its hold, and as soon as the fishermen feel that

it had done so they pull in the line and draw it over the water without

its loosening its hold, and the turtle, although very big and heavy, is so

dominated and tormented by the fish that it does not fight with it, but

lets itself be carried off easily because of the pain it suffers while they

are pulling it in, as at that time the fish grips it much tighter. Thus

the turtle is brought to the side of the boat, when the fishermen quickly

seize it in their hands and lift it in, and the fish they put back into its

basin. In this manner they catch a number of turtles.®

In the Bantu language of Zanzibar we have ‘‘kassa,

turtle; the kassa is caught by means of the taza fish,

d ‘

5 George McCall Theal, ‘‘Records of Southeastern Africa’’ (London),

1901, Vol. VII, p. 325 ff.

168 THE AMERICAN NATURALIST [Vou. LV

which the fishermen carry alive with them; when they

see a kassa, they let the taza go after it to stick fast to

the kassa; when the taza has seized it, the fisherman

throws a harpoon and takes the kassa out of the sea, the

taza letting go instantly when exposed to the air.’’® The

same dictionary gives tasa ‘‘a kind of fish which serves

as a bait for turtles,’’’ but the other dictionaries give for

it chazo, which name is also recorded by Professor

Gudger. Kassa for ‘‘turtle’’ is of extremely wide dis-

tribution and is not primarily a Bantu word, although it

is also found as kasi in Tete, that is in the region to

which dos Santos refers.

The oldest forms on record for this word are Sanskrit

kacchapas, kagyapas, Avestan kasyapa, hence Persian

keshef, Afganistan kasph, Singhalese keshew, Hindu-

stani kacchua, kaccha. It is therefore certain that the

turtle fishing was brought to the shores of Zanzibar

from somewhere in the Indian Ocean. This is in keep-

ing with the frequently recorded tortoise-shell trade in

the Indian Ocean, but ‘‘opposite the Ganges there is an

island in the ocean, the last part of the inhabited world

toward the east, under the rising sun itself; it is called

Chryse, and it has the best tortoise shell of all the places

on the Erythraean Sea.’’* The Chryse Island has been

identified with the Malacca peninsula,® hence the origin

of the practise of catching turtles with the remora is

most likely to be referred to the East Indies, whence it .

traveled eastward, to the Torres Strait and Melanesia,

and westward to the eastern shores of Africa.

The earliest definite reference to the remora fishing is

contained in a version of the cormorant fishing, as told

by Odoric of Pordenone and for the first time printed in

Ramusio in 1574, although it can be shown that it was

6 L. Krapf, ‘‘A Dictionary of the Suahili Language,’’ London, 1882, pp.

130 f.

7 Ibid., p.

8 W. H. bai ‘‘The Periplus of the Erythraean Sea,’’ London, 1912,

48.

9 Ibid., p. 259.

No. 637] ONCE MORE THE SUCKING-FISH 169

already in existence in the fourteenth century. Odoric

of Pordenone told how he came

to a certain great river, and I tarried at a certain city (called Belsa)

which hath a bridge across that river. And at the head of the bridge

was a hostel in which I was entertained. And mine host, wishing to

gratify me, said: “If thou wouldst like to see good fishing, come with

me.” And so he led me upon the bridge, and I looked and saw in

some boats of his that there were certain water-fowl tied upon perches.

And these he now tied with a cord round the throat that they might

not be able to swallow the fish which they caught. Next he proceeded

to put three great baskets into a boat, one at each end and the third

in the middle, and then he let the water-fowl loose. Straightway they

began to dive into the water, catching great numbers of fish, and ever

as they caught them putting them of their own accord into the baskets,

so that before long all the three baskets were full. And mine host then

took the cord off their necks and let them dive again to catch fish for

their own food. And when they had thus fed they returned to their

perches ane mee tied up as before. And some of those fish I had for

my dinn

This is followed by another kind of fishing :

The men this time were in a boat, wherein they had a tub of hot

water; and they were naked, and had each of them a bag slung over

his shoulder. Now they dived under water (for half a quarter of

an hour or so) and caught the fish with their hands, stowing them in

those bags that they had. And when they came up again they emptied

the bags into the boat, whilst they themselves got into the tub of hot

water, and others went in their turn and did as the first; and so great

numbers of fish were taken.**

The second kind of fishing is interesting from the fact

that it was much earlier told by Idrisi as in use at Zan-

zibar.

These people (at Meduna) fish in the sea without boats. They fish

y swimming, with small nets spun from grass and manufactured by

them. They tie these strings to their feet by means of knots which

they hold in their hands, they draw the strings of the net together the

moment they feel that the fish have entered, and this they do with an art

in which they excel, and with rules in which they have long experience.

10 Sir Henry Yule, ‘‘Cathay and the Way thither,’’ London, 1913, Vol.

II, pp. 188 ff.

11 Ibid., pp. 190 f.

170 THE AMERICAN NATURALIST [ Von. LV

To attract the fish they use land reptiles. Although they live in a

state of great distress and misery, these people (God loves those who

reside at their family hearths) are satisfied with their lots and with

what they have. They are under the government of Zanzibar.1?

Yule'® cites Fortune and Dabry for the same custom in

China, which once more shows the wide distribution of

identical maritime customs from Zanzibar to China.

The first kind of fishing has undergone all kinds of

changes in the very earliest Odoric manuscripts. Sir

John Mandeville, who cribbed so much out of Odorie,

tells of a fish-otter, instead of a cormorant, as the ani-

mal with which Sah, are caught.

In that country there be beasts taught of men to go into waters, into

rivers and into deep tanks for to take fish; the which beast is but little,

and men clepe them loirs. And when men cast them into the ae

anon they bring up great fishes, as many as men will. And i

will have more, they cast them in again, and they bring up as scary as

men list to have.14

It is interesting and important to observe that the

Italian version of Sir John Mandeville which came out

in 1491,'5 that is, one year before the discovery of Amer-

ica, has the same story, the French term loir for ‘‘otter’’

being here rendered by udria. The Latin version, of

- about 1500,1* simply says:

Tamed water dogs whom we call luteres, are here aplenty; every time

- they are sent into the river, they bring out

The substitution is everywhere from Vincent of Beau-

vais, who in his ‘‘Speculum naturale,” XX, 89, tells of

the same dukcotiogs with which fish are enaght: but the

substitution is unquestionably older than Sir John Man-

deville’s, who would not have omitted the strange story

of the cormorant if he had found it in his copy of Odoric.

12 Op. cit., pp. 55 f.

13 Op, cit., .

14 ‘í The Travels of Sir John Mandeville,’’ London, 1900, p. 136.

15‘ Tractato delle piu maravigliose cose e piu notabili,’’ Venice, Nicolaus

de Ferrariis, 17 Nov., 1491, cap Il.

16 ‘* Johannis de Montevilla Ttinerarius in partes Iherosolimitanas,’’ cap.

XXXI

No. 637] . ONCE MORE THE SUCKING-FISH — Hi

The Latin version of Odoric has the old cormorant

story’ where the bird is called mergus, while in the Ital-

ian version it is called smergo% The usual Italian

names for the cormorant which Odoric must have known,

are also mergo, maragone,’® so that the Latin mergus is

formed from the Italian mergo. Curiously, there are

two versions of Odoric in Ramusio. In the first the

whole cormorant fishing episode is omitted, while the

second has a totally different account. Here we read:

Mine host... took us to one side of the bridge where the river was

wider, and there we found many boats, and there was one of them em-

ployed in fishing by aid of a certain fish called marigione. The host

had another such, and this he took and kept it by a cord attached to

a fine collar. And this indeed is a creature that we have seen in our

own seas, where many call it the sea-calf. It had the muzzle and the

neck like a fox’s, and the fore paws like a dog’s, but the toes longer,

and the hind feet like a duck’s, and the tail with the rest of the body

like a fish’s. Mine host made him go in the water, and he began to

catch quantities of fish with his mouth, always depositing them in the

boat. And I swear that in less than two hours he had filled more than

two big baskets.?°

It is clear that the description of the sea-calf is an ex-

aggeration of that of the fish-otter, which is in Arabic

called ‘‘fox of the water’’ or ‘‘dog of the water.’’ Hence

there is most likely here an Arabic influence which

caused the substitution. And the reference to a fish

marigione, which was kept by a cord attached to a fine

collar, is similarly an attempt to bring the cormorant

story in keeping with the Arabic and Zanzibar account

of the fishing with the remora. We have here a transi-

tional stage from the cormorant story to the remora

story, as fathered by Columbus and permanently incor-

porated in all later accounts who drew upon the Colum-

bus story.

17 T. Domenichelli, ‘‘Sopra la vita e i viaggi del beato Odorico Da Por-

denone,’’ in Prato, 1881, p. 180 (cap. XLVI).

18 Ibid., p. 232,

19 Yule, op, cit., p. 352.

20 Ibid., p. 189.

172 THE AMERICAN NATURALIST [Vou. LV

Professor Gudger has shown, beyond any possibility

of cavil, that all the accounts of the remora fishing in

America recorded after Oviedo go back to this latter

source, and I shall now show that Oviedo’s account goes

back, through Bernaldez, to an Arabic source, which is

itself an evolution of the second Italian version of

Odoric’s cormorant fishing, as preserved to us in Ra-

musio.

Bernaldez* says: ‘‘For they call it hunting, and they

hunt one fish with others of a particular kind,’’ while in

the ‘‘ Journal of the Second Voyage’??? we read: ‘‘The

fishing consists in this that they take certain fishes which

they call revesos, the largest of whom are not larger

than pilchards,’’ from which Peter Martyr made his

‘*reversus fishes.’ In the Spanish the passage in Ber-

naldez runs as follows: :

Vino una canoa a casa de pezes que ansi le llamaban ellos caza, que

cazan con unos pezes otros.

It will be observed that all the Columbus accounts tell

of the invitation extended by the fishermen to Columbus

to see the peculiar kind of fishing, and the giving of the

catch to Columbus, according to Bernaldez, for a feast.

This is identical with the manner in which Odoric tells

of the invitation to watch the cormorant fishing. The

resemblance is striking. Now, in the second Italian ver-

sion in Ramusio the fish with which other fish are caught

is called marigione, ‘‘diver,’’ while others call it sea-calf.

We have here, side by side, cormorant, otter and remora.

I have already shown in my book, ‘‘ Africa and the Dis-

covery of America,” that much of the matter in the

‘Voyages of Columbus’? is apocryphal and comes from

Odoric of Pordenone’s ‘‘Itinerario,’’ a name which Ber-

naldez uses for the book of Columbus, from which he got

his information. There can be little doubt that the sec-

21 Loe. cit., p. 450.

22 See my ‘‘Africa and the Discovery of America,’’ Philadelphia, 1920,

Vol, I,

ceil anaes 3 loc. cit., p. 297,

No. 637] ONCE MORE THE SUCKING-FISH 173

ond Italian version was corrected or annotated by Co-

lumbus in the margin, where the true story of the remora

fishing at Zanzibar was given from an Arabic source,

from which Columbus retained two foreign terms. He

had found in his source kassa, the turtle caught by the

remora, and the name was apparently entered into the

margin from which Bernaldez got his threefold caza

*‘chase.’’ Indeed, it appears that in his ‘‘que ansi le

llamaban ellos caza,’’ it referred originally to the fishes

caught, that is, to the turtles, which from the resem-

blance to Spanish caza, ‘‘chase,’’ produced the unfortu-

nate pun. It will be noticed that in the ‘‘Journal of the

Second Voyage’’ the corresponding passage runs ‘‘they

take certain fishes which they call reversos,’’ where the

second Italian version says ‘‘fishing by aid of a certain

fish called marigione,’’ that is, ‘‘diver.’? Now the Arabic

word for ‘‘diver’’ is gavvdsah. Anciently the initial

guttural was rendered in Spanish by a simple g, but in

the fifteenth century this Arabic word would sound to a

European ear as reverso or reveso, which it actually as-

sumed in the Columbus story. No such Spanish word

is anywhere else recorded for the remora. Again, the

marginal gloss, from Bernaldez, ‘‘hunting with a fish,’’

must have been ‘‘caza con un pez,’’ which Peter Martyr

took to be the name of the fish, the remora, hence he mis-

read the first as guaicanum, and called this the Indian

name for the fish, a word which is only recorded as a

quotation from Peter Martyr.

From the previous discussion it follows:

1. The remora fishing is very old and originated in the

Indian Ocean, but did not get into literature before Co-

lumbus.

2. Odorie of Pordenone’s cormorant fishing was from

the start confused with the fishing by means of an otter

and, in Ramusio’s second version, was dimly identified

with the remora fishing.

3. Ramusio’s second version was, before the time of

Columbus, influenced by an Arabic source or explained

174 THE AMERICAN NATURALIST [Vou. LV

by an Arab acquainted with the remora fishing at Zan-

zibar, and this new form supplied Columbus with the

Zanzibar word for ‘‘turtle,’’ namely, kassa, and the

. Arabic word for ‘‘diver,’’ namely, the Spanish reves or

reverso, which was wrongly attached to the ‘‘remora.’’

4, Bernaldez and Peter Martyr created a non-existing

remora story for America out of Odoric’s much-revised

cormorant story, on the basis of some marginal notes in

Columbus’s ‘‘Itinerario,’’? which itself was based on

Odoric’s ‘‘Itinerario,’’ and referred to Zanzibar and not

to America.

5. There are in America no corroborative stories of

the remora fishing, except as derived from Oviedo’s

hearsay account, which itself is based on the accounts of

Bernaldez and Peter Martyr, which, in their turn, are

taken from a revised edition of Odoric’s cormorant

fishing story.

SHORTER ARTICLES AND DISCUSSION

REPORT OF THE COMMITTEE ON GENETIC FORM AND `

NOMENCLATURE

THE American Society of Naturalists at their meeting in 1919

appointed a Committee on Genetic Form and Nomenclature con-

sisting of Drs. S. Wright, G. H. Shull, O. E. White, A. H.

Sturtevant and myself as chairman. We were to consider the

matter of genetic nomenclature and submit constructive sug-

gestions for standardizing descriptive terms in this subject.

‘The following report of the committee was submitted to the

meeting of the American Naturalists at Chicago, 1920, as a foun-

dation intended to cover the cases of inheritance commonly met

with by the majority of experimental workers in genetics. It is

submitted in the hope that it may be published to invite dis-

cussion as to suggested modifications which would enable it to

include particular problems of the scores of investigators in this

field. In making such criticisms it is suggested that the primary

object of this report be continually borne in mind and that con-

structive suggestions based on it as a framework are more likely

to lead to beneficial results than purely destructive ones. The

` vast majority of workers in genetics will be concerned with

simple enough problems to be covered by the report. Those

whose material requires modification of the methods therein sug-

gested will undoubtedly see the justice in attempting to adapt

their particular needs to some modification of a system which

will meet the needs of the majority.” ;

C. C. LiTTLE, Chairman,

Committee on Genetic Form and Nomenclature.

In submitting this report your committee desires to call atten-

tion to certain matters of general interest in connection with it.

It is neither proposed nor supposed that those now familiar with

some characteristic or individual form of genetic nomenclature

will necessarily find it desirable to conform with the suggestions

contained herein. If they can and will cheerfully do so, so

much the better; if not, no intention to dictate is implied in this -

report.

175

176 THE AMERICAN NATURALIST [Vou. LV

It is, however, believed that a considerable number of geneti-

cists will agree to the main suggestions of the report, and will

thereby form a nucleus to which younger geneticists beginning

publication would in a majority of cases join themselves. Thus,

after a time, a far more uniform method of publication than now

exists would become established.

In order to give such an opportunity, your committee respect-

fully suggests that this report, if approved by vote of the mem-

bers present, be published at the earliest’ convenient time.

1. The Type—tIn most animals and plants it is convenient to

settle on a standard type, preferably the wild type, when this is

known. The effects of the various genetic factors are in general

to be measured by the departure from type which they bring

about. This recommendation involves no real departure from

the system now in use by most geneticists.

2. Series of Allelomorphs.—A single letter, with a subscript,

if necessary, is to be assigned to each series of allelomorphs.

This letter should, when possible, be chosen so as to give some

hint as to the nature of the effects caused by variations in the

series in question. The member of each allelomorph series pres-

ent in the type is to be represented by the symbol for that series,

capitalized and with no superscript. Factors dominant over the

type are to be represented by the same capitalized symbol as the

type, but with appropriate superscripts. Recessives are to be

represented by the same symbol in lower case also with appro-

priate superscripts (when necessary). The symbols for the type

factors may be omitted in formule where convenient. The

agouti series in mice A’, At, A, a”, in which two factors are domi-

nant over the wild gray type and one recessive is an example of

the use of symbols. [This series might properly have been given

a Y or B symbol in place of the A adopted. Since, however, it

is to be thought of in terms of modification of the agouti pattern,

thë symbol A is chosen. ]

3. Dominance —Dominance of genes is recognized to be largely

a matter of convenience. Factors may be considered dominant

which produce an easily recognized departure from type, when.

heterozygous.

4. Superscripts—It is suggested that both a literal and a

numerical superscript be assigned, upon the initial description,

to each factor except the type (at least in series of multiple

allelomorphs). EITHER SUPERSCRIPT MAY THEREAFTER BE USED

ALONE, The numerical superscript shall indicate the estimated

No. 637] SHORTER ARTICLES AND DISCUSSION 177

degree of divergence from type, produced by the factor in ques-

tion in a seale in which 10 is the apparent physiological or visible

limit and 0 is the type. Thus A! and a” represent factors

which cause deviations to the physiological limit in opposite

directions (self yellow and self black) from the type A (agouti).

A+ (light-bellied agouti) represents an estimated deviation be-

tween ticked bellied agouti (A) and yellow (A1). The order

„of effect is more important than a precise estimate of the degree

of effect. Decimals and numbers beyond 10 may be used when-

ever necessary, in event of grades not believed physiologically

possible. A superscript, once adopted, should not be changed,

which also applies to all other symbols. The value of making

provision for a system to indicate the order of a multiple allelo-

morph series is clear; the numerical symbols will only be used

when such a situation is encountered.

5. Independent Factors——Independent series of allelomorphs

should be represented by different letters or, if desired, by the

same letter with different following numbers. Symbols com-

posed of two or more letters should not be used. It is suggested

that factors with more or less similar effects be represented by

the same letter with different following numbers, as S1, 82, S3,

ete. The same symbol may conveniently be used for factors

with more or less similar effects in different animals and plants

without implying identity.

6. Doubtful Factors—In case the formula of an individual is

not fully known, the uncertain factor may, if desired, be repre-

sented by a superscript X (or ?) or the whole symbol may be

replaced by a dash. Thus C* C* (or C? C?) means complete

ignorance of the factor in series ©. CC*, CC*, or C — represents

ignorance as to one of the factors in the zygote. If there is

partial knowledge, a double (or triple) superscript may be used

to indicate the various possibilities. Thus the progeny of the

cross CC X ec? may be represented by C cè, a form which gives

more information than C —.

7. Modifiers—The symbol [ ] containing appropriate symbols

represents residual heredity of the kind indicated. Thus [S +]

is a convenient method of representative + modifiers of the effect

produced by the S (Spotting) series of allelomorphs. In a đe-

tailed study of a particular group of modifiers, the parenthesis

may well contain the grade of effect produced by the modifiers

in the case in question. Thus [+ 4.2] and [—2.5] might be

used to represent the modifiers of typical hooded rats of grades

178 THE AMERICAN NATURALIST [ Von. LV

+4.2 and — 2.5. The modifiers of a cross bred may be indi-

cated in some appropriate manner as [+ 4.2, —2.5].

8. Linkage is best represented by the fractional form used by

workers on Drosophila. The factors are written in the order of

linkage, omitting type factors.

OMMITTEE ON GENETIC FORM AND NOMENCLATURE

STANDARDIZED MICROPHOTOGRAPHY

SECOND CONTRIBUTION: THE OBJECT Factor

In my first contribution to the subject of standardized micro-

photography, published in the Anatomical Record, I have pointed

out the variables and the methods which I have pursued in treat-

ing them. One, or perhaps more correctly a group of variables

were, however, left out of consideration quite purposely because

of the difficulty in finding for them a standard of permanent

value. I have in mind the microscopical section itself or what

may be properly called the object factor. The following four

elements enter into its composition: (1) the thickness of the

section, (2) the light absorption coefficient of the tissue, (3) the

relative luminosity of the different stains and (4) the depth or

intensity of staining. The second and third of these component

elements may be disregarded since experience ‘shows us that ex-

posure is very little influenced by them. There remain, how-

ever, the first and fourth, and to determine the influence of these

on exposure the following experiments were undertaken.

First of all, to avoid all possible error, slides were chosen of

uniform thickness measured with a Ciceri Smiths Patent Mi-

erometer so as to be sure that the distance of the section from

the substage condenser should in every case be the same. The

cover-glasses were also of uniform thickness. The stomach of a

frog preserved in Zenker’s fixing fluid was sectioned into series

of 5, 10 and 20 micromillimeter thick sections on a Minot micro-

tome and care was taken to have in each case a ribbon of 100

even sections, thus more or less assuring their uniform thickness.

The sections were stretched on distilled water heated over a

flame and no cement of any kind was used. On one slide three

sections of each thickness were placed. On other slides sections

only of one kind were placed and their thickness marked in

every case by a carborundum pencil.

No. 637] SHORTER ARTICLES AND DISCUSSION 179

The slide with all three kinds of sections was stained for 12

hours in alumearmine, a stain which, as is well known, does not

overstain. They were then washed in water and again stained

for 12 hours in a weak alcoholic solution of Bleu de Lyon. All

sections on this slide received therefore the same treatment and

the difference in the depth of stain was entirely due to the thick-

ness of the section stained.

The other slides were treated in a different way. They were

stained in Delafield’s hematoxylin followed by tetrabromfiuo-

rescic acid. This stain was chosen because it is possible at will

to control the depth of staining. A set A of three slides, one

with 5 micromillimeter sections, one with 10 and one with 20,

was treated simultaneously in a Coplin’s staining jar. The sec-

tions were therefore stained, washed, destained in acid alcohol,

treated with ammonia alcohol, stained in a weak solution of

tetrabromfluorescic acid in 95 per cent. alcohol and washed in

pure alcohol the same length of time. In regard to depth of

stain these slides presented, therefore, the same conditions as the

slide stained in alumearmine and Bleu de Lyon.

Several other slides were treated individually in the same

stains. They were all first considerably overstained in Dela-

field’s hematoxylin, washed in water and destained in acid

alcohol until they had when viewed over a white surface, the

same shade of color to the naked eye, regardless of the thickness

of the section. If after treatment with ammonia alcohol the blue

color was not approximately of the same shade, the darker slide

was again transferred to the acid alcohol, until all sections

looked approximately alike. They were now stained in the

tetrabromfluorescie acid, the thickest sections remaining a short

time in the fluid, the 10 micromillimeter sections somewhat

longer and the 5 micromillimeter sections longest, and again

compared over a white surface. From several slides thus pre-

pared three were selected which to the naked eye were to all

purposes of the same appearange, although their respective

thickness were 5, 10 and 20 micromillimeters. Here then we

had a Set B of sections in which the depth of stain had nothing

to do with the thickness of the section, but was entirely depend-

ent upon the amount of stain absorbed by the tissues.

Fractional exposures on orthonon plates with Cramer ray-

filters were made on our standardized microphotographie appa-

ratus and developed by the time and temperature method with

180 THE AMERICAN NATURALIST [Vou. LV

fresh developer for each plate. In the ease of the first slide

stained in alumearmine and Bleu de Lyon as in the Set A, the

normal exposure for the 10 micromillimeter section was twice

that for the 5 one, and the normal exposure for the 20 micromilli-

meter section was four times that of the 5 one. In the case of

the Set B all slides required the same exposure.

In analyzing the results thus obtained we come to the con-

clusion that the thickness of the section, within the limits given,

has no influence whatsoever on the length of exposure; but that

the latter stands in a direct ratio to the amount of stain absorbed

by the tissue.

For practical purposes, especially for those who are using my

table of R-P factors, the results of this investigation may be

interpreted as follows: disregard the thickness of the sections

and the appearance of the stain under microscope and pay at-

tention only to the intensity as it appears to the naked eye. The

table refers to normally well stained sections of medium thick-

ness. Double the exposure for darker appearing and reduce the

exposure by half for lighter appearing sections.

It will be observed that in the experiments elements 2 and 3

remained of constant value for the simple reason that the tissue

used ‘was not only the same in kind, but actually from the same

piece of organ and was stained in the same stain in each set of

slides. A comparison of the length of normal exposure in the

case of sections stained in alumearmine and Bleu de Lyon with

those stained in hematoxylin and tetrabromfluorescic acid serves

to confirm my statement at the beginning of this article, that the

relative luminosity of different stains may be entirely disre-

garded in the matter of exposure.

ALEXANDER PETRUNKEVITCH

YALE UNIVERSITY.

SEX RATIOS IN PLATYGASTER

In the Journal of Heredity for last year (Vol. X, p. 344) I

published data on sex ratios in three species of polyembryonic

Hymenoptera. One of these is Platygaster felti, which para-

sitizes the eggs of two species of the Cecidomyiide, Walshomyia

tarana and Rhopalomyia sabine. These flies make their galls

on the mountain cedar, Sabina sabinodes.

Since the above data were published I have had an oppor-

tunity to breed out in the laboratory a total of 200 broods of

No. 637] SHOFTER ARTICLES AND DISCUSSION 181

Platygaster ; 73 from the carcasses of Rhopalomyia and 128 from

those of Walshomyia. The remarkable character of the sex

ratios, as revealed in the published data, is emphasized by the

additional facts secured from the more recent rearings. It

therefore seems worth while to publish the full data, which is

given in condensed form in the following table. In the first

column the total number of individuals in each brood is given;

in the second, the number of females; and in the third, the

number of males. In the fourth column are listed the number

of broods showing the combination of females and males in the

corresponding horizontal line.

The total number of individuals in the 200 broods is 2,722, of

which 2,346 are females and 376 males. The average per brood

is 13.61. The size of the brood reared from Walshomyia is very

much smaller than those from Rhopalomyia. There are 1,417

individuals in the 128 broods from Walshomyia, or an average

of 11.07 per brood. The broods from Rhopalomyia have 1,305

individuals, or an average of 18.12 per brood. This represents

an average increase of 63.6 per cent. The rate of increase in

the number of males per brood from Walshomyia to Rhopalo-

myta is almost the same as this. The average number of males

in broods from Walshomyia is 1.55, and in those from Rhopalo-

myia is 2.45. This represents an increase of 58 per cent.

The difference in the size of broods is, in all probability, due

to the difference in the size of the two host larve. The larva of >

Rhopalomyia is almost twice as large as that of Walshomyia,

and hence must furnish a more abundant food supply for the

multiplication of embryos at the time of their formation in the

polygerm.

One of the most striking facts in the data is the preponder-

ances of females. Approximately 86 per cent. of the individuals

are females. No male brood has been found, and in not a single

instance does the number of males exceed or even equal the

number of females in a brood. There are, however, nine pure

female broods. Of the 191 mixed broods 113, or 59.61 per cent.,

have a single male present in each brood. The other 78 broods

show the following numbers of males: Thirty with two males

each; twenty-four with three each; ten with four each; four

with ‘ave each; four with six each; ‘thie with seven cith; two

with eight each; and one with ten.

Another interesting fact is the frequent occurrence of broods

with a single male. Almost sixty per cent. of the mixed broods

[ Vou. LV

THE AMERICAN NATURALIST

182

TABLE I

BROODS oF PLATYGASTER FELTI

f3]

2 .

mM MARENN N A r AA N A AN i O 00 a rA A OU A tt O o m N O OD tt O OI O ei CI H r OO e ttt tt OR tt GA e r

- .

t=]

Š NANNAN HANAHAN ONNANMHOMNA HANHAN NOHH NNOMN AHOMAA NO HNA m HA m HAN O HA HNNAN m EO H HO

A MHTMOWOOCOKhHHDHALMDRHOCDHOCCHHBOCHAMOCAHANAAMAHAMMHMOOMHINOHNOKRMHOKOQOKKDKROHOGOOK

5 rH y r m rA ri r r i rd ri o o or r or o o o e o o o o o e p p et et St oe d edt NN ANN m

ics .

kz]

Zl SSOP DDBADBOCSCOOCHAAAAHMHMDHOOMHHHOHMOMOOOORKREERDDNDABARSOOH AANA 9 W196

s PA ore or o o e p e e ed e e e e e e ee ee o e e e d r NANANNANNAAAN

No. 637] SHORTER ARTICLES AND DISCUSSION 183

TABLE I (Continued)

BROODS OF PLATYGASTER FELTI

No. of Individuals Females Males No. of Broods

26 20 6 1

26 21 5 1

26 22 4 1

26 23 3: 1

27 21 6 1

28 25 3 1

29 24 5 1

31 24 7 2

34 26 8 1

34 27 t 1

36 26 10 1

36 30 6 1

37 31 6 1

are of this character. Certain combinations, such as nine

females and one male, occur with very great frequency, sug-

gesting that a single male is produced at some definite point in

development. Statistical treatment of the data supports this

suggestion. My colleague, Dr. Muller, has kindly calculated

the standard error and standard deviation of the numbers of

males per brood, and finds that they are 1.17 and .64 respec-

tively. These results show that the number of males produced

per brood does not vary as much as it would if males were

formed at random. If males were produced at random, the

amount of variation in the number of males would be expressed

by a standard deviation of 1.17. This means that there is a

tendency to have the production of males confined to particular

cells in the embryonic mass, so that only one or two males are

usually formed in a brood.

The fact that the parasite deposits one egg at each oviposition

makes it practically certain that the mixed brood of Platygaster

is the product of a single fertilized egg. The important ques-

tion is how the single male originates during the course of de-

velopment. I have elsewhere discussed this question, and have

suggested that the appearance of one or more males in a brood

may be due to an abnormal behavior of the sex chromosomes.

An abnormal division causing the loss of an x-chromosome from

one of the early blastomeres would explain the appearance of a

mixed brood, for such a cell could become the progenitor of one

or more males.

J. T. PATTERSON

UNIVERSITY OF TEXAS

184 THE AMERICAN NATURALIST [Von. LV

REPRODUCING POWER OF WELL-FILLED VS. POORLY-

FILLED EARS OF MAIZE .

THESE tests were conducted to determine the effects, if any, on

the progeny, particularly as to productivity, when a stalk of corn

is caused to produce comparatively few kernels instead of a nor-

mal-sized, well-filled ear. In other words, the object was to learn

whether artificially reducing the possible number of progeny

kernels would have any influence on their viability, vigor or

ability to yield.

In selecting seed corn, ears are occasionally found which evi-

dently would have been much larger and better filled had not

something such as an overhanging blade or an insect interfered

with pollination. Are such ears suitable for seed?

Similar tests were conducted with three varieties, one being a

cross-bred variety. The three classes of seed for these tests were

grown in 1914 and their comparative productiveness tested in

1915. The seed of U. S. Selection 77 was grown and: tested at

Piketon on river-bottom soil in Southern Ohio, and that of the

other two varieties at Broad Run on Piedmont clay of Northern

Virginia,

METHODS OF PROCEDURE

Two methods were used to control the pollination and conse-

quent seed production of the poorly filled ears. In one case, the

first silks to appear were about an inch beyond the end of the

shoots when the shoots were bagged to prevent further pollina-

tion. In the other case, the ear shoots were bagged before the

silks began to appear. When all the silks had protruded several

inches the bags were removed for half an hour and then re-

placed. This was done when pollen was falling freely. A few

of the uncovered silks thus became naturally pollinated.

The first method produced ears the butt ends of which were

fairly well filled for about one fourth the length of the cob. The

second method gave ears that had a few large rounded kernels

seattered over the cob. As check seed for the tests, large, well-

filled, typical, seed ears that had been allowed to mature un-

molested were selected. The seed ears of the three lots of each

variety were selected from the same rows from similar stalks

grown under like conditions as far as possible. The drying, care,

ete., were the same for each of the three lots.

185

S AND DISCUSSION

R ARTICLE

SHORTE

oe,

rp edd hy

ek?

pet

bog

ANAK

Ñi

}

arative productivity tests were conducted with lots of seed taken

1. Comp

Fic.

from the three kinds of ears here represented.

186 THE AMERICAN NATURALIST [ Vou. LV

The seed from the check ears will be referred to as such while

that from the ears filled only at the butt ends, as Lot 2, and that

from the ears with kernels scattered over the cob as Lot 3.

The number of ears used in making up each lot of seed fol-

lows : Selection 77, Check, 62 ears; Lot 2, 67 ears; Lot 3, 54 ears;

Selection 119, Check, 26 ears; Lot 2, 19 ears; Lot 3, 26 ears; and

` Cross 182, Check, 26 ears; Lot 2, 14 ears; and Lot 3, 19 ears.

In preparing for planting each lot was composited in the fol-

lowing manner. The same number of kernels was taken from each

ear of a lot and these kernels combined made just enough to

plant one row 50 hills long. The comparative weights of the

three lots of seed are given in Table I.

TABLE I

WEIGHTS IN GRAMS OF 265 KERNELS OF EacuH LOT oF SEED

Variety | Check Seed | ets | Lot 3

SOSOMON: or sok. poke es 124 141 137

Selection 210.3555 kras 118 122 127. |

Re PREE NE PE A S A T 126 148 151

DATE AND METHOD OF PLANTING

U. S. Selection 77 was planted May 1, 1915. The three lots

were planted in adjacent rows and the test repeated 14 times.

Selection 119 and Cross 152 were planted May 3, 1915. The lots.

of these two varieties were planted in the same order as the lots

of Selection 77, but only 15 rows each of these two varieties were

planted. The planting was done by hand and later all the rows

were thinned to a uniform stand.

OBSERVATIONS DURING GROWTH

At no stage during the growth of the corn was there any

noticeable difference among the three lots. Neither was there

any difference in the time of silking and tasseling, in the height

of stalk, nor in time of maturing. There was a slight difference

` in the field germination of the three lots of seed as shown in

Table II., but these differences within the varieties are not great

enough is be significant nor are they consistent for the three

varieties,

No. 637] SHORTER ARTICLES AND DISCUSSION 187

TABLE II

Ave. Weight | | Corrected

Per Cent. | “pounds | Pounds | pushes

Selection 77..... Check | 87.6 0.763 | 1,725 | 1,2920 | 96.7

2 86.6 | 0.736 | 1,738 | 1,228.0 | 89.9

3 86.6 0.761 1,723 | 1,277.0 | 94.5

Selection 119....| Check 84.5 0.592 481 308.1 | 74.8

2 84.5 0.586 472 283.1 | 70.9

3 79.7 0.585 494 290. 6 69.7

Cross 182;..:... Check 83.8 0.682 503 80.1

2 88.3 0.635 507 I $ 74.8

3 80.8 0.662 510 | 3314 | 77.0

RESULTS

The yields in Table II. are all based on field weights at harvest

time. With all three varieties the well-filled seed ears produced

the highest yields, the increase being from 2.2 to 3.9 bushels per

acre over the next highest lot. In the 25 comparisons between

the two lots, Checks outyielded Lot 3, its nearest competitor,

16 times, with one tie; and it outyielded Lots 2, 20 times with

two ties.

Previous work had proved Cross 182 more productive than

Selection 119. In the tests reported in Table II. they occupied

- the same amount and kind of soil and Cross 182 is consistently

more productive than Selection 119.

Regarding all three varieties the ears harvested from each of

the three lots of seed were equally well-filled and of the same

general appearance. These tests warrant the conclusions that

ears poorly-filled by reason of withheld pollen will not transmit

this character to their progeny, and can be expected to supply

seed almost as productive, if not as productive, as they would

have supplied if completely pollinated.

C. P. HARTLEY,

H. S. GARRISON

U. S. DEPARTMENT ‘OF AGRICULTURE

ANTHEROPHAGUS OCHRACEUS MELS. IN THE NESTS

OF BUMBLEBEES

THE recent appearance of two articles on ‘‘The Phoresy of

Antherophagus,” one by W. M. Wheeler (1919) and the other

by H. Donisthorpe (1920) have prompted me to publish some

additional observations on the habits of Antherophagus ochraceus

Mels. in this country.

Wheeler writing in December, 1919, recorded the capture of

an adult of this beetle near Colebrook, Connecticut, attached to

the proboscis of a worker bumblebee (B. vagans). When the

bumblebee, which vainly tried to rid herself of the beetle, was

placed in a cyanide jar, the beetle still maintained its hold. In

this same article there is a discussion of the phenomenon of

‘‘phoresy’’ as defined by Lesne (1896) and expanded by Janet

(1897), together with an account of the known habits of An-

therophagus and a bibliography. Donisthorpe in October, 1920,

published a résumé of Wheeler’s paper, presenting further in-

formation concerning ‘‘phoresy’’ in general, the habits of An-

therophagus, and additional references. Scott (1920) has also

contributed to our knowledge of the biology of these beetles.

Many of the rather numerous European references report the

finding of Antherophagus (pallens, silaceus and nigricornis) and

of Cryptophagus (setulosus, and sp.) in the nests of various

species of bumblebees. The only American record actually citing

an instance of finding Antherophagus ochraceus Mels. in the nests

of bumblebees is that given by A. S. Packard (1864) based on

observations made by F. W. Putnam in Massachusetts and Ver-

mont. J. B. Smith (1909), without giving any data, says that

Antherophagus occurs in the nests of bumblebees. This latter

note is probably based on the statements by Packard, or else on

accounts of the habits of the European members of this genus.

While in Wisconsin last summer (1920), I was able to examine

many bumblebees nests in various parts of that state, through the

kindness of Dr. S. B. Fracker. On two different occasions I

found Antherophagus ochraceus (C. A. Frost det.) in the nests.

In one nest of Bremus (Bombus) fervidus (Fabr.) examined

August 12, 1920, were eighteen adult specimens of this beetle.

188

No. 637] ANTHEROPHAGUS OCHRACEUS MELS. 189

At the same time and place I took thirty-four larve of a small

beetle in various stages of development. As these larve were

associated with the beetles and as they agree with the figure and

brief description of A. ochraceus given by Packard (1883), I

assume them to be the same. In another nest of Bremus (Bom-

bias) auricomus (Robt.) opened on July 26, 1920, at Clyman

Junction, Wisconsin, I found a single adult of A. ochraceus.

Again on October 3, 1920, near White Heath, Illinois, I collected

about a dozen small beetle larve in a surface nest of B. Pennsyl-

vanicus (DeGeer). These last-mentioned larve differ slightly

from those found in the nest at Baraboo, Wisconsin, and if not

the same species may represent another species of Antherophagus.

There has been much discussion as to the feeding habits of the

adult and larval Antherophagus. Wheeler, after a survey of the

literature of the subject, came to the conclusion that the larve

were ‘‘in all probability merely scavengers in the Bombus nests.’’

Wagner (1907) expresses the idea that they ‘‘will occasion enor-

mous destruction in the nest,” but without giving an instance

of the same. I believe that these insects are purely scavengers,

not only feeding on the excrement of the bumblebees as suggested

by some, but also on all kinds of refuse as maintained by Reuter

(1913). In the nests I examined containing Antherophagus

ochraceus, the beetles and larve were never on that portion of

the comb then being used by the bees. They were always either

on the old decaying empty cocoons on the bottom of the nest,

or in the débris directly beneath or surrounding the comb. Such

are not the habits of the true parasites and harmful inquilines

of bumblebees. The larve of Vitula (Nephopteryx in litt.)

edmansii described by Packard from the nests of bumblebees feed

on the pollen, honey, wax or cells of the comb. To escape being

killed by the bumblebees or carried out of the nest, the larve

of this moth spin a regular labyrinth of silken tubes or cases

and never expose themselves to the bumblebees. The larve of

Antherophagus do not spin protective cases and are in no sense

of the word parasitic on the adult bees, larve or pupe. If they,

thus unprotected, should crawl conspicuously over the comb to de-

stroy the eggs, larvæ or pup, or to eat the new comb and stored

food, they could easily be combated by the bumblebees. Further-

more, the nest containing the thirty-four larve and eighteen

adult beetles taken at Baraboo, Wisconsin, showed no signs of

the great destruction mentioned by Wagner. For that time of

190 THE AMERICAN NATURALIST [Von. LV

year, August 12, it was in fact a strong colony, containing ninety-

one workers, fifty-six pupal cocoons, and large stores of honey

and pollen. It is possible that the upper part of the comb of a

bumblebee nest might develop so swiftly in some cases, as to

cause some cells either filled or not filled with pollen and honey

on the lower part of the comb to be neglected, and thus infested

with Antherophagus. This last statement, however, would cer-

tainly be the exception rather than the rule. In the cases that

have come under my observation A. ochraceus played the rôle

of a scavenger, in the débris beneath and about the nest, feeding

on the refuse comb, feces, honey, or bits of pollen and wax that

perchance had fallen to the bottom of the nest.

Wheeler voices the opinion of Sharp (1899) that the instincts

of the beetle permit it to recognize the bumblebee, but not to en-

able it to find the nest. Therefore the beetle waits on flowers

until it can attach itself to a bumblebee and be conveyed to the

nest of the latter. Donisthorpe suggests that ‘‘it is not so much

that they [Antherephagus] lack the instinct to find the bee’s

nest, but rather that it gives them protection from their hosts

when they arrive there.” . Antherophagus may or may not be

able to find the nests of bumblebees of its own accord, but I am

inclined to doubt whether the occasion of the ‘‘phoresy’’ is pro-

tective, in that it gives ‘‘them protection from their hosts when

they arrive there,’’ by their having acquired the nest ‘‘aura.’’

lf Antherophagus is a scavenger, as the evidence seems to indi-

cate, and keeps well hidden in the débris on the bottom or sides

of the comb, why is there a need for a nest ‘‘aura’’? One of

these beetles carried to a bumblebee’s nest, in all probability,

soon after arriving there, releases its hold and falls down to the

lower part or bottom of the nest. Many other beetles are acci-

dental visitors or inhabitants of such nests, and living thus in the

material about and beneath the comb are not noticed by the ever-

watchful bees and go unmolested. W. H. Tuck (1896, 1897) lists

over sixty species of beetles from the nests of various species

of bumblebees in England, most of which are undoubtedly only

casual intruders. I have taken specimens of Harpalus sp. and

Onthophagus hecate Panz. in the nests of bumblebees. Such

les are much larger than Antherophagus, are not even con-

sidered as ‘‘anthophilous’’ (Lovell, 1915), nor have they ever

been accredited with habits of ‘‘phoresy.’’ Evidently then, such

beetles gain entrance to the nest and live there for a time at

No. 637] ANTHEROPHAGUS OCHRACEUS MELS. 191

least without having first acquired a nest ‘‘aura.’’ I believe that

Antherophagus often, if not always, forces the bee to carry it

simply in order to find the nest, and not to acquire a nest ‘‘aura,”’

such as all the bees of each and every colony possess. If

Antherophagus had habits similar to those of the inquiline-bee

Psithyrus, there would be an advantage in having a nest ‘‘aura.”’

Scott (1920) says that

Presumably these [A.: pallens] beetles are double-brooded, with a

short summer generation intervening between the emergence of the

adults in May and the assumption of the resting condition by the larve

in autumn.

I have taken adults of A. ochraceus by sweeping flowers on

May 7 and 23, 1917. The insect collections of the Illinois State

Natural History Survey contain adults taken on July 19, 23

and 30, 1891, and August 15, 1893. Blatchley (1910) mentions

the species as occurring on flowers, June 24 to September 21.

As previously mentioned I took one adult in a bumblebee nest on

July 26 and eighteen more on August 12. The adults taken on

May 7 and 23 certainly represent the hibernating brood. Those

found both out of doors and also in a nest on July 19-30, are in

all probability the adults of the first brood or summer genera-

tion. Those taken on August 12-15, may represent a true second

summer generation, but more than likely belong to the same

brood of July 19-30. Scott found that A. pallens hibernated as

larvæ, pupating in early summer. Some of the larvæ presumably

those of Antherophagus which I took on October 3, when ex-

amined on November 11, 1920, had constructed cells in the earth

on the bottom of the rearing jar; thus indicating that they

hibernated as larve. Summarizing these records: A. ochraceus

is probably double-brooded, hibernating as larve in cells in the

soil or material about or under the bumblebee nest.

THEODORE H. Frison

UNIVERSITY OF ILLINOIS

BIBLIOGRAPHY

Biátohloy, w. 8.

1910. An Illustrated Descriptive Catalogue of the Coleoptera or

eetles (Exclusive of the Rhynchophora) Known to Occur in

Indiana. Nature Publ, Co., Indianapolis, Indiana.

Donisthorpe, H.

1920. The Phoresy of Antherophagus. Ent. Record, Vol, 32, No. 10,

Pits nee»

192 THE AMERICAN NATURALIST [Vou. LV

Janet, C. 3

1897, Etudes sur les Fourmis, les Guépes, et les Abeilles. Note 14.

Rapports des Animaux Myrmécophiles avec les Fourmis. Li-

moges, pp. 1-9.

Lesne, P.

1896. Moeurs du Limosina sacra. Phénomènes de transport popes chez

les animaux articulés. Origine du parasitisme chez les

Diptères. Bull. Soc, Ent. France, T, 45, pp. 162-165.

Lovell, J. H.

1915. A Preliminary List of the Anthophilous Coleoptera of New Eng-

land. Psyche, Vol. 22, No. 4, pp. 109-117 i

Packard, A. 8.

1864. The Humble-bees of New England and their Parasites; with

Notices of a New Species of Anthophorabia, and a New Genus

of Proctotrupide. Proc. Essex Inst., Vol, 4, pp. 107-140, pl. 3.

1883. Guide to the Study of Insects. Eighth Ed. New York, Henry

Holt and Co., pp. 446—447.

Reuter, O. M.

1913. Lebensgewohnheiten und Instinkte der Insekten bis zum Er-

wachen der sozialen Instinkte. Berlin, Friedländer und Sohn.

Seott, H.

1920. Notes on the Biology of Some Inquilines and Parasites in a Nest

of Bombus derhamellus Kirby; with a Description of the Larva

and Pupa of Epuræa depressa Illig. (== aestiva Auctt.: Cole-

optera, Nitidulidæ). Trans. Ent. Soc. London, pp. 99-127,

Figs. 1-8. ;

D. `

1899. Insects. Cambridge Natural History, Vol. 6, p. 235.

Smith, J. B.

1909. Insects of New Jersey.

W. B

1896. Inquiline and other Inhabitants in Nests of Aculeate Hymenop-

tera. Ent. Mon. Mag., Series 2, Vol. 7, pp. 153-155.

1897. Coleoptera, ete., in the Nests of Aculeate Hymenoptera. Ent.

=a Mag., Series 2, Vol. 8, pp. 58-60.

Wagner, W.

1907. EERE Untersuchungen an Hummeln mit Bezugnahme

auf der Frage der Geselligkeit in Tierreiche. Zoologica, Bd.

z p. 145.

Wheeler, W.

1919. me R of Antherophagus. Psyche, Vol. 26, No. 6, pp.

45-152.

THE

AMERICAN NATURALIST

Von. LV. May-June, 1921 No. 638

THE INTERNAL SECRETIONS IN GROWTH AND

DEVELOPMENT OF AMPHIBIANS

EDUARD UHLENHUTH, Px.D.

(From the Laboratories of the Rockefeller Institute for Medical

; Research, N. Y.)

Ware up to 1910 the higher vertebrates were used pre-

dominantly in the study of the internal secretions, during

the last decade the larvæ of the amphibians have been

found an excellent material, suitable for the investigation

of many problems of endocrinology. To-day the results

obtained in this work seem to form a solid mass of trust-

worthy evidence, from which may be derived not only

valuable information as to the mechanism of growth and

development in amphibians, but also important knowl-

edge as to the functions of certain endocrine glands.

In the field of internal secretion, these experiments have

attracted increasing interest from the begięning. It is

evident at present that further clarification of many of

the more important problems of internal secretion will

come from the work on amphibians, as it can be and has

been carried on with methods far superior to those avail-

able in the work on higher vertebrates.

Before entering into details the most prominent facts

as revealed in the amphibian experiments may be pointed

out. In the control of growth and development of the

amphibian organism, the thyroid and pituitary glands

play the most important roles. The thymus is not con-

cerned with the growth and development of amphibian

larve. The functions of the thyroid and hypophysis

glands, as far as they are revealed in the processes of

. 193

194 THE AMERICAN NATURALIST [von LV

growth and development, exhibit a remarkable resem-

blance, and the secretions of these two glands can replace

each other to some degree, but for the most part are

specific.

. Among those who have worked out these facts, J. F.

Gudernatsch, Leo’ Adler, Bennet M. Allen and his pupil

W. W. Swingle, E. R. Hoskins and M. M. Hoskins, and

P. E. Smith deserve the greatest credit. Since much of

their success is due to the extirpation of the endocrine

glands in early embryonic stages, we should mention here

also the names of three investigators, namely Gustav

Born, Herman Braus and Ross G. Harrison, who have

elaborated the delicate technique employed in the extir-

pation experiments and thus have made possible the

progress which has been derived from them.

We will begin with the thyroid mechanism, as it has

been studied more thoroughly than other glands, and in

fact seems to be the chief factor in the control of growth

and development. Its study in the amphibians as well

as the entire work on amphibians was initiated by the

well-known experiments on thyroid-feeding to tadpoles

as carried out by Gudernatsch (1).

If tadpoles are fed fresh thyroid gland or are kept in

water to which minute amounts of thyroid extract are

added, a remarkable acceleration of development takes

place. This development is the more conspicuous as it

may occur with complete absence of growth. In tadpoles

it is characterized especially by the sudden development

of the fore limbs, by the atrophy of the tail, a sudden

protrusion of the eye-balls (2), by the rapid shortening

of the spiral gut, by the precocious atrophy of the organs

of the larval mouth, which are replaced by the frog mouth

(3), and by precocious ossification (4). These experi-

ments have been repeated with the larve of salamanders,

in which the precocious occurrences of the first moult, the

atrophy of the gills and absorption of the fin of the tail,

are most conspicuous effects of the thyroid applica-

tion (5).

No. 638] INTERNAL SECRETIONS OF AMPHIBIANS 195

The rapidity with which these processes may take place

is one of the most remarkable features of the thyroid

effect. Two normal larve of the species A. opacum, for

instance, metamorphosed at an age of 86 days and meas-

ured 60 mm. at this time. Six other larve of the same

brood were placed in an emulsion of iodothyrine at an

age of 35 days, at which time they measured 30 mm. on

the average. One week later, at an age of only 42 to 43

days and a size of 24 mm., all -had metamorphosed.

Moreover, 5 days after metamorphosis, i.e., at an age of

47 days in one animal which was examined in sections,

the visceral skeleton had undergone all those complicated

changes through which the gill arches of the larve de-

velop into the hyoid apparatus of the adult. The effect

of the thyroid hormone is quantitative; the acceleration of

the amphibian metamorphosis increases with increasing

concentration of the thyroid emulsion, as shown in

Table I.

TABLE I

QUANTITATIVE EFFECT OF THE THYROID HORMONE IN THE ACCELERATION OF

THE METAMORPHOSIS OF Ambystoma maculatum

Age at Metamorphosis

Quantity of Iodothyrine |

Control | Iodothyrine | Difference

0.1 gm. iodothyrine in 1,000 c.c. of a 101 days

0.01 e iodothyrine in 1,000 c.c. of

CO Oe Cee Oe 8 ee 6b em EENE 8s

33 days | 67 per cent.

é 58 “ee | 28 as

In one experiment the larve of Ambystoma maculatum

were kept in water, to which 0.1 gm. of iodothyrine per

1,000 c.c. of water had been added; in another experiment.

only 0,01 gm. of iodothyrine was added to 1,000 c.c. of

water. In the first experiment all larve metamorphosed

13 days after the first application of iodothyrine; in the

second experiment metamorphosis took place 39 days (on

the average) after the first application of iodothyrine.

The difference between the normal time of metamor-

phosis and the time of metamorphosis of the experi-

mental larve was 67 per cent. in the first experiment and

28 per cent. in the second.

196 THE AMERICAN NATURALIST [Vou. LV

It is remarkable that the administration of the same

amount of iodothyrine causes metamorphosis of sala-

mander larve of different species in nearly the same in-

terval of time. Thus, 0.1 gm. iodothyrine per 1,000 c.c.

of water caused metamorphosis of A. opacum larve in 7

days, of A. maculatum larve in 13 days and of A. tigrinum

larve in 13 days. The time required to induce metamor-

phosis in thyroid-fed tadpoles decreases with increasing

age of the tadpoles. Gudernatsch (1) found that thyroid

feeding caused metamorphosis in 20 days if tadpoles of

a certain age were employed, in 6 days, if tadpoles 7 days

older than the first lot were employed, and in only 4 days

if the tadpoles were 14 days older than the first lot.

As pointed out above, the larve, when fed thyroid sub-

stance, may undergo the most remarkable development,

although no growth may take place. This seems to be of -

great importance in many ways. In all organisms de-

velopment and growth, under normal conditions, proceed

in a parallel way. The behavior of the thyroid-fed larve

suggests that the reason why no development takes place

without growth is the fact that, under normal circum-

stances, the substances which cause development of cer-

tain organs are supplied through the same reactions

which control the growth of the organism. If these sub-

stances are supplied to the organism from without,

development may proceed at a higher rate than growth

or may proceed even in the complete absence of growth

and thus the relation between growth and development

may become changed as in the thyroid-fed larve.

The changes of the relation between growth and de-

velopment furnish an important link in the chain of facts

that we must know in order to understand the mechan-

ism of the thyroid apparatus as well as that of the am-

phibian metamorphosis. Although under certain condi-

tions growth may be inhibited completely upon the

feeding of thyroid, this is not always the case. Both the

rate of development and the rate of growth are dependent

No. 638] INTERNAL SECRETIONS OF AMPHIBIANS 197

on the quantity of thyroid substance administered to the

larve. Up to-a certain quantity, growth as well as de-

velopment is accelerated; if the quantity administered is

further increased, growth becomes more and more in-

hibited, while differentiation is increasingly accelerated.

With very large doses the thyroid substance may effect

even a decrease in the size and weight of the larve; while

development of the limbs is greatly accelerated in the

beginning, it finally stops and the animals die a

emaciation (6).

Kendall (7) has shown that in man the thyroid hormone

increases the basal metabolism in a strictly quantitative

way. Determinations of metabolism have not been made

in amphibians, but the behavior of the thyroid-fed tad-

poles as described above indicates that the thyroid hor-

mone also increases highly the metabolism of the cold-

blooded organism. If too much .of the hormone is

administred, metabolism is increased in such a manner

that catabolism becomes higher than anabolism, since

the organism no longer is capable of supplying enough

food materials from outside to maintain a positive meta-

bolic balance, and consequently the body substance itself

is broken down and a decrease in size and body weight

takes place. Finally even development becomes impos-

sible. For this reason, as Lenhart (6) showed, more.

thyroid substance can be administered without leading

to a check of development if the thyroid-fed larve are

either kept under conditions which decrease metabolism,

ie. in low temperature, or are fed on substances

(carbohydrates) which can be made easily available for

metabolism.

From these facts it seems ev cident that the amphibian

metamorphosis is the result of a highly increased metab-

olism, or more correctly, metamorphosis seems to result

if metabolism is increased in such a degree and manner

that catabolism becomes higher than anabolism. The

question arises whether substances or agents other than

198 THE AMERICAN NATURALIST [Von. LV

thyroid substance can cause such an increase of metab-

olism as to bring about metamorphosis. Several experi-

ments have been carried out to answer this question. But

until thyroidectomized tadpoles have been employed in

these experiments, no definite conclusions are possible;

in larve possessing a normal thyroid gland it can not be

decided whether the experimental conditions employed

have caused metamorphosis by raising the metabolism

directly or merely through the intermediation of the thy-

roid by precociously releasing the thyroid hormone.

Powers as well as Barfurth has shown that a sudden

cessation of food supply results in precocious metamor-

phosis. Although this is certainly true, it does not decide

the point in question, but may mean that sudden starva-

tion may precociously release the thyroid hormone. At

any rate, starvation in itself does not cause metamorpho-

‘sis, but is effective only if well-fed larve which are ap-

proaching metamorphosis and possess a thyroid capable

already of functioning are suddenly starved. The same

criticism applies to Kaufman’s (8) recent experiments,

in which an advanced neotenous larva (axolotl) of Am-

bystoma tigrinum was given salicylic acid, whereupon it

metamorphosed promptly. This result is extremely in-

- teresting as it raises most urgently the question whether

the action of iodothyrine is specific and whether the

changes of metabolism resulting from thyroid adminis-

tration are merely quantitative or also qualitative. As

pointed out, Kaufman’s experiment, however, does not

answer any of these questions.

In accord with the highly increased catabolism as

caused by the action of the thyroid hormone is the fact

that metamorphosis, in its initial stages, appears to be

more a process of profound atrophy than one of construc-

tive development, although phenomena of the latter

kind frequently accompany the degenerative processes.

Among the most conspicuous processes of destruction

are the complete atrophy of the gills and the entire vas-

cular apparatus which serves the gill circulation, a con-

No. 638] INTERNAL SECRETIONS OF AMPHIBIANS 199

siderable destruction of the cartilaginous visceral skele-

ton, the atrophy of the larval mouth in anurans, the

reduction of the intestinal coils in anurans, the complete

atrophy of the tail in anurans and the atrophy of the fin

in urodelans. Not before this extensive breakdown of

the larval tissues has taken place and out of the remnants

of the destroyed organs the new organs of the adult

develop. Particularly instructive in this regard is

the development of the epithelial bodies in the larve of

salamanders; these develop from the epithelium of the

destroyed gills and in the midst of the masses of detritus

which result from the destruction especially of the gill

vessels.

The fact that metamorphosis can be brought about by

feeding mammalian thyroid substance to the amphibian

larve, does not of course prove that the amphibian meta-

morphosis, under normal circumstances, is the result of

the function of the amphibian thyroid gland itself. This,

however, is the case, as demonstrated especially by the

work of Allen (9) and of E. R. and M. M. Hoskins (10).

If in an early embryonic stage of the anuran organism

the thyroid is extirpated, metamorphosis can not take

place at all and the tadpoles remain permanently (as far

as the observations go) in the stage of an aquatic am-

phibian larva. Growth likewise is ultimately interfered

with, although the thyroidectomized tadpoles may grow

more rapidly in the beginning and even grow larger than

normal tadpoles. On the other hand, if the thyroid of

metamorphic tadpoles is grafted to tadpoles which are

in early larval stages, metamorphosis of the latter, up to |

the stage of the larve from which the thyroid graft was

taken, is caused (11). The metamorphosis of the am-

phibian eye is likewise impossible if it is removed from

the influence of the thyroid hormone which controls the

development of the eye. If eyes of old salamander larve

are grafted to young larve, the metamorphosis of the

graft may be retarded by as many as 7 months and will

200 THE AMERICAN NATURALIST (Vot. LV

not take place before the eyes of the host metamorphose.

On the other hand, eyes of young larvae, if they are

grafted to old larve, can be made to metamorphose earlier

than they would under normal conditions (12).

It has been said that the thyroid substance does not

actually produce new characters, but merely accelerates

the rate of their development which is predetermined by

heredity. There can be little doubt, however, that the

advance of the higher vertebrates from an aquatic stage,

with open gill slits and internal or external gills, and in

particular all the characters distinguishing the terres-

trial amphibian from the aquatic larva, could not have

developed if the thyroid apparatus had not attained, at

some evolutionary stage of the amphibians, its present

function. For the benefit of those who might think that

the relatively short time (about 14 years) of observation

in Allen’s and Hoskins’s experiments does not justify this

statement, I may refer to the Texan cave salamander,

Typhlomolge rathbuni which illustrates in a most vivid

manner the effect of the absence of the thyroid gland.

This salamander never develops beyond the larval stage,

retaining permanently its external gills and other larval

organs. An examination of the endocrine system of this

animal was made by Emerson (13); it revealed the com-

plete absence of the thyroid gland. It is worth while to

mention briefly another interesting condition observed in

this animal, namely the almost complete lack of pigment,

a condition somewhat similar to that observed by Smith

and by Allen in hypophysectomized tadpoles, and the

highly atrophied state of the eyes. TZ yphlomolge is a

white, blind salamander. These latter peculiarities have

been attributed frequently to the absence of light in the

caves, a theory which at first seems very plausible. It

would not be surprising, however, if some day these char-

acters should be found to be the result of endocrine dis-

turbances. Similar to T'yphlomolge in all the charac-

teristics mentioned above is a European salamander,

Proteus anguineus, which inhabits the Austrian lime-

No. 638] INTERNAL SECRETIONS OF AMPHIBIANS 201

stone caves; nothing, however, is known about the endo-

crine glands of this animal.

If: the thyroid substance is capable of causing the de-

velopment of the characters of a terrestrial amphibian,

the administration of thyroid substance should cause

metamorphosis of Proteus anguineus. Jensen (14)

subjected Proteus to the action of thyroid substance, but

did not get any demonstrable results. Many causes may

have been responsible for this failure, in particular the

fact that the animals were too old when they were sub-

jected to the thyroid feeding.

It has been known for some time that the effect of equal

doses of thyroid substance on the amphibian metamor-

phosis is the greater, the more iodine there is contained

in the thyroid gland (15). Recently, Swingle (16) has

demonstrated that the feeding of common inorganic

iodine to tadpoles or the keeping of the tadpoles in iodine

solutions accelerates metamorphosis in the same way as

does the thyroid. This effect of iodine is strictly quanti-

tative; if there is no iodine contained in the food of the

tadpoles, metamorphosis is inhibited, while with an in-

creasing amount of iodine metamorphosis is increasingly

accelerated. Moreover, the effect on the relation between

growth and development is the same in iodine solutions

and in thyroid feeding. Weak solutions of iodine increase

-not only the rate of development, but also the rate of

growth, while high concentrations prevent growth.

There can be no doubt that at least in the metamorphosis

of tadpoles, iodine is an indispensable constituent of the

thyroid hormone.

Swingle (16) found that potassium iodide and do

-form had an effect on metamorphosis similar to that of

iodine, while bromine had no effect on metamorphosis

and growth. Thus the effect of iodine appears to be very

specifie when comparison is made with so nearly related

a substance as bromine.

The feeding of iodine to mammalians does not produce

the same effects as the administration of thyroid sub-

202 THE AMERICAN NATURALIST [Von. LV

stance. This fact has formed the basis for the opinion

(7) that the characteristic action of the thyroid hormone

is not directly caused by the presence of iodine in the

thyroid hormone. It seems, however, more probable that

the feeding of iodine has no effect on mammalians, be-

cause the mammalian organism, for some reasons, can

not utilize an excess of iodine. It is well known that the

mammalian thyroid gland is capable of storing large

amounts of iodine (17). If, under normal conditions,

only a definite amount of hormone could be excreted by

the thyroid gland, the feeding of excess amounts of iodine

would have no effect in the healthy individual, since every

excess of iodine would be retained and stored by the thy-

roid tissue. If in the mammalian organism the thyroid

gland should be the only organ capable of elaborating the

thyroid hormone, the feeding of iodine could have no

effect in the absence of the thyroid, or in persons whose

thyroid function is insufficient. Conditions are different

with tadpoles. Swingle (16) has shown that even in

thyroidectomized tadpoles, iodine solutions are capable

of causing metamorphosis. Apparently the thyroid

gland is not the only organ of the tadpole which can pro-

duce the thyroid hormone. —

It should be pointed out, however, that a fundamental

difference exists between frogs and toads on the one

hand, and salamanders on the other, as regards their .

reaction to iodine. Salamanders behave much like mam-

malians. Although I was able to confirm the accelerating

action of iodine at least in the development of the limbs

of the tadpoles, I have not been able to cause precocious

metamorphosis by placing salamander larve in iodine

solutions. Table II will illustrate this statement.

Two larve of the species A. maculatum were kept first

in a solution of 5 drops %o m. iodine per 1,000 c.c. water

and then, up to metamorphosis, in a 3-drops-iodine solu-

tion. No acceleration of metamorphosis took place; the

larve metamorphosed at an age of 122 days, while the

controls were only 101 days old when they metamor-

No. 638] INTERNAL SECRETIONS OF AMPHIBIANS 203

phosed. It is interesting to note that while tadpoles of

Rana sylvatica are killed by a 5-drops-iodine solution and

upon a 3-drops-iodine solution respond promptly with

development of the hind limbs, the larve of A. maculatum

showed no other effect from a 5-drops solution than a

slightly decreased food intake. The latter cireumstance

may account for the longer duration of the larval period

of the experimental larve. Since it was believed that in

this experiment the solution was too weak, 2 larve of the

same species, after a short sojourn in a 3-drops solution,

were placed in an 8-drops-iodine solution; but as Table

II shows, in this experiment also metamorphosis was not

accelerated by the iodine solution. Several larve were

fed directly crystals of iodine to make sure that the in-

effectiveness of the iodine solution in salamanders was not

due to a possible impermeability of the larval skin for

iodine. In one case two crystal-fed larve metamor-

phosed at 124 days, while the controls metamorphosed at

the age of 101 days. In another experiment, in which 3

larve were employed, one metamorphosed at the age of

89 days, while the controls metamorphosed at 80 days.

Of the two other larve, one did not show any signs of

metamorphosis when it was killed for histological ex-

amination; the other one died from an overdose of iodine,

but did not show any sign of metamorphosis.

TABLE II

Iopins Has No EFFECT ON THE METAMORPHOSIS OF A. maculatum

Age at Metamorphosis.

Quantity of Iodine. Iodine Sol.

Normal. | Iodine Solution. 4 Crystals.

poe lin aieaat a iodine in 1,000 c.c.

Oe 0 Be Ree 28 Soe Oe OW CR e wee Bl ee

101 days | 122 days | 124 days

? |

80 “a 79 ‘ 89 Pst

Three old larvæ, all of the axolotl type, and one neo-

tenous, of the western race of Ambystoma tigrinum,

which were collected in the Rocky Mountain lakes last

204 THE AMERICAN NATURALIST [Vou. LV

summer, were subjected to an iodine treatment. They

were placed in water containing 5 drops of a 20 m. solu-

tion of iodine per 1,000 c.c. of water and, as they showed

no reaction of any kind, this concentration was increased

gradually to 8 drops and in one larva to even 13 drops of

iodine (3 drops of a %o m. solution of iodine per 1,000 c.c.

of water is enough to cause growth of the hind limbs in

larve of Rana sylvatica), which is more than 0.2 c.c. of

a %o m. solution of iodine per 1,000 c.c. of water. Although

these larve have now been in the iodine solution for 2

months, none of them has developed any tendency

towards metamorphosis, while 3 other control larve,

among them a neotenous specimen, metamorphosed 13

days after being placed in an emulsion of 0.1 gm. of

Bayer’s iodothyrine per 1,000 c.c. of water. Evidently

the assumption suggested by Swingle (16, III), that lack

of iodine prevailing in the lakes is causing the inhibition

of metamorphosis of the axolotl and other urodelans, is

unwarranted. I will show presently that in the inhibition

of metamorphosis and in neoteny of axolotls and probably

certain European urodelans we are confronted with an

entirely new phase of internal secretion, namely with the

differential action of temperature upon the development

of various components of the endocrine system.

In a former article (20) I suggested that the inhi-

bition of metamorphosis in thymus-fed amphibian larve

may be caused by lack of iodine in the thymus. Swingle

(16, III) has accepted and unfortunately repeated, with-

out further testing, this suggestion. But recent experi-

ments show that this view must be abandoned, since

addition of iodine to a pure thymus diet does not enable

the salamander larve either to grow or to metamorphose.

Similarly the retardation of growth and metamorphosis

of salamander larve kept on a pure diet of posterior lobe

of hypophysis remains unaffected if iodine is added to

the water.

The iodine requirement of salamanders must be ex-

tremely slight, since anterior lobe of hypophysis, a nearly

No. 638] INTERNAL SECRETIONS OF AMPHIBIANS 205

iodine-free diet, does not in any way retard growth or

metamorphosis.

There are several species of salamanders (Autodax

lugubris, Autodax iecanus) whose young do not emerge

from the eggs before metamorphosis is completed.

Although the larve of these species have no opportunity

to obtain iodine from outside, these cases do not prove, of

course, anything against the importance of iodine in the

amphibian metamorphosis; very likely the eggs of

Autodax contain enough iodine to permit metamorphosis

of the larve within the egg.

Still another difference between anurans and salaman-

ders has made itself apparent in this work. While in

tadpoles, of at least certain anuran species, the develop-

ment of the legs is, in some as yet unknown way, dis-

tinctly under the control of the thyroid, the leg-develop-

ment in salamanders is independent of the thyroid gland.

Both hind and fore limbs develop in a normal way after

thyroidectomy in salamander larve, as shown by E. R.

and M. M. Hoskins (10). Moreover, the development of

the legs is not accelerated if the larve are kept in solu-

tions of iodothyrine (18); this is the case even if the

administration of iodothyrine is commenced soon after the

eggs have been deposited. Consequently, it is very com-

mon to find that the larve metamorphose in the iodo-

thyrine solution before the legs are completely developed.

It is evident that in tadpoles part of the larval develop-

ment is controlled by the thyroid function, since neither

the hind limbs, from a certain stage on, nor the fore limbs

can develop in the absence of the thyroid (9, 10, 19).

Apparently the anuran thyroid gland begins to secrete

already in the larval period. In salamanders the larval

development seems to be highly independent of the thy-

roid function and it is quite probable that the salamander

thyroid does not begin to function much before the first

moult. This can be demonstrated in the following way

(12). If eyes of old larve which, however, are still far

enough from metamorphosis, are grafted on to young

206 THE AMERICAN NATURALIST [Vor. LV

larve, their metamorphosis is inhibited until the host

metamorphoses. If the eye graft, however, is taken from

larve which are near metamorphosis, such an inhibition

is no longer possible. Apparently shortly before meta-

-morphosis actually occurs, the thyroid begins to excrete,

and after the circulating hormone has reached the eye

metamorphosis of the eye takes place, even if the organ

is transferred to an animal in which the thyroid hormone

has not yet been secreted. ,

It is quite possible, that the late beginning of the thy-

roid function in salamander larve is one of the causes

why the administration of an excess of iodine is ineffec-

tive in the metamorphosis of these amphibians. Probably

the thyroid merely stores up the excess of iodine, but

does not release the hormone till shortly before the first

moult.

Allen (19) has recently examined the condition of the

thyroid of Colorado axolotls and has found that they

possess a thyroid corresponding in size, structure and

colloid content to the thyroid of adult specimens of A.

tigrinum. The thyroid of the larve of other salamander

species likewise seems to be mature much before meta-

morphosis actually takes place. Allen concluded from

his observations that the thyroid of salamanders begins

to function at an early stage of the larve. The inde-

pendence of the larval development of the salamander

larve as demonstrated by the facts mentioned above

shows, however, that the presence of a mature thyroid

before metamorphosis must be interpreted in a different

way. The most conspicuous character in the salamander

metamorphosis is the fact that, although it certainly is

dependent on the thyroid hormone, it does not necessarily

take place in larve whose thyroid is mature. This can

only mean that two factors are required in order to bring

about the metamorphosis of salamander larve, namely a

mature thyroid gland and a factor which releases the

thyroid hormone from the follicles of the gland.

This conception, which is now supported by several

No. 638] INTERNAL SECRETIONS OF AMPHIBIANS 207

facts, is also capable of explaining the problem of neoteny

of the so-called axolotl. In the course of experiments

carried on during several years in the laboratory, and by

inspection of the conditions prevailing in the Rocky

Mountain lakes, the natural habitat of the American

axolotl, I have become convinced that the neoteny of this

species is due to the effect of low temperatures. We have

in the amphibians an experimental material in which the

relation between the development of the body and certain

endocrine glands can be changed by the influence of tem-

perature, owing to the differences of the temperature

coefficients of the processes governing the development

of different glands. _ .

Although my experiments are not yet finished, they

seem to permit the following conclusions in connection

with my field observations:

1. The thyroid gland of salamanders undergoes a de-

velopmental change consisting of two periods, one of

early development, lasting at least 63 weeks, in the course

of which the thyroid becomes more and more sensitive to

the action of a releasing factor (called excretor substance

in my earlier work) and one of aging in the course of

which the thyroid loses gradually its sensibility to the

releasing factor.

2. In order to release the hormone of the thyroid gland,

a particular releasing factor is required (the nature of

which is entirely unknown); the quantity of this factor

necessary to release the thyroid hormone depends on the

sensitivity of the thyroid gland. Metamorphosis can

take place only if the thyroid is sensitive and is acted

upon by the proper quantity of the releasing factor.

3. The temperature coefficient for the elaboration of

the releasing factor is higher than the temperature coeffi-

cients for growth and the thyroid change.

The following facts seem to warrant these assumptions:

1. Salamander larve, kept at an identical temperature,

are nearly all of the same size when they metamorphose.

Larve kept at low temperatures grow considerably larger

208 THE AMERICAN NATURALIST [Vou. LV

than those kept at high temperature, before they can

metamorphose. This is shown in Table III (2). The

temperature coefficient for the releasing factor is higher

than that for growth.

TABLE III

TEMPERATURE AND SIZE OF THE METAMORPHOSING LARVÆ

| Size in Mm.

Species Series | ie ea Series

ORI OS OES ATE RT i ee C 1916

XIV 1918 | 61 -n XVIII 1918

Bonum o a e ae S1 | 100. 296 U 191

XLVI.19190. | ..108..| 122 XLVIII 1919

Maculatum.............. LAXV 1920 | a | = 89 LXXVII 1920

2. In very low temperatures (6° C. to 10° C.) growth

is greatly slowed down and consequently the elaboration

of the releasing factor must be still more retarded; vet

larve kept at 6° C. grow less and less large before meta-

morphosis, when they are transferred, at increasing

ages, to 15° C., as shown by an experiment lasting 63

weeks thus far. Apparently the thyroid has gone on to

mature at a relatively high rate and at 63 weeks is highly

sensitive and responds to smaller quantities of the releas-

ing factor. The temperature coefficient for the thyroid

change is considerably lower than those for growth and

for the elaboration of the releasing factor.

3. If the thyroid can continue to develop in the absence

of growth, it probably can also commence to age. Should

this assumption be correct, the larve kept at 6° C. should

finally become unable to metamorphose, if the time dur-

ing which they are kept in 6° C. is sufficiently long. At

present this assumption would explain why many speci-

mens of the Colorado axolotl yield only slowly, if at all,

to the influence of high temperature, and the Mexican:

axolotl frequently loses completely its ability to meta-

morphose.

4. The Colorado axolotls reach frequently a size con-

siderably in excess of the normal maximum size of the

No. 638] INTERNAL SECRETIONS OF AMPHIBIANS 209

species as calculated from the largest known terrestrial

specimens of the eastern race of this species; the Colo-

rado axolotls are giants. Since sexually mature speci-

mens of the eastern race of A. tigrinum become giants if

they are fed anterior lobe of hypophysis, the gigantism

of the sexually mature axolotl could be explained if any

indications of hyperpituitarism of these animals could be

discovered. On the assumption that in spite of the pres-

ence of a large thyroid the function of this organ is sup-

pressed by the absence of the releasing factor, the over-

function of the axolotl hypophysis would be very

plausible, since, as will be pointed out later on, the ab-

sence of the thyroid function causes hypertrophy of the

hypophysis in amphibian larve.

5. The maturing of the sex organs of the axolotl is not

incompatible with the assumption of an athyroidism,

since, as will be discussed presently, there can be no

longer any doubt that the development of the sex organs

of amphibians is entirely independent of the thyroid

hormone.

6. The assumption that the temperature effect can ac-

tually produce the complex phenomenon of neoteny is

supported by the fact that the species A. tigrinum be-

comes neotenous only in the high and cold regions of the

Rocky Mountains and the Mexican high plateau, while

in the eastern part of the United States all individuals

of this species metamorphose in a normal manner. I

have examined the conditions prevailing in the Rocky

Mountains; to summarize briefly my observations, the

axolotl is regularly found only in those lakes which are

permanently exposed to low temperatures, while in the

shallow lakes of lower altitudes axolotls are found only

during some years and are absent during other years;

apparently a succession of several years favorable in

temperature conditions is required to produce the axolotl

state.

7. A. tigrinum is the only species of North American

salamanders which becomes neotenous. This is prebably

210 THE AMERICAN NATURALIST [Vou. LV

not due to differences existing between the endocrine

system of the numerous species inhabiting the United

States, but is explained by the fact that A. tigrinum,

among the closely related species which I had an oppor-

tunity to test, is the only species that can stand tempera-

tures low enough to bring about the necessary differ-

ence between the rate of the thyroid development and

that of the elaboration of the releasing factor.

8. The fact that many individuals among the offspring

of female specimens of the Mexican axolotl do not meta-

morphose even if they are brought, immediately after

hatching, into conditions permitting normal metamor-

phosis of other salamander species, is not necessarily

related to the factors discussed above, but may be due

to the development of congenital thyroid disturbance in

the young born by an athyroidous female.

It is, of course, well known that many structural

changes, only a few of which have been studied, are re-

quired to make, out of the aquatic larve, the terrestrial

amphibian. ‘This is true for the anurans as well as for

the urodelans. Since we know that the complex phe-

nomenon of metamorphosis is initiated by the thyroid

effect, the question arises now which of the component

changes are directly caused by the action of the thyroid

hormone. The fact that certain developmental processes

frequently take place upon thyroid administration and

therefore are a very convenient indicator in studying

quantitatively the effect of thyroid substance, of iodine

or of any other metamorphosis-causing agent, does

not mean, in itself, that these developmental processes

are caused directly by the action of the thyroid; it is pos-

sible and indeed supported by many facts, that certain

of these changes will follow automatically, after the

initial changes have been effected by the thyroid action.

Thus, while under normal conditions, the pigmentary

pattern, the legs, the tongue, the palatal teeth and the

sex organs mature in salamanders: during metamor-

phosis, they can be shown to be highly independent of

No. 638] INTERNAL SECRETIONS OF AMPHIBIANS 211

the thyroid action at least in salamanders and may, under

certain conditions, occur in the absence of this action or

not occur in the presence of it. I have pointed out in

former articles (18) that among the many changes occur-

ring during the salamander metamorphosis there are two

which seem to be particularly closely related to the thy-

roid action, namely the first shedding of the skin and the

reduction of the gills to mere stubs. While the succession

of all the other changes enumerated above seems ex-

tremely variable, the order in which the first moult and

the reduction of the gills follow each other could not be

changed as yet by any of the procedures employed, inas-

much as the shedding of the skin always is followed by

the atrophy of the gills. Moreover, these two phenomena

have never failed to occur in the metamorphosis of the

many hundreds of metamorphosing larve observed in

the laboratory, and even in such larve as were forced

at a very early date into precocious metamorphosis by

the administration of iodothyrine and in which other

changes did not occur. And furthermore, neither the first

moult nor the reduction of the gills could ever be ob-

served in larve, whose metamorphosis was inhibited by

dietary or other means. Thus I have come to look upon

the first moult and the atrophy of the gills as two of the

primary components of the salamander metamorphosis.

I have not enough personal experience with the larve of

anurans, but feel encouraged through the experiences

reported by other investigators to believe that in anurans

these phenomena play a similarly important role. Cer-

tainly the first shedding of the skin seems to accompany

true metamorphosis in salamanders and tadpoles as well

(38), and substances other than thyroid hormone or

iodine, such as the anterior lobe substance, although they

may cause the limbs to grow, do not bring about atrophy

of the gills in thyroidectomized tadpoles (27).

It is different with the limbs, the pigmentary pattern,

the tongue, the palatal teeth and the sex organs; these

five groups of organs, at least in salamanders, have

212 THE AMERICAN NATURALIST (Von. LV

proved to be little influenced by the thyroid action. That

the development of the limbs of salamanders is not de-

pendent on the thyroid gland has been pointed out above;

here I may add that Typhlomolge is a further illustration

of this fact, as in this salamander the legs develop in a

normal manner in spite of the complete absence of the

thyroid gland. The relation of limb development and

thyroid action in tadpoles is by no means definitely

settled as yet. In tadpoles the development of the limbs

seems to be highly dependent on the action of the thyroid

gland; but attention has been called to this surprising

difference between two groups of organisms so closely

related otherwise and the suggestion has been made in a

previous article (18), that this difference as far as the

fore legs are concerned may be due merely to the fact

that in tadpoles the limbs grow beneath the skin and

consequently can not break through unless the changes

are initiated which finally lead to the shedding of the skin

and that these changes and not the thyroid action are the

primary factor in the development of the anuran fore

limbs. Whether or not this assumption is correct can

not be decided at present, but certainly deserves renewed

attention in view of recent discoveries which demonstrate

that the development of the limbs of tadpoles, at least in

certain species, is not as dependent on the thyroid secre-

tion as some investigators were inclined to think. Allen

(34), who has made prolonged observations in thyroidec-

tomized tadpoles, has recently found that not only the

hind limbs, but even the fore limbs in the thyroidectom-

ized larve of Bufo ultimately attain a size and differenti-

ation not only equal but superior to those attained in nor-

mal metamorphosing larve. The only difference, how-

ever, is that in the absence of the thyroid gland the fore

limbs can not break through the skin.

As to the skin pigmentation, it is well known that larve

in which metamorphosis has been inhibited for some

reasons may develop a nearly adult pigment pattern. In

larve of A. opacum which were fed thymus gland, and

No. 638] INTERNAL SECRETIONS OF AMPHIBIANS 213

consequently did not metamorphose, the coloration of the

skin advanced to a stage very similar to that of an adult

animal. On the other hand if young larve of A. opacum

are made to metamorphose precociously by means of the

application of iodothyrine, metamorphosis takes place,

while the color pattern remains in an early larval stage.

Through the observations of Cope (35) it has become

known that otherwise completely metamorphosed indi-

viduals of the species A. tigrinum may exhibit either a

larval condition of the tongue or larval characters of the

palatal teeth or larval characters in both the tongue and

the palatal teeth.

In nature it is not uncommon that the sex glands of

salamanders develop to complete maturity while the rest

of the organism remains in a larval stage (18). This

phenomenon, known by the name of neoteny, illustrates

that the sex organs can develop in the absence of the thy-

roid function. The same fact has been shown in the

larve of anurans by B. M. Allen and his coworkers. In

thyroid-fed frog larve, which have undergone precocious

metamorphosis, the sex organs do not seem to be further

developed than those of normal larve of the same age

(3). Moreover, if the thyroid is removed from the larve

and metamorphosis inhibited, the sex organs develop at

the same rate as in normal larve (21). Hoskins (22

and Allen (21) showed that the testicle of thyroidectom-

ized tadpoles may develop ripe spermatozoa. These

facts, however, can not be interpreted to mean that the

germ plasm is independent of the somatic plasm, in the

Weismannian sense. The characteristic feature of the

amphibian development is not the independence of the

germ plasm from the somatic plasm, but the inde-

pendence of various groups of organs from one another,

due to the fact that the development of each of these

groups is controlled by substances different from those

controlling the other groups, and that each of these sub-

stances separately may be supplied to or withheld from

the organism either by the experimenter or by conditions

214 THE AMERICAN NATURALIST [von LV

not fully known as yet (18). One of these conditions is

the temperature as has been pointed out above.

I will discuss briefly now the rôle of the hypophysis in

the growth and development of amphibians. The most

noteworthy fact seems to be the existence of a remark-

able resemblance between the functions of the amphibian

thyroid and hypophysis glands during the larval period.

If the hypophysis gland is extirpated in early embryonic

stages, the tadpoles stop to develop at a stage at which,

in normal tadpoles, metamorphosis begins. Growth, too,

is inhibited in the hypophysectomized, tadpoles (23, 24).

In a series of extremely interesting experiments Allen

(25) showed that both growth and development can be

restored to the hypophysectomized tadpoles, if the an-

terior lobe of the hypophysis of an adult frog is grafted

to such larve. No other part of the hypophysis when

grafted to the hypophysectomized tadpoles can restore

growth and development, and it is certain, therefore, that

it is the anterior lobe of the hypophysis which controls

the growth and development of the larve. In tadpoles

the feeding experiments as made by P. E. Smith (26)

seem to corroborate the extirpation experiments. Feed-

ing of anterior lobe to hypophysectomized tadpoles in-

creases the rate of growth to such an extent that growth

becomes as vigorous as in normal larve. Moreover, at

the time when the normal tadpoles metamorphose and

growth ceases for a time, the anterior lobe-fed hypophy-

sectotnized tadpoles continue to grow and finally attain

a size in excess of that of normal larve. Ultimately, how-

ever, the growth of these larve stops and before the size

is reached characteristic of the normal adult animal.

The effect of feeding anterior lobe to normal larve is a

matter still under discussion at present. Smith (26)

found that normal tadpoles when fed anterior lobe grew

apparently at a slightly higher rate and also metamor-

phosed at a slightly earlier date than normally fed tad-

poles. Recently, however, Smith (36), on account of the

considerable variation in the rate of growth and develop-

No. 638] INTERNAL SECRETIONS OF AMPHIBIANS 215

ment of normal larve, seems to be inclined to consider

these differences as being of no significance. Certainly it

is of no small importance that normal and hypophysee-

tomized larve react so differently to a diet of anterior

lobe substance; apparently part of the active principle

of the anterior lobe introduced, by the diet, into the organ-

ism is made ineffective in the presence of a normal

hypophysis. Not yet completed experiments on sala-

: mander larve seem to suggest that the larval growth of

salamanders at least can not be affected by feeding an-

terior lobe of hypophysis; this may be due either to a

destruction of the active principle in the digestive tract

or to some peculiarity in the metabolism of the salaman-

der larve, and is of particular interest with regard to the

fact that the adult salamanders react very markedly to

an anterior lobe diet, as will be discussed presently.

One of the most pertinent and yet most difficult prob-

lems of endocrinology is presented by the existence of

interrelations and interactions between the various endo-

crine glands. There can be no doubt that in tadpoles

such an interrelation exists between the hypophysis and

thyroid glands. Thus Rogers (31) and later on Hoskins

and Hoskins (22) found that upon thyroidectomy per-

formed in early embryonic stages of anurans the anterior

lobe of the hypophysis shows a tendency towards hyper-

trophy. On the other hand if the buccal anlage of the

hypophysis is removed, the thyroid soon ceases to grow

and to differentiate and finally presents a state of hypo-

plasia, as shown by Allen (30) and by Smith (36). Since

the effects of the extirpation of either of these glands on

general body growth and on development are quite simi-

lar and since the behavior of each of these glands after

the extirpation of the other one demonstrates the exist-

ence of an interrelation between them, the question might

well be asked, if the function of each of these glands can-

not be replaced by the hormone of the other one of them.

Although this question can not be satisfactorily answered

thus far, it seems highly ‘probable that these hormones

216 THE AMERICAN NATURALIST [Vou. LV

are strictly specific in as much as neither of them can re-

place the function of the missing one. To quoting the in-

hibition of metamorphosis and growth following hypoph-

ysectomy as proof in favor of this view one could ob-

ject that in this particular case the thyroid can not effect

metamorphosis and growth merely on account of its

atrophic condition. Smith (36), however, found, that in

certain cases of partial hypophysectomy the thyroid re-

mains completely unaffected and, yet no metamorphosis

takes place; only if the remaining fragment of the epi-

thelial hypophysis grows large enough to come in contact

with the neural hypophysis, metamorphosis can be ef-

fected. For this reason Smith takes the view that the

function of the hypophysis is indispensable in metamor-

phosis and that the secretion necessary for this purpose

can only be elaborated, if epithelial and neural hypoph-

ysis are in contact with each other. That neither the

anterior nor the posterior lobe of the hypophysis contains

the substance necessary for metamorphosis and that this

substance can be produced only in the body itself, requir-

ing for its elaboration the contact between neural and

buccal hypophysis, seems much supported by the fact

that, although growth may be maintained up to a certain

size, by feeding anterior lobe to hypophysectomized

tadpoles, metamorphosis can not be effected in such tad-

poles by feeding either anterior or posterior lobe. As

to the possibility of replacing the function of the anterior

lobe substance by introducing into the organism thyroid

hormone or iodine, Allen (28) fed iodine to hypophysec-

tomized tadpoles and obtained some, but not all of the

changes induced by iodine in normal and thyroidectom-

ized larve and seemed to be tardily inclined to the view

that the lack of the hypophsis could be compensated for

by feeding iodine. Smith, however, in his last publica-

tion (36), claims that neither thyroxin nor thyroid gland

itself causes metamorphosis, when fed to pituitaryless

tadpoles. Quite similar are the results of feeding hypoph-

ysis to thyroidectomized larve. Hoskins and Hoskins

(27) were able to cause growth of limbs and emaeiation

No. 638] INTERNAL SECRETIONS OF AMPHIBIANS 217

by feeding anterior lobe substance to thyroidectomized

tadpoles, but could not obtain complete metamorphosis;

especially the atrophy of the tail and of the gills could not

be enforced. Similarly Allen (37) points out that feed-

ing anterior lobe of cattle does not result in metamor-

phosis ‘of thyroidectomized tadpoles.

If taken together, all these results seem to indicate that

although certain resemblances exist between the hor-

mones of the thyroid and the hypophysis glands, they are

nevertheless specific and can not replace each other as

regards at least certain functions.

As pointed out above, the metamorphosed salamanders

react on anterior lobe feeding quite differently from the

larve. Such differences in the reaction upon the same

principle in different stages have been observed quite

frequently and are apt to throw an important light on the

nature of the chemical reactions involved in growth and

development of different stages. The salamander larve

show no appreciable effect from an anterior lobe diet,

whether the anterior lobe be fed alone or in small quan-

tities added to normal food. If metamorphosed sala- `

manders of the species A. opacum or A. tigrinum are fed

anterior lobe, the rate of growth becomes almost imme-

diately accelerated and growth continues after the ani-

mals have reached the specific maximum size of the

species; they become giants. The latter result must be

attributed to the action of a specific growth promoting

hormone contained in the anterior lobe (32).

The thymus gland apparently has ‘no effect on growth

and development, although it has been believed that it

contains specifice growth-promoting and development-

retarding substances. It is true that in larve which are

fed on thymus only, growth as well as metamorphosis

are inhibited. The inhibition of metamorphosis, how-

ever, is due to the fact that in the absence of growth the

releasing factor of the thyroid can not form, as has been

mentioned above. Moreover, the inhibition of growth is

not caused by specific hormones of the thymus, but is

merely a deficiency phenomenon. The more normal food

218 THE AMERICAN NATURALIST [Vou. LV

there is added to the thymus, the less marked does the

inhibition of growth become; small amounts of thymus

added to a normal diet have no effect (33). It is un-

known at present which of the food substances necessary

for growth are missing, although it is certain that the

deficiency of the thymus is not caused by a deficiency in

iodine, calcium, sodium or potassium. Many other glands,

such as the spleen, prescapular lymph-gland, parathy-

roids, and posterior lobe of the hypophysis are more or

less Canat in the growth of salamander larve

REFERENCES

L s oP,

p S an Amphibienlarven. Vorl. Mitt. Zen-

abl. Physiol., XXVI, 323.

1912. Pram me noma on i I. The Influence of Spe-

peA rgans Giv Food, on arima and Differentiation.

h. f: Purpitiedaia: XXXV, 457.

1914, rodine Experiments on Tadpoles. II. A Further Contribu-

ion to the PEO of Organs of Internal Secretion.

ede r. Anat., XV, 431.

1916. ae on fatai cae IV. Treatment of Tadpoles

with Thyroid and Thymus Extracts. Anat. Rec., XI, 357.

2, Gudernatsch, J. F

1912. Feeding Experiments on Tadpoles. I. The Influence of Spe-

cifie Organs Given as Food, on Growth and Differentiation.

cil f. Entw. Mech., XXXV, 457.

3. Swingle, W

1918. The a i aieia of Metamorphosis in Frog Larve by Thy-

oid Feeding, and the Effects upon the ag aA Tract

i Sex Glands. Jour. Exper. Zool., XXIV, 521

4. Terry, G. 8.

1918. Effects of the Extirpation of the Thyroid Gland upon Ossifi-

tion in Rana pipiens. Jour. Exper. Zool., XXIV, 567.

. Uhlenhuth, E.

1918. Is the Influence of Thymus Feeding upon Development, Meta-

morphosis, and Growth Due to a ni eee Action of that

land?

en, Physiol., I, 473.

1919. Relation between Metamorphosis and other Developmental

Phenomena in Amphibians. Jour. Gen. Physiol., I, 525.

6. Lenhart, C. H.

1915. The Influence upon Tadpoles of Feeding Desiecated Thyroid

Gland in Variable Amounts and of Variable Iodine Contents.

our. Exper. Med., XXII, 739

Kendall, E. C.

1918. The Active Constituent of the Thyroid. Jour. Amer, Med.

Assn., LXXI, 871,

þa

—

. Lenhart, C. H.

1915

. 638] INTERNAL SECRETIONS OF AMPHIBIANS 219

. Kaufman, L.

1918 ure on the Artificial poa s of Axolotls. Bull,

ad, Sc. Cracow, Ser.

. Allen, B. =x

1916. The Results of Extirpation of the Anterior Lobe of the Hy-

ee and of the Thyroid of Rana pipiens larve. Sci-

s T06:

ce, XLIV

1918. The Faila of Thyroid Removal in the Larve of Rana pipiens.

Exper. V, 4

Jour,

. Hoskins, E. R., and Hoskins

1917. On Thyroidectomy in inhibi Proc. Soc. Exper. Biol. and

Me 74.

"7 ,

1919. Growth and Development of Amphibia as Affected by Thy-

cha etomy. Jour. Exper. Zool., XXIX, 1.

> EN W:

1920. aA Thyroid Transplants. Anat. Rec., XVIII, 263.

; doang h, E.

913. Die Tokes Metamorphose transplantierter Salamander-

gen. Arch. Entw. Mech., XXXVI, 2

1913. ie Ens sey Shen auf das transplantierte Amphibien-

auge. i oi halmologie, ITI, 343.

1917. A Kan a to the Metamorphosis of Amphibian

Organs. (The Metamorphosis of Grafted Skin and Eyes of

Pipe punctatum.) Jour. Exper. Zool., XXIV, 23

. Emerson, E

1905. rA Anatomy of T rathbuni. Proc. Soc. Nat.

History, Boston, XXXII, 4:

. Jensen,

Cc. O.

1916. Ved rp natal sired fremkald Forwandling hos

t. klg. Danske Vidensk., Selsk. For-

handl., itest is

The Influence upon Tadpoles of epn Loren Thyroid

Gland in Variable Amounts and of Variable Iodine Con-

tents. Jour. Exper. Med., XXII, 739.

Graham, A.

1916, A Study of the iy tree Activity of Adenomata of the

Thyroid Gland, n to Their Iodine Content, as

Evidenced by Peeing “Experiment on Tadpoles. Jour.

Exper, Med., XXIV, 3

. Swingle, W.

1919. Studies on the Relation of Iodin to the Thyroid. I. The Ef-

fects of Feeding Iodin to Normal and Thyroidectomized

Tadpoles. Jour. Exper, Zool., XX VII, 397

1919. II. Comparison of the Thyroid Glands of Todia- we and Nor-

mal Frog Larve. Jour. Exper. Zool., XXVII, ;

1919. Iodine and the Thyroid. III. The Specifie Action ot Todine

in cae te oon Amphibian Metamorphosis. Jour. Gen

siol., I, 593.

1919. IV. _ auntie Experiments on Iodine Feeding and Meta-

osis. Jour. Gen. patie II, 161.

220 THE AMERICAN NATURALIST [Vou. LV

17. rae D.

915. Demonstration in vitro of the Specifie Affinity of Thyroid

Cells for Iodine. Proe. Soc, Exper. Biol. and Med.

XII, 132.

Marine, D., and Feiss, M. O.

1915. The Absorption of Potassium Iodid by Perfused Thyroid

Glands and Some of the Factors Modifying it. Jour.

Pharm, and Exper. Therap., VII, 557

18. Uhlenhuth, E.

1919. Relation between Metamorphosis and Other Developmental

Phenomena in Amphibians. Jour. Gen. Physiol., I, 525.

19. Allen, B.

1918. The Relation of Normal Thyroid Gland Development to Bodily

Growth and Differentiation in Rana, Bufo and Amb ystoma.

Anat. Record, XIV, 26

20. Uhlenhuth,

1919. Relation between pase gy “ti Metamorphosis and Growth.

Jour. Gen. Physiol., I,

21. Allen, B. M.

1918. The Results of Thyroid Removal in the Larvæ of Rana

pipiens. Jour. woe Zool., XXIV, 499.

22. Hoskins, E. R., and Hoskins, M

1919. Growth and seus of Amphibians as Affected by

dectomy. Jour. Exper. Zool., XXIX, 1

23. Allen, B. M.

1916. The Results of the Extirpation of the Anterior Lobe of the

Hypophysis and of the Thyroid of Rana pipiens larve.

Science,

1917. Effects of the Extirpation of the Anterior Lobe of the Hy-

pophysis of Rana pipiens. Biol. Bull., XXXII, 117.

24. Smith, P. E. :

1916. Experimental Ablation of the Hypophysis in the Frog Em-

bryo. Science, N. 8S., XLIV, p. 280.

1916. The Effect of Hypophyeestomy in the Early Embryo upon

the Growth and Development of the Frog. Anat. Record,

XI, 57.

1920. The Pigmentary, Growth, and Endocrine Disturbances Induced

in the Anuran Tadpole by the Early Ablation of the Pars

bucca’is of the Hypophysis. Amer. Anat. Memoirs, Phila-

XI.

delphia,

25. Allen, B. M.

1920. P in the Transplantation of the Hypophysis of

dult Rana pipiens to Tadpoles. Science, N. S., LII, p. 275.

Pee E

1918. The Growth of Normal and Hypophysectomized Tadpoles as

Influenced by Endocrine Diets. Univ. Calif. Publ. in Phy-

siol., V, 11.

27. Hoskins, E. R., and Hoskins, M. M.

1920. The Interrelation of the Thyroid and Hypophysis in the

ee e SEY N Cpu em Endocrinology,

A 1

və

had

iv)

ie)

. Allen, B. M.

. Dickerson, M. C.

1907.

. 6838] INTERNAL SECRETIONS OF AMPHIBIANS 221

. Allen, B. M.

1919. The Relation of the Pituitary and Thyroid Glands of Pea

and Rana to Iodine and Metamorphosis. Biolog.

XXVI 405

. Adler, L.

1914. Metamorphosestudien an Batrachierlarven. I. a gta

endokriner Drüsen. Arch. Entw. Mech., XX XIX,

. Allen, B. M.

1917. Effects of the Extirpation of the Anterior Lobe of the Hy-

pophysis of Rana pipiens. Bio. Bwl., XXXII, 117.

. Rogers, J. B.

1917-18. The Effect of the Extirpation of the Thyroid upon the

Thymus and the Pituitary Glands of Rana pipiens. Jour.

Exper. Zool., XXIV, 589.

. Uhlenhuth, E.

1920. Hapurimiitad py pn Decorah by Feeding es

Gland. Proc. Soe. Exper. Biol. and Med., XVIII, 11

1921. GAPE Prodú a pe Gigantism by Feeding the An-

or Lobe of the Hypophysis. Jour. Gen, Physiol, ITI,

Uhlenhuth,

18 A the Influence ee Development, Metamorphosis and Growth

of ym en Taken as Food Due to a a Ac-

of the ai Jour. Exper. Zool., XXV, 1

1917. jan is Thymus Gland of Mammals „hon RL as Food to

Amphibians Exert any Specifie Influence fot the Organ-

ism? Proc. Soc. Exper. Biol. and Med., (E

1919. Nature of a_i Influence of the <a upon Amphib-

an Metamorphosis. Jour. Gen. Physiol.

1919. Relation between Thyroid pear Melia koni ana Growth.

r. Gen, Phystol., I,

1919, ane Function of the ta a. Endocrinology, III, 285.

Allen, B. M.

1918. The ETEN of Thyroid and Hypophysis Removal upon Gen-

Body Growth and upon the Development of the Limbs

$ rOn and Bufo. Anat. Rec., XIV, 86.

. Cope, E. D.

889. The Batrachia of North America. Washington.,

. Smith, P. E.

1920. The Pigmentary, Growth, and Endocrine Disturbances Induced

in the Anuran Tadpole by the Early Ablation of the Pars

` bueca’is of the Hypophysis. Amer. Anat. Memoirs, Phila-

delphia, XI.

1918-19. Miscellaneous Notes Regarding Experimental Studies upon

the Endocrine Glands of Rana and Bufo. Anat. Rec., XV,

The Frog Book. New York, 205.

CONTRIBUTION TO THE KNOWLEDGE OF THE

NUDIBRANCHIATE MOLLUSK, MELIBE

LEONINA (GOULD)*

PROFESSOR H. P. VON W. KJERSCHOW-AGERSBORG?*

UNIVERSITY OF WYOMING

INTRODUCTION

THe writer’s attention was drawn to this rather

unusual type of molluskan life (Melibe leonina) in ob-

serving the living animals at the Puget Sound Biological

Station, located at Friday Harbor, Washington, during

the summer of 1914, and inasmuch as little was known

concerning the species, an effort was made to assemble

such data as might be of interest relative to its habits

and development. The results of this study are pre-

sented in the following pages.

It was originally intended to publish ices in connec-

tion with work on the morphology of the species, now

under preparation, but owing to the extent of the morpho-

logical data and the crowded condition of the morpho-

logical journals it seems best to have this other matter

appear separately.

TAXONOMY

The genus Melibe (Rang), together with Tethys (Linné),

constitutes the family T'ethymelibide, which forms one of

the numerous groups of family rank included in the

sub-order Nudibranchiata of the opisthobranchiate Mol-

lusca. The type of the genus Melibe was discovered at

the Cape of Good Hope and described by Rang in 1829.

Since that time eleven species have been added by various

* Received by the Editor on May 5, 1920.

1 In my previous writings, Biol. Bul., Vol. 35, No. 4, 1918; School and

Society, Vol. 9, No. 232, 1919; Pub. Puget Sound Biol Sta., 2, No. 49, 1919;

AMERICAN NATURALIST, Vol. 54, 1920, my name is written Kjerskog-... .

which is the modern form of the original (Kjerschow) substituted hence-

222

No. 638] MELIBE LEONINA (GOULD) 223

authors. Gould,1852,deseribed Melibe leonina, the species

upon which this paper is based, from Puget Sound, found-

ing for it the genus Chiorera now merged in Melibe.

Bergh, in a series of papers between 1863 and 1907 revo-

lutionized the classification of the Nudibranchiata. He

divided the nudibranchs into two sections: the Klado-

hepatica and Holohepatica porostomata. Although his

work was primarily systematic, it was based on morpho-

logicalstudies. He added six species to the genus Melibe,

including M. pellucida, from the coast of Washington

near the mouth of the Columbia River. His description

indicates such a close similarity to M. leonina that it may

be questioned whether this species is entitled to specific

rank; material from the type locality may be necessary

to settle this point. The more recent work on taxonomy

by Sir Charles Eliot, 1910, modifies Bergh’s work to

some extent.

No detailed study has been made of M. leonina pre-

vious to the present work and our knowledge of the

species rested largely with the brief description and

figure presented by Gould.

DISTRIBUTION

Melibe, as far as known, is restricted to the Pacific and

Indian Oceans, and to the South Seas. On the eastern

Fic, 1. Melibe (Chioraera) leonina Gould. From a drawing by Mr. Bert Elliot,

private artist to Professor Trevor Kincaid, University o

Washington. MENY changed. x 1/3.

224 THE AMERICAN NATURALIST [ Von. LV

Fic. 2. Photograph, by erga of sperie specimen of M. leonina eet! ing pro-

Pana branching of hepa system in the body-wall; the hood is contracted

a knob as mean the animal swallows. Slightly Ai

side of the Pacific, two species have been described,

M. leonina Gould, 1852, and M. pellucida Bergh, 1904.

On the western side of the Pacific, various species have

deen found by Bergh from the Japanese Island to the

Straits Settlements. One species has been described by

Angas, 1864, from Australia; one from the Pacifie by De

Filippi, 1868; one from the Sandwich Islands by Pease,

1860; and several species from the Indian Ocean by Rang,

1829, Alder and Hancock, 1864, and Eliot,1902. So far no

species of this genus seems to have been found on the

coasts of the Atlantic Ocean. In these parts of the world,

and in the Mediterranean and Caribbean Seas, it seems,

according to works of Bergh, 1877, 1890a, Locard, 1885,

Viguier, 1898, and_ Eliot, 1910, to be replaced by Tethys

Linné. Eliot 1910 (pp. 12-13), fails to record Melibe

among the American Pacific fauna. Neither Tethys nor

Melibe seems to occur in the northern waters. All the

Melibide recorded by Agassiz, 1852, G. O. Sars, 1878,

Johnston, 1838 (Melibea), are, in fact, not Melibe, but

No. 638] MELIBE LEONINA (GOULD) 225

Doto. Von Marten’s (1879) Melibea does not seem to

come under either type, although from this author’s

description certain characteristics are in common with

that of Melibe. For some time there was considerable

confusion in regard to these genera. Bergh, 1863, and

again in 1871, makes this clear when he writes:

Das Geschlecht Doto wurde von Oken (1815) auf der Gemlin’schen

Doris coronata (Bomme) aufgestellt. Jahre nachher (1829) bildete

Rang (Man., p. 129, pl. 3, f. 3) eine neue Geschlechtsform Melibe ab,

dessen Typus eine Nacktschnecke war, die er im Meere des Vorgebirges

der guten Hoffnung (Cap) am schwimmenden Meeresgrase fand. Das

Rang’sche Geschlecht, das von späteren Verfassern gewöhnlich Melibaea

genannt ist, ist am meisten mit dem Oken’schen Doto identificirt worden,

in der Art, dass alle bisher bekannten Melibaeen—eben mit Ausnahme

der “Melibe” rosea von Rang—der Doto-Gruppe gehören. Die Melibaea,

autt. sind in der That mit den Doto’en identisch und mussen diesen

Namen bekommen. Die Meliben Rangs dagegen werden, wie früher von

mir hervorgehoben (Schiödte Naturh. Tidskr., 3 , 3, 1863, p. 480),

einen ganz verschieden eschlechstypus pw den ich als mit den

hioraeren von Gould verwändt betr: einer späteren

t aber ein

Mittheilung von Alder und Hancock (1864) ERO den Seyllaeen

näher (?) k

The genus Tethys Linné, 1758, the nearest relative of

Melibe Rang, 1829, was also confused with other forms,

e.g., Aplysia. Pilsbry, 1895, has, however, cleared up

this point.‘ He shows that the various specific names

attached to Tethys, such as fimbria, fimbriata, and

leporina, are to be considered synonyms. Bergh, 18804,

and Krause, 1885, do not record Melibe m their nudi-

branchiate collections from the north Pacific. Likewise,

investigators in the north Atlantic, on both sides of the

ocean, including a number of explorers and’ independ-

ent workers such as Alder and Hancock, 1842 and 1845;

Meyer and Möbius, 1865; M. Sars, 1870; Aurivillius,

1885; Garstang, 1890; Krause, 1895; Ohdner, 1907;

Walton, 1907; Théel, 1908; and Johnson, 1915, do not

record Tethys or Melibe in their collections. These

species are, therefore, restricted to the warmer seas.

When the Puget Sound Biological Station’ was estab-

2 This station was first known as the Puget Sound Marine Station; its

founder, Professor Trevor Kincaid, remained its director until 1914

226 THE AMERICAN NATURALIST [Von. LV

lished in the San Juan Archipelago in 1904, it was found

that Melibe leonina was not uncommon in the vicinity,

although like many pelagic organisms its abundance was

subject to great fluctuations. In the summer of 1912 it

was particularly abundant, great numbers appearing

among the fronds of Nereocystis drifting past the float-

ing dock in front of the station. At this time Prof. H. L.

Osterud, of the University of Washington, gathered and

preserved a considerable number of specimens. ‘The

largest specimen obtained was six cm. in length, the time

of collecting being the latter part of July. During the

season of 1913 very few were seen.. In 1914, the writer

found several specimens of large size, 8 to 13 em. in

length. These were taken among the floating eel-grass,

Zostera marina. In the summer of 1915 only two speci-

mens were found. It appears that the genus Tethys is

of spasmodic recurrence in the Mediterranean (Viguier,

1898). The appearance of Melibe does not seem to be

determined by any particular season, as Prof. Osterud

found specimens spawning when visiting the Biological

Station early in March of 1916. The period of existence

of this nudibranch must be more than one year, or one

season, judged from the sexual condition in this species

which has shown maturity in individuals ranging in size

from two to fourteen centimeters. Alder and Hancock

(1845), p. 24, think the period of existence for a large

number of species ‘‘not much exceeding one year.’’

EcoLoey

There is a striking similarity between the Klado-

hepatice, not only between the members of any given

family, but also between members of different families

within the section. The former is well illustrated by

Tethys and Melibe, members of the family Tethymelibide.

This similarity is not only morphological, but equally

true as to manner of living and general behavior, for

instance, the method of swimming. M. leonina may

crawl on the leaves of the eel-grass; or it may float with

No. 638] MELIBE LEONINA (GOULD) 227

the back up, the hood having air under it, or the papille

serving as floats. In the latter case Melibe alternately

bends laterally the anterior end to angles of 45 degrees.

During these alternating bendings of the body the move-

ments are swiftest when the body is relaxed from the

45-degree bend. By this method it conveys itself slowly

through the water. However, on the surface tension of

the water or on the eel-grass, the movements are caused

by the ciliary action of the ventral surface of the foot,

because progressive movements occur without visible

bodily contortion. Tethys leporina Linn. moves through

the water by similar means (Bergh, 1877 and 1883). In

its case, however, the large veil, as well as the lateral

bendings of the body, plays an important part. Gould,

1852, in his original description of Melibe leonina, says:

This antmal swims by lateral flexions of the body, the foot being then

folded; and when crawling it is able to flex its enormous head laterally

with considerable force (p. 310

Scyllea pelagica Collingwood, 1879, swims very much

like M. leonina and T. fimbria; and Garstang, 1890, men-

tions that Lomanotus Vérany, swims ‘‘vigorously through

the water in the dish . . . by lashing the body from side

to side. But paither Gould, Bergh, Collingwood, nor

Garstang mentions ciliary action as a factor in locomo-

tion.

Collingwood, 1879, says in part:

Considerable numbers of this pelagic species were found upon Sar-

gassum bacciferum, floating in Lat. 25 N., Long. 37 W., most species

of weed having one or more specimens. The animals were in constant

movements of contracting and writhing. _ In the water they swam freely,

moving the head and tail from side to side alternatingly, so as nearly

to touch one another; and when thus swimming were always, owing to

the weight of the papillary prolongations and tentacles, back downward,

and bore grotesque resemblance to a four-legged animal with ears, such

as a Skye terrier. :

Another similarity between Scyll@a and M. leonina is

the manner of dropping from the surface to deeper

water. In the case of leonina this is a sort of death

feigning. S. pelagica may be found at the surface or in

228 THE AMERICAN NATURALIST [Vou. LV

deeper water. Collingwood mentions it as assuming

certain aspects when it falls through the water to a con-

siderable depth, where it is frequently found. M. leonina

possesses the same habit. In so doing the muscles of

leonina are absolutely relaxed and the animal appears

dead. The means by which the animal gets to the sur-

face are not known, unless it be by its general mode of

swimming. One striking difference between these two

forms is: M. leonina may swim with the back upward,

‘‘the weight of the papillary prolongations,’’ important

in S. pelagica, being apparently of no account.

The eel-grass offers an excellent feeding ground for

Melibe. Here the water must not only be calm, but may

abound in small Crustacea. Zostera, which grows in

large beds in the bays near the Biological Station at

Friday Harbor, offers also a suitable assembling place

for Melibe where it may pair and lay its eggs. At low-tide

the eel-grass floats on the surface of the water and leaves

many inclosures of open water. In these open spaces

M. leonina collect and copulate, as was observed in the

summer of 1914. In such spaces, as described, a con-

siderable number of Melibe had collected, and some of

them were copulating, being united head to head, the foot

of one mate facing the surface. The excellent condition

of the water offered an ample opportunity to study the

mode of swimming and the manner of feeding. The

former has partly been described above, and will further

be discussed under the topic on observations in the labo-

ratory; the latter corresponds to Eliot’s description

(1902) of M. fimbriata Ald. and Hane. M. leonina is not

so definite in its movements during its feeding as is

M. fimbriata, yet some similar method of feeding is pur-

sued. Both species have a large hood. In the case of

leonina the hood is extended very widely (Fig. 1), when

the animal is searching for food, and is periodically con-

tracted into a knob (Fig. 2), when food is obtained.

When the hood is open, leonina tosses it sideways, hold-

ing it in direct position for the capture of small horizon-

No. 638] MELIBE LEONINA (GOULD) 229

tally swimming crustaceans. Eliot, 1902, says of M.

fimbriata:

In spite of its want of jaws, Melibe fimbriata is a most voracious ani-

mal, and I more than once found in the stomach whiċh I examined limbs

of Crustacea more than an inch long. . . . The movements of the animal

are rapid and energetic, whether it crawls or swims. It can float on the

surface, foot uppermost (p. 70).

Melibe leonina is actively predaceous also; its gizzard

has been found completely filled (Agersborg, 1919) with

minute Copepoda, Amphipoda, and larger and smaller

Isopoda, until the gizzard would bulge out into almost a

perfect sphere; ordinarily the gizzard has only a partial

enlargement; its normal size is a little larger than that of

the proventriculus, and the anterior part of the intestine.

Daugherty and Daugherty, 1912 and 1917 (p. 83),

refer to nudibranchs as vegetable feeders; having men-

tioned Eolis and Pleurophyllida, they say:

These soft naked sea-slugs live in shallow water near the shore, erawl-

ing about and feeding upon the sea-weeds.

However, only a few of the rarer species are phytivorous ;

the majority are carnivorous, a fact which is recognized

by the authorities on nudibranchiate Mollusca. Thus

Bergh, in most of his descriptions of nudibranchiate

fauna, reports in favor of animal diet: 1880a (Akiodoris

lutescens) p. 56, (L llidoris bilamellata) p. 64, (L.

luptricina,Acanthodoris pilosa) p.101,(Triopha modesta)

p. 116; the food of these forms consisted of ‘‘indeter-

minable animal matter, mixed with some diatomacea,

. and with some Polytholamia, . . . with larger and

smaller pieces of small Crustacea, . . . and a little inde-

terminable worm, of the length of 2.0 mm., . . . spongiary

masses and different Radiolarie of a diameter of 0.09

mm.” In regard to Tritonia reticulata, the same author,

1881, says:

Die Tritoniaden sind Raubthieren und scheinen sich hauptsachlich von

Aleyonien und ähnlichen Thierformen zu ernähren.

Again, 1883, referring to Tethys leporina, he says:

230 THE AMERICAN NATURALIST [Vot. LV

Tethys ist ein Raubthier und sein Nahrung besteht namentlich aus

kleinen Ophiuren, deren Reste oft ganz den Magen erfiillen.

1890a:

Der Magen und der Darmen von Nahrung vollgeslofft ; dieselbe bestand

us Massen von kleinen Decapoden, mit Bruchstiicken von kleineren

Gasteropod Schalen und Sandkornern vermischt (p.

1894 (Dendronotus robustus) :

In der verdauungshdle unbestimmbare thierische Masse, mit

Diatomeen vermiseht (p. 144

And finally, describing the food of Melibe rosea, 1907,

he says:

The contents of the alimentary cavity (specimen 1.5 em.-3.8 em.)

were animal matter with remains of small Hydroids (p. 98).

Alder and Hancock, 1845, p. 23, say:

But, though so patient and long-suffering in the endurance of hunger,

these little animals are very voracious. The greater number of them are

carnivorous; living principally upon zoophytes and sponges. The

Alcyonium digitatum is a favorite food with the Tritonie; and the

Actinie and Lucenarie often fall prey to the attacks of the Eolides.

These latter, indeed, do not seruple occasionally to devour the weaker

among their own brethren, as we have recorded elsewhere. Sir J. G.

Dalyell states that his Kolis histrix (Drummondi) ‘fed voraciously on

mussel, and on common periwinkle, whereof’ Ped portions were swal-

lowed entire’; and he thinks that Goniodoris nodosa feeds upon Ascidia

papilla (Cynthia rustica), to which he attributes the reddish colour ob-

served in the viscera. This colour, however, is caused by the liver and

ovary. We have taken from the stomach of Eolis papillosa minute

Tethys. The more common food of the tribe, however, is the flexible

zoophytes. Until lately the Dorides have been considered vegetable

feeders, but this would appear not to be the case. Doris tuberculata

feeds upon common enerusting sponges (Halichondria panicea), and

sponges and zoophytes seem to constitute the food of most of the

others. A few of the gregarious Nudibranchs, such as Polycera quadri-

lineata, Hermea dendritica, and Alderia modesta, which congregate on

marine algae, appear to be phytivorous; but Eolis despecta, and E.

exigua though not unfrequently gregarious on the fronds of Laminaria

digitata, are only found on those parts of the plants that are covered

with the parasitie zoophytes, Laomedea geniculata and L. gelatinosa, on

which they feed and deposit their spawn. —

Meyer and Möbius, 1865, are of the opinion that it is

No. 638] MELIBE LEONINA (GOULD) 231

rather difficult to determine whether Nudibranchs are

carnivorous or phytivorous, that is, that the food of

Nudibranchs is very variable. In part they say of Elysia

viridis Montagu, p. 10:

Sie nahrt sich wahrscheinlich von Pflanzen.

Page 23:

Eolis alba frisst, wie die anderen Kieler Arten ihrer Gattung,

thierische Stoffe.

Page 31 (Eolis papillosa) :

Thre Nahrung sind Thierstoffe; besonders liebt sie Actinien. Kleinere-

Exemplare der Actinia plumosa greift sie am Fussrande an, und frisst

ein halbmondformiges Loch hinein, das sie immer mehr vergréssert.

This is also the opinion of Hecht, 1895, p. 621:

Il west done pas possible d’établir à ce point de vue une division bien

tranchée. Ou peut dire seulement, que les familles les plus franchement

herbivores sont les Hermaeidae et les Elysiidae, et en général les Asco-

glosses qui, comme Ihering et d’autres lont remarqué, ont une masse

buccale disposée pour exercer une succion. .. .

Page 622:

Les Eolidiens sont tous franchement carnivores et présentent, parmi

les grandes espèces, quelques types d’une voracité extraordinaire. Eolis

coronata . . . devore des Elysia viridis; à l'autopsie j’ai trouvé des

radulas dans son tube digestif. Les petites espèces Eolis despecta, E.

exiqua, E. olivacea peu faites pour de grands déplacements, vivent à

demeure, comme je l'ai dit plus haut, sur des colonies d’ Hydroides.

Page 223:

Calma glaucoides, qui, pendant une période de sa vie tout ou moins,

se nourrit, je l’ai dit, d’embryons de Poissons. Le régime des Doridiens

est moins uniforme; certains genres sont probablement herbivores. Plu-

sieurs espéces de Doris se nourrissent d’Eponges ealeaires dont on re-

trouve les spicules dans les excreta. . . . Plusieurs espèces de Goniodoris

se nourissent de Bryozoaires. I] est probable que Polycera quadrilineata

mange des Algues. Il faut signaler ici les observations de Prouho, sur

la façon particulière dont Idalia elegans se nourrit de certaines Ascidies.

Quant aux Ascoglosses, j’ai indiqué plus haut que Hermaea dentritica

dévore les couches superficielles des Codium tomentosum, qu’elle réduit

à l'état d'un petit moignon verdâtre. Elysia viridis se nourrit aussi de

Codium tomentosum, mais sans marquer de préférence pour telle ou telle

région; j’ai du reste observé qu’elle s’accommode aussi d’autres Algues.

232 THE AMERICAN NATURALIST [Vou. LV

Jeffreys, 1869, contributes to this subject and says:

Although most of the order are zoophagous, Limapontia and others of

a simpler kind feed on seaweeds.

And von Ihering, 1876, p. 37, referring to Tethys states:

Das der Magen eines so gefriissigen, jeder Bewaffnung des Mundes

baaren Raubthieres wie Tethys eines solehen Schutzes ganz besonders

bedarf, wird sofort verstiindlich, wenn man den Mageninhalt desselben

noch die Otalithen eines andren Fisches, welche diejenigen des eben-

bezeichneten um das Doppelte iibertrafen.

M elibe fimbriata Ald. and Hanc., is, according to Eliot,

1902,

in spite of its want of jaws, a most voracions animal.

This same author says he more than once found in the

stomach he examined limbs of Crustacea more than an

inch long. And, in 1910:

Thus the red British Dorids Rostanza coccinea and D. flammea eat

red sponges, such as Microciona atrasanguinea (p. 5).

In fact, this author thinks that most Dorids feed on

sponges (Eliot, 1910, p. 39). Step, 1901, referring to the

crowned sea-nymph, Doto coronata, says it feeds upon

Hydroids (Sertularia and Plumularia) and Corrallines

(Antennularia antennina). The marble slug, Lomanotus

marmoratus, feeds upon corrallines which it closely re-

sembles in color and ornamentation. Eolis feeds upon

anemones, Sagertia, Lucernari ia; Sea-mats, Tubularia,

various sponges and Obelia. The crimson Hermæa

(Ascoglossa), Hermea bifida, feeds on small crimson

weeds (Bryopsis, Codium, Enteromorpha and Ulwa).

Vayssiere, 1901, p. 84, referring to Tethys fimbria,

Bohascht, Delle Chiaje (Synon. T. leporina Linné,

Cuvier), writes:

Dans l'intérieur du premier renflement stomacal (jabot), je trouvais

d'ordinaire une grande quantité de filaments fibreux de Zostéres; ces

mollusques doivent en aspirant avec leur trompe, absorber des débris de

No. 638] MELIBE LEONINA (GOULD) 233

ces végétaux et dissocier leurs fibres par les contractions répétées des

parois musculaires de cette poche.

Page 85:

Au milieu ies ces débris, ils onc de petits crustacés (Entomostracés,

Amphipodes, Isopodes, jeunes Décapodes brachyures) et autres petits

Invertébrés, adian parmi les Zostères, qui doivent former la base

de leur nourriture.

Thus it is seen that even the carnivorous Tethys may be

phytivorous. Vayssiere, Da, p. 43, says of Halgerda

willeyi C. Eliot, 1903:

La poche stomacale était remplie de gros débris Crayeux constitués

par des fragments de madrépores et de bryozoaires que ‘ce mollusque

arrache et broie avee sa forte radula.

Finally, MacFarland, 1912, p. 530, says in regard to the

Dironidæ (Diron allolineata) :

Diatom shells and minute spicules, these made up a very small ‘por-

tion of the total contents.

Nudibranchs may be said to be omnivorous; as seen

above, a species which is void of radula, e.g., Tethys,

may at one time be carnivorous, at another time phytiv-

orous; likewise forms possessing radula (Acanthodoris)

are omnivorous. The largest number seem to be carniv-

orous notwithstanding; a few, the Hermeide and

Elysvide, are phytivorous.

Means oF DEFENSE

Upon being first encountered, Melibe leonina appears

brown, but when examined in the aquarium one can.

easily see that the brown coloring is rather superficial in

comparison to the marked transparency which the body

possesses. This transparency is so great that the inter-

nal organs, such as the alimentary canal, the organs of

reproduction, and the heart, can be easily seen. The

possibility of actually seeing these organs through the

body wall is due to the arrangement of the muscles and

the connective tissues, and because the body-fluid con-

tained in the perivisceral cavity, between the visceral

234 THE AMERICAN NATURALIST [Vou. LV

organs and the muscles, the connective tissues, as well as

the blood, are colorless. ‘This characteristic is also

common to M. pellucida Bergh, 1904; and M. vexulifera

Bergh, 1880.

Upon touching the curious-looking animal it gives off

a peculiar odor. This is rather strong, and resembles

that of oil of bergamot. It is caused by a secretion from

small compound saccular glands lying immediately under

the ectoderm (Fig. 12). These glands are distributed all

over in the external parts of the body: in the body wall,

the papille, under the ectoderm of the exterior part of

the hood, and in some cases, under the ectoderm of the

foot. None of these glands seem to be present under the

ectoderm of the ventral side of the hood. The extent of

distribution of the odoriferous glands seems to indicate

that they have a definite use and purpose, e.g., that of

defense.

Meckelii, 1838, describes the odor exuded by Tethys

leporina as resembling citron, or being rather pleasant.

And Bergh, 1877, says:

Von ... toten Thiere habe ich irgend eine Spur bemerkt, dagegen

einen nicht starken, etwas besonderen, aber nicht wesentlich unbehab-

lichen Gestank.

Hecht, 1895, discusses the various means of defense pos-

sessed by Nudibranchs. He mentions protective colora-

tion, nematocysts, mucous glands and death feigning.

The last will be discussed presently; the first two may

only be referred to, as M. leonina possesses no protective

coloration, and has no nematocysts as a defensive means.

It may, however, be stated that in the aquarium, Melibe

leonina became even more transparent than it was when

seen in its natural environment. This change of color

was also observed by Alder and Hancock, 1845, on a num-

ber of Nudibranchs kept in captivity: ‘‘ In such cases

they generally lose a good deal of colour and become

very transparent,’’ and that coloration is not caused by

the color of the food taken, but by the color of the liver

and gonads. Eliot, 1910, says:

No. 638] MELIBE LEONINA (GOULD) 235

The colour of Dorids is to some extent affeeted by their food, though

less than that of Eolids. The brightly coloured species often frequent

and feed on similarly bright sponges or Ascidians, and when they do not

obtain their usual food in confinement they lose their colour (p. 5).

An increased transparency when kept in an aquarium,

€.g.,a glass-jar,may be designated adaptive coloration.

The odorous substance that the animal exudes when

touched by an enemy is its main protection. However,

Step, 1901, records some very interesting facts relative

to protective coloration among Nudibranchs:

Dendronotus frondosus is obviously adapted to life among sea-weeds

and coralline, resembling some small red-brown sea-weeds (Collitham-

nium). It is said to be highly edible, having nothing in its flavor to dis-

please the taste of the most fussy fish; and therefore its disguise is

absolutely necessary to the species (p. 288-289).

Evidently Melibe leonina needs no protective coloration,

having, as said above, odoriferous glands. One reason

why it does not possess nematocysts is perhaps because

it does not live on Hydroids, but mostly on Crustacea.

Glaser, 1903, in his discussion of the origin of nemato-

cysts in Nudibranchs gives a historical review, including

a number of citations all of which refer to the nemato-

cysts as having been taken in with food. The food then

seems to be the origin of the nematocysts in the Nudi-

branchs; M. leonina, which does not feed on hydroids, has

no nematocysts in its system. The marble slug, Loma-

notus marmoratus, according to Step, ‘‘Feeds upon

Corallines pup it closely resembles in colour and orna-

mentations ’’; Gamble, 1892, however, says that enido-

cysts are ca

Bergh, 1890a, mentions experiments by Krugenberg,

1880, who tried to determine the physico-chemical con-

stituents of the odoriferous glands, as well as of the liver

and blood of Tethys fimbriata (s. leporina), and says

that T. leporina has a peculiar musk-like nauseous odor

which it uses as a means of defense against its enemies.

This is, without doubt, the office of the odor in the case of

M. leonina also. The actual nature or constituents of

236 THE AMERICAN NATURALIST [Vou. LV

the substance which causes this defensive odor have not

been determined.

Another means of protection is self-mutilation, exem-

plified by Discodoris fragilis, that according to Eliot,

1899, throws off part or the whole of its mantle edge.

Collingwood, 1868, also records this habit of self-

mutilation of a Doris.

EMBRYOLOGY

The Egg-body (Nidosome)

Bergh, 1902, describing the egg-body of Melibe bu-

cephala says:

The spawn forms a large heap of a diameter of 3.5 cm., composed of

the innumerable windings of a dull yellow tube of a diameter 0.75 mm.

The tube contains inside of the tough transparent covering several

series of displaced, more or less cleft eggs.

From this description it is clear that this nidosome is

quite different from that of M. leonina; it seems, indeed,

strange that the egg-body of two closely related species

can differ so widely. The external features of M. bu-

cephala, according to Bergh’s description, are not much

different from those of M.leonina. It may be a question

whether the egg-body attributed to M. bucephala, by

Bergh, actually belongs to this species.

The writer during the summer of 1914 found several

nidosomes among the eel-grass, but it was not known to

which animal they belonged until Melibe leonina was

seen to lay the same kind in the aquarium of the labora-

tory (Agersborg, 1919). These nidosomes of transparent

mucous, or gelatinous substance, were funnel-shaped,

when suspended in the water, with the apex attached to

some solid object (Fig. 3). The average slant-height of

these conical structures was 5 em., with a perimeter of

about 28.2 em. and a convex surface, therefore, of about

70.50 sq. em. From the adhering point of the nidosome,

dotted lines, the capsules radiated to the periphery of

the conical body. This radiation was not so regular in

band at right angle to the fold is a g

No. 638] MELIBE LEONINA (GOULD) 237

some as in others, yet there was a prevailing regularity

in this respect. In Fig. 3, the arrangement of the cap-

sules does not represent the prevailing regularity, as it

was necessary to select

an entire nidosome for photo-

graphing

most of the other having been broken. The

capsules contained from 10 to 22 eggs (Figs. 4-6). The

actual method of deposition has not been observed, but

yet

met

e ty

me Pat

Leal" teres *

8. Photograph of a | nidosome of M. leonina, natural size. The dark

yiece of eel-grass to which the egg-body is

attached. hues: many one white dots are "eae canals containing from

15-22 egg

it is conjectured from the knowledge of the anatomy of

the animal that the capsules are imbedded in the gela-

tinous mass as the nidosome is deposited. The mucous

gland, which consists of (1) albuminous gland, (2) nida-

mental gland (Lang, 1896), is in Melibe in direct connec-

238 THE AMERICAN NATURALIST [Vou. LV

Fig. ys

Fics. 4-7. Drawing of egg-capsules from a nidosome sie in apen show-

ing a Aaii number of eggs. Figs. 5—6 show the eggs in t and four-cell stage,

three hours after the nétosome was laid. In Fig. 7 clea ve has reached the

here di stage. The more oblong embryos are moving within the capsule

tion with the vaginal orifice. In copulation, the penis,

which is long, twisted like a screw, and of tough muscu-

lature, is inserted into the posterior genital pore of the

mate, and so firm is the union that separation may not

occur even though the couple be dipped from their natural

abode and placed in a vessel.

Observations in the Laboratory

A study of Melibe leonina from an embryological stand-

point, was made at the Biological Station. From the lot

collected, some were preserved, others were kept alive in

an aquarium. One morning, however, all save one were

No. 638] MELIBE LEONINA (GOULD) 239

dead. Later, this one also seemed dead, and it was

thought that the water had become stale. Melibe lay

absolutely motionless on the bottom, all its muscles com-

pletely relaxed, and showed signs of life only after the

water had been oxygenated for several minutes. It is

muscular relaxation of this sort that Melibe assumes

when it sinks from the surface to deeper water. After

this the writer became used to its death feigning and

needed only to oxygenate the water for it to become active

again, crawl along the bottom and side of the aquarium,

and after a while start swimming in the vessel. Changing

of the water could not be done indefinitely, as on another

morning a nidosome was found to have been deposited

by the animal during the night. It was a funnel-shaped,

transparent, gelatinous body adhering by the tapering

end to the side of the jar. Viguier, 1898, when trying to

prepare a specimen of Tethys fimbriata for fixation, ob-

served the same phenomenon. He, however, did not

change or oxygenate the water; he left the animal in it

for about two weeks, when he found that an egg-body had

been deposited, the eggs having ceased to divide in the

four-cell stage. The nidosome of T. fimbriata is quite

different in shape from that of M. leonina. In the case

of Melibe the eggs continued to divide until they were

transformed into larval forms, which actually turned the

whole nidosome into a vibrating mass. The development

continued, apparently normally, until the larve left the

capsules, when they soon died. That the delicate mol-

luskan young should die when coming in direct contact

with the water of the aquarium was expected, as the

renewing of the water was stopped after the nidosome

was deposited, it being thought undesirable to disturb it

too much; after the deposition of the egg-mass the water

was simply kept at constant level, and oxygenated from

day to day, so that the animal should not die. When the

young Melibes were hatched the mother animal, without

being further inseminated, laid another nidosome, which

also hatched two weeks later. The eggs of the first nido-

240 THE AMERICAN NATURALIST [Vor. LV

some developed into distinctly living creatures, moving

about in the capsule on the fifth day after the setting

(Fig. 7); it took two weeks for the complete development

of the young. Alder and Hancoek, 1845, p. 25, say:

The embryo matures after deposition of spawn, from a few days to

a month or more, according to species; the actual time appears to be

about ten days or a fortnight.

Temperature, no doubt, plays an important part in the

speed of development. Stuart, 1865, says:

Gewöhnlich wird von den Opisthobranchiereiern angegeben, dass die

Dauer der Entwickelung des Embryo ein Monat ist, in meinem Falle war

sie circa zwei Monate; dabei war die vorherrschende Witterung, die fiir

Sicilien jedenfalls eine kalte zu nennen war, gewiss von grossem

Einflusse (p. 96).

The second nidosome was to some extent abnormal, com-

pared with the first, and with those collected from the eel-

grass. It showed a variation of the number of eggs in

the capsules, from one to fifteen (Figs. 8-9). This abnor-

mality was perhaps an indication of the decline in vitality

of the mother animal. In fact, the adult specimen had

greatly decreased in size since its capture.

One diffculty was that of keeping the water at a con-

stant density. In order not to break the nidosome the

water was only oxygenated and kept at a constant level.

Each time when the eggs were examined, small portions

of the nidosome were removed, and by so doing the mem-

brane of the egg-body was broken. This did not seem to

affect the development, however. Yet it was thought

safer to keep the water at the same temperature as

hitherto, than to change it daily, as the latter might cause

too great physical shock. The abnormality of the water,

as said, did not affect the embryos as long as they were

within the capsules of the nidosome. Perhaps thus far

in their development they were not affected by the abnor-

mality of the water; even though the egg-body was punc-

tured and broken in the examination of the eggs, the

embryos seemed all to develop, as far as could be detected,

No. 638] MELIBE LEONINA (GOULD) 241

and to pass through the normal developmental changes

of a typical gasteropod.

Ic. 8. Three abnormal egg-capsules shiv the oap EER PEE after

insemination; a, nd. no cleavage. pb, r body; b, pb. first polar body has

divided, but no cleavage in the egg; a no 5 lavas os egg although the polar

bodies were give

Fic. 9. An average sized ae ser the second setting ajag henna ee aa

n, — ganglion; ft, ; pb, polar body; sh, sh

m; all the poe are im the veliger erel

Abnormality in the second nidosome was marked by

the reduced number of eggs within the capsules and in

the early development in the eggs. The variation of the

number of eggs in normal capsules was most marked at

242 THE AMERICAN NATURALIST [Vor. LV

the end of the egg-belt. This abnormality was much

greater in the second nidosome, in that there was a much

greater number of small capsules with only one or a few

eggs inthem. There were also capsules actually without

eggs. Some eggs failed to develop; some gave off the

polar bodies and then did not advance any farther; others

did not form polar bodies. It was of great interest, in-

deed, to watch the development of the embryos in the

large capsules which in some cases contained more than

twenty eggs (Fig. 7). Within a few hours, there would

be twice or even three times as many polar bodies as eggs

(of course within the capsules), because the first polar

body sometimes divided. A detailed study of the blasto-

meres was not undertaken. Figures 4 to 6 show a few

early developmental stages, and early and late larval

stages are shown in Figs. 7 and 9. An embryonic shell

is shown in Fig. 10; veliger larve in Fig. 9. :

When the embryos reached the gastrula stage they

swam about within the capsules. On the fifth day after

being laid the whole egg-body was practically alive with

imprisoned swimming larve (Fig. 9). Nine days later,

the larve began to leave the capsules. It was surprising

to see how rapidly the embryos advanced from day to

day, going through the trochophore and veliger stages.

During the latter stage the shell was very prominent; it

resembled the shell of Natica russa, in that it had a blunt

apex and short body; the posterior or tapering part of

the shell had no spiral turns; the posterior edge of the

aperture had a small indent; the edge of the aperture’

of the shell was otherwise without any modifications

(Fig. 10). The animal itself did not assume the adult

_ Shape before it lost the shell, but when it left the capsule

it shed the shell, and the young began the life of a

so-called naked mollusk. The presence of a shell with

operculum in embryonic life of Nudibranchs has been ob-

served by various authors: Alder and Hancock, 1845;

Pelseneer, 1893; Smith, Bell and Kirkpatrick, 1905; Boas,

1916, and others.

No. 638] MELIBE LEONINA (GOULD) 243

A large number of eggs were present in each capsule

of the first nidosome (Figs. 4-7), and all developed into

Fic. 10. Embryonic shell of Melibe leonina.

embryos which finally went through metamorphosis. In

the second nidosome, some of the eggs failed to develop.

The cause was perhaps lack of spermatozoa. During

copulation, the spermatozoa become stored in the sperma-

totheca but also wander up the uterus as far as the pros-

tate gland (Fig. 11, spt., ovd., pr.). The eggs are prob-

ably fertilized while passing down the uterus, ut., or

while in the spermatotheca, spt., as many eggs actually

pass into this out-pocketing of the uterus, the spermato-

theca may therefore be termed ovo-spermatotheca. The

only means of regulating the fertilization process in the

egg must be the physical condition of the egg, which de-

termines the reception or the function of the sperm; as

all eggs, under natural conditions, contained in one and

the same capsule, and, indeed, in the entire nidosome, go

through simultaneous development, although they all

(perhaps more than 100,000, in one normal deposit, nido-

some) can not possibly have been fertilized at the same

moment. This primary part of fertilization must take

place before the eggs are incapsulated. If, however, an

insufficient number of spermatozoa are present during the

flow of the eggs, some eggs may become incapsulated

without having been fertilized. One thing noted was

244 THE AMERICAN NATURALIST [Vou. LV

that the eggs in one end of the nidosomic belt, and in the

main part of it, were all fertilized, while the other end

(the last end) of the belt showed lack of fertilization.

Another fact noted is that Melibe leonina carries over

spermatozoa in its genital vessels; that more than one

egg-body is deposited after insemination.

pericardium; Pr, prostate; Si, small part of the intestine; Sto, stomach; Ut,

uterus ; Vd, vas deferens ; Le, branch of liver; Spt, ovo-spermatotheca.

Reid’s observations, 1846, on Doris bilamellata, D.

tuberculata, Gonidoris barvicensis, Polycera quadri-

lineata, Dendronotus arborescens, Doto coronata, and a

species of Eolis, bring out the same fact, viz., more than

one deposit takes place after insemination. From 26

hours after coitus deposition may begin; ‘‘ it does not,

however, appear to be absolutely necessary for the

production of fertile ova in all, if in any of the individuals

of the nudibranchiate Mollusea, that coitus should have

so shortly preceded spawning as was observed in Poly-

cera, for an Eolis which was kept strictly confined in a

vessel by itself, deposited, on the tenth and again on the

thirty-second day of its isolation, abundance of fertile

ova.’’ Crozier, 1919, claims that the larger animals of

Chromodoris zebra Heilprin, lay several more egg-masses

in a given time than do small ones; that it is conse-

quently of advantage to the species that large individuals

No. 638] MELIBE LEONINA (GOULD) 245

should mate together; that there is, in fact, selective pair-

ing which is of a distinctly advantageous or ‘‘purposeful’’

character, since it makes for the multiplication of the

species. The same author, 1917a, records the unique find-

ings relative to a rather high degree of correlation

between the sizes of the two pairing members of Chromo-

doris zebra. The writer has observed the same fact rela-

tive to copulation among the Eolidæ. As for Melibe

Fic. Micro-photoegraph of the inner side of the ectoderm of the body-wall,

baat the odoriferous glands, the small dots among the bré Popua

(black) hepatic cæca, and the crossing muscle (pale) fibe

leonina, it is also true that those found copulating were

of the same relative size. But even so, as seen in the

second nidosome of Melibe, the animal may run short of

spermatozoa during ova-deposition. To guard against

this, there is, as shown by Crozier, among certain species,

selective pairing between individuals of nearly the same

size. Garstang, 1890, finds a considerable variation be-

tween the offspring of Lomanotus Vérany, ‘‘the indi-

viduals apparently showing a tendency to unite rather

with those of their own variety than with those unlike

themselves.’’

246 THE AMERICAN NATURALIST [Vou. LV

To understand the process of insemination of the eggs,

and the conditions controlling the number of eggs in the

capsule, a little speculation is necessary. It seems as if

the processes of fertilization and ineapsulation are ef-

fected during the emission of the eggs; that when the

eggs pass down the uterus or pass the spermatotheca

they are fertilized, and immediately after that incapsu-

lated, and that the size of the capsules and the number

of eggs present in each capsule are regulated by the speed

of the outflow of the eggs. The size of the capsule, as a

rule, varies according to the number of the eggs present

within the capsule. It seems puzzling, however, when

capsules are found without eggs, and with eggs which

show no indication of being fertilized, but this abnor-

mality is limited to the last part of the nidosomic belt,

and is of course so controlled that an entire egg-body

may not be deposited without some eggs at least being

present, and being fertilized. It is a matter for future

observation to determine whether individuals in ova-

maturity are capable of depositing normal nidosomes,

without being stimulated by an individual in ripe male-

phase. It is a question whether the mere pressure of

ripening eggs will cause egg-flow. Crozier, 1917b, re-

ports, however, that Chromodoris zebra, if left alone,

deposits fragments of egg-bodies which are not fertilized.

The writer has noted the same phenomenon relative to

Eolis olivacea when it is kept alone; but also in this case,

as in the case of Chromodoris, no normal nidosomes were

deposited.

The question of cross-fertilization becomes of interest

since spermatozoa are found in both genital ducts of the

same individual, from the ampulla of the penis and all the

way down the penis to the end of it; from and including

the prostate which surrounds the uterus, to and including

the ovo-spermatotheca. If self-fertilization takes place,

should there be any shortage of spermatozoa during ova-

deposition? The presence of spermatozoa in the female

genital tube is undoubtedly the result of coition and the

No. 638] MELIBE LEONINA (GOULD) 247

wandering of the spermatozoa up the uterus, against the

outward current of that organ. Alder and Hancock,

om p. 20, say:

e Nudibranchs, notwithstanding that they are androgynous, fre-

ane copulate during the breeding season. The conjoined indi-

viduals lie side by side, their heads turned in opposite directions. Thus

the right sides of the two animals are brought in close contact, and

mutual impregnation is effected. They remain in this position for some

ime, but in a short period after separating, generally about the first or

second day, the spawn is deposited.

Crozier, 1919, claims that Chromodoris zebra is function-

ally b cciaphvodita and effective reciprocal insemina-

tion is practised. But this is not practised among the

species Melibe leonina; although semen may be present

at the same time in both genital ducts, insemination is

not reciprocated simultaneously. That is, in all the indi-

viduals examined, coitus was effected by the introduction

of the penis of the one mate; the penis of the other mate

was completely withdrawn. Whether spermatozoa were

present in the members whose external genital organs

were not visible was not determined. It looks, however,

as if Melibe leonina is protandric, a condition, according

to Pelseneer, 1895, common among Eolis, Elysia, and

Clione limocina. Eliot, 1910, says:

Pairing, according to Hecht, is reciprocal, and though hermaphrodite

Mollusea are incapable of self-impregnation both individuals spawn

after mating.

_The writer has observed on Eolis olivacea, at Woods

Hole, that one mate may start spawning while copulating.

Spermatozoa, according to Reid, may be carried in the

female genitals (Eolis) for more than thirty days before

being used. That is, cleavage does not start in the eggs of

Eolis until after deposition; fertilization, therefore,

may not occur before the time of incapsulation. Sperma-

tozoa are kept alive in, and stimulated by, secretions of

the female genital organs, as shown by Eliot and Evans,’

1908, p. 287:

248 THE AMERICAN NATURALIST [Voi LV

The walls of the spermatotheea (of Doridoides gardineri) are thick

and produce a secretion. In some specimens small clumps of sper-

matoza are embedded in this secretion. In others all the spermatozoa

form a central mass in the main cavity of the spermatotheca. It i

possible that the secretion serves to form small packets of spermatozoa

or spermatophores.

SuMMARY

1. Melibe leonina is a large carnivorous Nudibranch

reaching sometimes 14 centimeters in total length; it is

an actively predaceous animal; it practically gorges

itself, feeding mainly on small Crustacea; it is gregarious.

2. It seems to live more than one year; its recurrence

is spasmodic.

3. It swims freely in the water, backward, upward or

downward; it crawls on the surface by the surface ten-

sion, and on sea-weeds, by the help of its highly ciliated

foot.

4. Its defensive means are an offensive odor and death

feigning.

5. It drops to deeper water by relaxation of its

muscles.

6. It collects in groups among sea-weeds, where copu-

lation takes place.

utual insemination does not seem to be simul-

taneous.

8. It spawns as early as March and as late as July;

sexual maturity is reached quite early, as young ones

two centimeters long were found with ripe spermatozoa.

9. Spermatozoa from another individual are stored in

the ovo-spermatotheeca but wander up the uterus as far

as the prostate.

10. Eggs are also stored in the spermatothaca; hence

the name ovo-spermatotheca.

11. Copulating individuals are of the same relative size.

he same individual deposits more than one nido-

some, after insemination; spermatozoa may be carried

over in the ovo-spermatotheca at least two weeks.

No. 638] MELIBE LEONINA (GOULD) 249

13. The eggs are deposited in capsules, normally con-

taining from 15 to 22 eggs. The capsules are arranged

in rows within a gelatinous mass, sometimes quite regu-

larly; the gelatinous mass is formed into a belt from 3 to

5 em. wide; the mucous flow is greater in one side of the

belt than in the other, so that one side of the belt is shorter

than the other, and the belt curves into a funnel-shaped

mass, the apex adhering to some sea-weed, near the sur-

face of the water.

14. Eggs may become incapsulated without being fer-

tilized; no cleavage of such eggs follows.

15. Normally the embryo develops within two weeks.

ACKNOWLEDGMENT

The writer is deeply grateful to the librarian, Dr. R.

W. Tower, of the American Museum of Natural History,

and his staff, for ever-ready assistance, relative to the

use of the excellent Library of the Museum; to Professor

Trevor Kineaid of the University of Washington for kind-

ness and favors while at the Puget Sound Biological

Station.

BIBLIOGRAPHY

Agassiz, L

851. Report, Proe. Bost. Soc. Nat. Hist., 3.

Agersborg, H. P. Kjerskog

19 eee on Merite leonina, Gould. Pub. Puget Sound Biol, Sta.,

No.

IF J., a Tag

1842. Description Z Several New Species of Nudibranchous ip

w AE the Coast of Northumberland. Ann

Nat. es

1845. A Monograph of a British Nudibranchiate Mollusca. The Ray

, Londo

1864. re, Ald. snd Hanc., Notice of a Collection of Nudibranch-

te Mollusca Made in Tadia, by Walter Elliot, Esq., with de-

E of several genera and species. Trans. Zool. Soc.,

London, 5.

Angas, G. F.

1864. pur australis. PER d’espèces nouvelles appartenant à

rs genres mollusques nudibranches d de

Port pe (tyeri -Galles du Sud) accompagnée de des-

sins faits d’après nature. Jour. de Conch., 3 sér. T. 4, pp.

62-63.

250

Aurivillius,

THE AMERICAN NATURALIST [Vor. LV

1885. tase öfver de af Vega-expeditionen insamlade Arktiska

Bergh, R.

1863

1871.

18800.

1881.

1883.

18844.

1884b.

1888.

1890a,

1890b.

1892a,

18925.

1894.

1898.

1902,

1904,

Hafmollusken, II, Placophora och Gastropoda. Vega Exped.

ts a aa en T PRE, a Dendronotidernes gruppe,

bem inger familie. Prof J. ©.

seat "Nature danse: 3 Række, 1 bind, 3 hefte.

Kjébe

Doto, NEN ; Seyen poden Beiträge zur Kenntniss der Mol-

lusken fies Sar ere: Sie zool.-bot. Gesell. in Wien,

Bd. 21, Seit. 1875-1876, 1288,

C. Semper, Reisen im Avehined a pra en If, 2, Heft 9.

Notizen iiber Tethys leporina, Linn, JB. mal, Gesell., 4, p. 335.

On the Nudibranchiate Gasteropod Meineke of the ‘North Pa-

ee Ocean, with Special eer e to Those of Alaska, IT.

c. Acad, Nat, Sci, Phila.,

sien zur Kenntniss ra Japanischen yE I

(Melibe vesilifera). Verh, zool. bot, Gese

Beiträge Kenn der es nischen s ay

(Tritonia etioata), ji ibid.

Beitrage zur a ie One. nies VII (Tethys leporina

Linn.). Ibid

Beiträge zur Kenntniss e rarai Melibe, Rang or

papillosa). Z. s Zoo

Report on the Nadine. Zool, Challeng.

Beit r Kenn T ANEN IX et ocellata).

räge zu niss de

Verh. zool.-bot. Dai Wien,

On the Anatomy of Tethys ce L. (var.), in the Report on

the Result of Dredging in the Gulf of Mexico, and in =

Sea (1879-1880; 1887-1888). Bull. Mus. C.

Die Nudibranchien des ‘‘Sunda-Meeres.’’ C. Semper, Reisen

C. Semper, Reisen im ESN ipel der Philippinen II, Wiss

schaftliche Resultate, III, Malacologische Untersu A

Heft 15-18 (1880- 199

Die Nudibranchiata holehepatica E Verh. zool.-

bot. Gesell. Wien, 42 (Abh

Report on the Dredging Operations off the West Coast of Amer-

ica. Bull. Mus. C. Z. Harvard College, 25.

Ergebnisse einer zoologischen Fotiuhiangntclen in den Molukken

und neo. SARTE Abh. Senckenb. naturf.

Gesell., Heft I,

Den Danske Siar ion til Siam I (1899-1900), or

branchiata (Melibe pedut D. kgl. Danske Videns

Selsk. Skrif., 6 Række,

Nudibranchiata EAE, (Melibe pellucida), on the Co-

lumbia River, Washington. C. Semper, Reisen im oe

Philippinen, wiss. Resultite, 9, VI. Lief I.

No. 638] MELIBE LEONINA (GOULD) 251

1907. ae Opisthobranchiata of South Africa; Tethymelib‘de.

Trans, 8. Africa Phil, Soc., 1

Boas, J. E.

1916. S EE fjerde, forögede udgave. Gyldendalske Bog-

handel, Kjöbenhavn.

Collingwood, G.

868. Observations on ok oe ee of Nudibranchiate Mollusea in

he China Sea. d Mag i

1879. g e PRE oa from the China e and Spawn De-

ribed. Trans. Linn. Soc., London (2), 2. (Read March 7,

1878.)

Crozier, W. J.

1917a, Evidence of Assortive Mating in a Nudibranch. Proc. Nat.

Acad, Sci., Vol. 3, No. 8.

1917b. oe the Periodic Shoreward Migration s nen Nudibranchs.

E AMERICAN NATURALIST, Vol, 51, 606.

1919. Assortive Mating ina Nodibranch, ELRES zebra Heilprin.

r. Exp. Zool., Vol. 27, No. 3.

Cuvier, G.

1868. Les yey a (Tethys fimbria, Lin.).

Daugherty, L. S., and M

1912. Pineiies of ak Zoology, p. 83. Philadelphia.

1917. Ibidem.

Eliot, ©. N. E;

1899. Notes on Tectibranches and naked Tiun from Samoa. Proc.

Acad. Nat. Sci. Phila. (51), Ser. 9.

1902. On Py Nudđibranchs from Sioa Proc. Zool. Soc., Lon-

don, 2.

1910. A Monograph of Nudibranchiate Mollusea. The Ray Society,

London,

Eliot, C., and Evans, T. J.

1908. penn Tagen a Doridiform Cladohepatie Nudibranch.

NE r. Mier. Sci., 52 (N. 8.).

De Filippi,

1865. Melibe FERAN Zool. del veaggio intorno al globo della reg.

freg. Magenta durante gli anni 1865-1868. Bees della R. Acc.

delle Sc. di Trino, 2 sér., T. 28, pp. 219-

Gamble, F. W.

1892. Observations on Two Rare British Nudibranchs. Ann. Mag. Nat.

Hist., 9 (6).

Garstang, W. A.

1890. A Complete List of the Opisthobranchiate Mollusca’ Found at

Plymouth—with Further Observations on Morphology, Colours,

and Natural History. J. Mar. Biol, Ass., 1 (N.8.).

Glaser, O. C.

1903. The pers ag of Nudibranch Molluses, Johns Hopkins

Uni

Gould, Augustus A.

1852. Chioraera leonina, Gould. U. S. Explor, Exped. Mollusca and

Shells, p. 310,

252 THE AMERICAN NATURALIST [Vou. LV

eae?

80. Sous die Organisation von ce fimbriata, Linn, (s. fimbria).

Sitzber, naturf. Freunde,

Hecht, Emile.

1895. Contribution a 1’étude des Nudibranches. Mem. Soc. Zool.,

v. Thering, H

1876. Tethys, pa cot zur Phylogenie der Gastropoden, Morph.

Jahrb.,

Jeffreys, J. G.

1869. History of Naked Marine Gastropoda. Brit. Conchol., 5

Johnson, Charles W.

1915. Fauna y New England 13, Occasional papers. Boston Soc, Nat.

Johnston, Georg

1838 Misostlines Zoologica. Ann. Nat. Hist., 1.

Krause, A

1885. Ein Beitrag zur Kenntniss der Mollusken-Fauna des Berings-

meeres II, Gastropoda und Pteropoda. Arch. f. Naturgesch.,

t6.

1895. Nudibranchiaten von Tromsö. Tromsö Mus. Aarshefter, 18.

Lang, Arnold.

1896. Text-book on Comparative Anatomy, Part II, London.

Locard, A.

1885. ero de malacologie francaise. Ann. Soe. Agrit, Lyon,

8 (5).

ear a F. M.

1912. The ETRE Family Dironidæ. Zool, Jahrb., Suppl., 15, 1.

von Martens

1879 aids der von W. Peters von 1843 bis 1847 im Mossambique

gesammelten Mollusca. Monatsber. Ak. wiss. Berl.

Meckelii, Joannis

1832. Additamenta ad historiam Molluscorum Piscium et Amphibiorum,

Acad. Frideri P Halensis.

Meyer, H. A., und Möbi

1865. Feuá der eae: Bueht, Opisthobranchiata I.

Oken, Dr.

1815. Lehrbuch der Zoologie, I, p. 278.

Odhner, Nils

1906. Northern and Aretie Invertebrates in the Collection of the

Swedish State Museum. III, Opisthobranchia and Pteropoda.

Kungl. Sv. Vet. dandloeaiaiis Handlingar, Band 41, No, 1

Pease, W. Harper.

1860. Melibe pilosa, Description of a New Species of Mollusca from

the —* Islands. Proc. Zoo!. Soc., London, Part 28, p. 34.

Pelseneer, P.

1895. Hermaphroditism in Mollusca. Quart. Jour. Mier. Sci., Vol. 37.

1898. Sur la condensation embryogénique chez un Nudibranche. Congr.

intern. Zool,

No. 638] MELIBE LEONINA (GOULD) » 253

Pilsbry, H. A.

1895. On the Status of the Names Aplysia and Tethys. Ann. Mag.

Nat. Hist., 16 a Ja

Rang, Sander.

1829. ea des ihiihi. pp. 129-130, Paris.

Reid, Joh

1846, On pegs a et ha = PE of the Nudibranchiate Mollusca.

. Nat. His

Sars, G. O.

1878. Oversikt over de i Norges arktiske region forekommende bléddyr.

Moll. reg. Arct. Norv., 1, pp. 315-316.

Sars, M.

1870. Kristiania fjordens Mollusker. Bidr. til kunsk. om Kristiania-

cae auna,

Smith, E. A., Eda sad. Kirkpatri ick

1905 À PaA to the Shell and ae Galleries, London,

Step, Edw.

1901. Shell-life, pp. 276-303, London.

Stuart, A.

1865, pie sa a ap O einiger Opisthobranchier. Zeitschr. f.

, 15, pp. 94-102.

Théel, Hamar

1908. On Djurlifvet i omgränsande fjorder och haf ved Kristineberg.

Ark för Zool., Stockholm, 4

Tryon, Jr., George

1882. Genie and Systematic E Mollusca, Vol. 1.

Vayssiere, A.

= 1901. Recherches zoologiques et pac sur ies Opisthobranches

golfe de Marseille, 3 partie, Nudibranches. Ann. du Mus.

de Marseille, 6.

1911. Recherches zoologiques et anatomiques sur les Opisthobranches

de la Mer Rouge et du Golfe d’Aden. Ann. fac. sci. Marseille,

Viguier, C.

1893. pedere a l'étude du developpement de la Tethys fimbriata.

i . Zool. exp., 6 (3).

Walton, C. L

1907. Nudibranchiate Fauna from the North Sea. Jour. Mar. Biol.

oc. Plymouth, 8 (N. S

TYPES OF MUTATIONS AND THEIR POSSIBLE

SIGNIFICANCE IN EVOLUTION.

DR. ALBERT F. BLAKESLEE

STATION FOR EXPERIMENTAL EVOLUTION

Tux beginning of the twentieth century saw the rise of |

two concepts which have profoundly affected biological

thought and been of increasing influence in the trend of

experimental study of plants and animals. The mutation

theory of deVries based on the evening primrose, and the

laws of Mendel based on the garden pea, settled the date

of birth of the modern science of genetics. The studies

on these two plants have together formed a basis for

the main bulk of our present genetic investigations.

While the garden pea stands intimately associated with

a conception of inheritance of wider application than was

at first imagined, the evening primrose and the theory of

mutation connected with it are by many considered to

furnish an example of a valuable theory founded upon

incorrect interpretations. The belief is growing that

most of the new forms which have appeared in cultures

of the (Mnotheras are not mutations at all and that the

evening primroses, as an abnormal group of plants, are

not to be seriously considered as representative of the

processes of evolution in normal forms.

In the short time at my disposal, I wish to outline some

recent findings in the jimson weed (Datura Stramonium)

which it is hoped may throw incidentally some light on the

more highly involved phenomena in the Œnotheras, and

which may serve as a basis of a brief discussion of their

possible ev olutionary significance.

The jimson weed is not supplied with a wide range of

obvious Mendelian characters. The early studies of

1A paper presented before the American Society of Naturalists at the

Chicago meeting, December 30, 1920.

254

No. 638] _ TYPES OF MUTATIONS 255

Naudin (9) and Godron (7) as well as the later investi-

gations of deVries (13), Bateson and Saunders (1) and

the writer and Avery (4) on this species have shown that

purple color in flower and stem is dominant to lack of

purple in those parts, and that spiny capsules are dom-

inant to smooth capsules. The writer with Avery (5)

has been able to add a third pair of contrasting char-

acters: ‘‘many nodes,’’ causing tall stature, in contrast

to ‘‘few nodes,’’ causing low stature. These are all the

allelomorphic pairs actually determined. Moreover, no

variation has arisen in the writer’s cultures during the

last seven years’ study of this species -which gave evi-

dence of differing from the present stock by a single

Mendelian factor. ` Distinct variations, provisionally

termed mutations, have, however, regularly recurred

whenever a sufficiently large number of plants have been

subjected to observation (5). So far as investigated,

they have been found to be connected with a duplication

of one or more of the normal chromosomes (6). The

normal quota is 12 pairs: 12 being therefore the gametic,

haploid, or x number; and 24 the somatic, diploid, or

2x number. The simplest type of duplication is the ad-

dition of an extra chromosome, probably by non-disjune-

tion in one of the pairs of the diploid complement, giving

24+-1, or 25 chromosomes as the somatic number. In

such plants there will be 11 sets of 2 homologous chro-

mosomes each and 1 set of 3 homologous chromosomes.

We have on the chart, Table I, 12 recurrent mutants

of the type just discussed; which, while perfectly dis-

tinct from each other and from the normal stock, have

certain characteristics in common. All have been found

to produce gametes with 12 and 13 chromosomes (there-

fore with 25 as the calculated somatic number); all have

a relatively large proportion of bad pollen grains, vary-

ing in the different mutants from 8 per cent. in the

Globe to 21 per cent. in the Spinach, as indicated in the

chart; all fail to transmit the mutant complex to any

considerable extent through the pollen, while they do

256 THE AMERICAN NATURALIST [Vou. LV

TABLE I

Somatic NUMBER OF CHROMOSOMES AND PERCENTAGE OF Bap POLLEN

FOUND IN NORMALS AND IN DIFFERENT MUTANTS

Somatic Number Per Cent,

Types of Chromosomes Bad Pollen

yrs CUR Sewanee teens Ce es E x. lee 2i

Mur

Teall (CT NOW BPOS i) 6 5.065 e ke 48 3.3

aE A E E EE envi een ve ce eee s 36 34.1

Simple Trisomic

Ea E O T E Ware ecdiees 25 7.9

rR n 0. ec hss ee rs so 4s 25 12.9

Behe kre VER WAY oe oa ee co is S eae oop 25 9.3

3 cae Sus area E ARA a 6a 25 18.3

ED A e Sea NERO E near gd purer a gn a 25 12.2

5 atin Bi eal T S EE sores vile 25 20.7

0O Dugat UME ya ack ec eae 25 16.1

Fo POE eS Pea ie ee os Cees 25 8.4

S: BOA er a oa i ee es 25 10.7

P. Bokne sc asians san Eht Sa 25 10.4

AOS? GIOI e oameni ee ons Cae ee 25 18.0

Ii. Mierocarpio Nae ee ee eee 25 12.8

12: pannel ciere oh oe Soin ee 25 20.7

transmit it through the egg cells, although to only about

one quarter of the offspring. That the offspring of these

mutants repeat the parental type regularly in less than

the 50 per cent. expected is probably due to the lessened

vigor of growth of mutants in comparison with normals.

If the presence of an extra chromosome in a given set

causes a specific mutation due to the constitution of this

particular chromosome, rather than to the mere pres-

ence of an extra chromosome irrespective of its origin,

there are at least two consequences to be expected. First

there should be as many possible mutants of this type as

there are chromosome sets which may undergo duplica-

tion. In other words there should be 12. Twelve, as a

matter of fact, is the actual number which we had Head

before the nuclear condition had been determined. In

addition, we have two or three mutant forms apparently

belonging to this class for which it has not yet been pos-

sible to obtain chromosome counts. In appearance they

are combinations or modifications of members of the

No. 638] TYPES OF MUTATIONS 257

recurrent twelve. Secondly, it should be possible by

breeding tests to connect up mutants with as many

chromosome sets as there are known Mendelian factors,

or factor groups. This connection we seem to have estab-

lished between the mutant Poinsettia and the set of

chromosomes which carries the factors for purple pig-

mentation in flower and stem.

TABLE IT

GAMETIC AND SOMATIC FORMULAE

TYPES OF CHROMOSOMAL DUPLICATION,

Pair A AND a AND

LANTS HETEROZYGOUS FOR FACTOR

sucH PL ARE SELFED, TOGETHER

IAGRAMS ILLUSTRATING THE CHROMO-

SOMAL CONDITION IN SOMATIC CELLS

No.

Extra No. |

ml Sets | Gametic Formula Selfed Ratios Somatic Formula = arom nit

somes Affect

in Se

PANI

AA + Aa 1A : Oa AAAa N N

AA + 4Aa + aa 35A : la AAaa = Z. =

Aa + aa 3A : la Aaaa me WSs

2 12 (12 + 12) (12 + 12) XG NY

+a2 +12) WU ZF

24 +e FAA NOR. 8A : la N

2Aa| MUT. 9A : 0a } Ada a 4 Peay

A +2a + 2Aa LN

| + aa| NOR. 5A : 4a ZA pee

MUT. 7A : 2a | Aaa mr N $

1 1 | 12,0241 orao A || 7

| |

rip a2 + 12) + 12 ZAUN

No. of chro- | | | | | N “=

mosomes | = semis

aR 3) 7| 8 9/10 11 |12| ESSR

Frequencies | 1 |12 6 0p 408 792 |924 792/495 220 66 |12| 1| ST 7

The set of 3 chromosomes in the diagram, Table II,

may be called the Poinsettia set, or the purple set. A

Poinsettia plant may, to speak in terms of the dominant

factor, be considered nulliplex with no dominant genes,

or simplex, duplex or triplex with, respectively, 1, 2, or 3

dominant factors.. There are therefore two types of

heterozygotes, and under greenhouse conditions these ap-

parently can be distinguished from each other as well as

258 THE AMERICAN NATURALIST [Vou. LV

from the homozygous dominants by different intensities

of pigmentation. Simplex heterozygotes when selfed

throw offspring with 5 dominants to 4 recessives among

the normals, and 7 dominants to 2 recessives among the

Poinsettias; while duplex heterozygotes should give a

ratio of 8:1 among the normals, and all dominants among

the Poinsettias. There is evidence which seems to indi-

cate that the mutant Cocklebur is conditioned by dupli-

cation in the chromosome set which carries both the fac-

tors for spiny capsules and also those for number of

nodes. If this is actually the case, we must assume that

these two. factor pairs are loosely linked in the same

chromosome with about 50 per cent. crossing over, since

they appear to segregate independently of each other.

We have been discussing duplication of a single mem-

_ber in only one of the 12 chromosome sets. On the lower

part of the chart (Table II) is represented the only plant

we have yet found with an extra chromosome in every

one of its 12 sets. Such a plant is triploid. What its

breeding behavior will be, can not be told before another

season. If the chromosomes assort at random, the

gametes theoretically should have the chromosome num-

bers indicated in the chart, and the counts which my col-

league, Mr. Belling, has made from figures in pollen

mother cells are not inconsistent with the distribution

of the theoretical frequencies. One might expect such

triploid plants to give rise to individuals intermediate

between triploids and mutants of the Poinsettia type; in

other words to mutants with duplication of chromosomes

in 2, 3, 4, ete., up to duplication in all the 12 sets. Such

compound mutants we have not yet been able to surely

identify in our cultures; but we have never before this

past season had a triploid plant, which from the wide

range of gametic types in its egg cells would seem a

likely source of such mutations.

Tetraploid plants have been discussed at yesterday’s

session of the Botanical Society of America. They rep-

resent a further duplication over those of the triploids

No. 638] TYPES OF MUTATIONS 259

already mentioned in that there are 4 homologous chro-

mosomes in each set in somatic cells. The homologous

chromosomes therefore form tetrasomes, to use a new

term,” instead of disomes as in normals or trisomes as in

triploid plants. Members of these tetrasomes appar-

ently assort at random in the reduction division. In

consequence, certain peculiarities in breeding behavior

result. Plants duplex for a dominant factor (AAaa)

will, when selfed, give a ratio of 35 dominants to 1 re-

cessive in the offspring. Plants simplex for the domi-

nant (Aaaa) will give a 3:1 ratio in their offspring; but

a third of the dominant offspring will throw 35:1 ratios

in the next generation. Plants triplex for the dominant

(AAAa) will give in the immediate offspring all domi-

nants; one quarter of which, however, may be expected.

to give a 35:1 ratio in a later generation. The results

expected from selfing the 5 zygotic types are shown in

Table III.

It might be expected that mutant forms would be

found in which doubling of the chromosomal number had

involved only a single one of the 12 sets. Such mutants

would bear the same relation to tetraploid plants with

all the sets involved that the Poinsettia type of mutants

bear to triploid plants. They have not yet been found,

however.

The following terms are suggested to designate sets with numbers of

oriana from 1 to 12: monosome, disome, trisome, tetrasome, penta-

some, hexasome, heptasome, oktasome, enneasome, dekasome, hendekasome,

dodekasome.

The number of sets affected by duplication may be indicated by the

_ terms: pranoi double, triple, qua Pa quintuple, sextuple, septuple,

octuple, Pee decuple, undeeuple, duodecuple

The settia and Glebe: are simple EEE hratants. If the Globe and

boiie eis be combined to form a mutant with 3 chromosomes each

in two of the 12 sets, such a mutant would be called a double trisomie

mutant. If differential aes of gametes does not interfere, the triploid

plant already ea should produce, theoretically, offspring of all the

trisomie types from simple to duodecuple. Haploid, diploid, triploid, tetra-

ploid, ete., are sities already employed to designate plants with the sam

number of chromosomes in all the sets.

260 THE AMERICAN NATURALIST [Von LV

TABLE III

TETRAPLOID PLANTS. RESULTS. OF SELFING TETRAPLOID PLANTS ARISING

OM THE Cross oF A Homozygous DoMINANT (AAAA) BY A

RECESSIVE (aaaa), CARRIED TO THE F, GENERATION

In the F,, only phenotypes are represented.

Pi— AAAA and aaaa

Fı—A Aaa (Gametes of Fi — AA + 4Aa + aa)

F,—1AAAA ote 8AAAa + 18AAaa ES npa + laaaa

F:—AAAA AAAA + 2AAAa + AAaa AAaa + 2Aaaa + ange er

| | |

Fy—A A A 35A: la 35A: la 3A: 1a a a

It is possible that a single set in an otherwise tetra-

ploid plant may have an extra chromosome, giving 5

chromosomes in one set and 4 in the remaining eleven.

At least we have a single plant in a tetraploid pedigree

which strongly resembles the Globe—the best known of

our simple trisomic mutants. The cytological evidence

shows that its chromosomal number is at least tetraploid,

but is not yet sufficient to prove that its Globe-like ap-

pearance is determined by the addition of a fifth member

to the chromosomal set responsible for the Globe mutant.

The occurrence of mutations of the types discussed in

` the foregoing paragraphs is bound up with the causes of

chromosomal duplication. Knowing the mechanism to be

affected, we may be able ultimately to induce chromo-

somal mutations by the application of appropriate

stimuli.

We have outlined the types of chromosomal duplica-

tion already found in the jimson weed, and have shown

some of the peculiarities in the breeding behavior of the

mutant forms which they condition. It will be well to

consider for a moment this process of duplication as it

affects the individual plant and as it may have a possible

significance in our theories of mutation and evolution.

The mutants of the Poinsettia or Globe type, in which

but a single chromosomal set is involved in the duplica-

tion, should enable one to discover something in regard

No. 638] TYPES OF MUTATIONS 261

to the influence of each specific chromosome upon the

morphology and physiology of the datura plant. While

there seems to be but a single chromosomal set respon-

sible for the presence or absence of purple pigmentation,

probably each chromosome has an influence upon the

strength of expression of the pigment since the several

mutants appear to differ widely in color when homs-

zygous for the main purple factor. Thus Glossy is

darker purple than normals, while Cocklebur is distinctly

lighter. In normal plants there is a balanced adjustment

between the modifying factors in the different chromo-

somes. When this balance is disturbed by the addition

of only a single extra chromosome to one of the 12 sets,

profound changes are brought about in the ontogeny of

the resultant plant. When all of the sets have an extra

chromosome, however, as is the case in triploids, no great

disturbance of the balance is brought about and the plant

is not greatly different from normals. Even in tetra-

ploid plants where all the sets are equally affected, al-

though the total number of chromosomes is doubled, the

difference from normal is not so great as in mutants of

the Globe and Poinsettia series. The leaves of tetraploid

plants, when carrying the factor for many nodes, may be

distinctly larger than those of normals. Few-noded tetra-

ploids, however, are less easily distinguished. The best

diagnostic character has been the globose shape of the

eapsule, and yet plants known to be tetraploid from cyto-

logical evidence have been found this past seasen with

capsules perfectly normal in appearance.

What is the bearing of the phenomena of chromosomal

duplication in datura upon the mutation theory? In the

first place, the mutants of the Globe type apparently cor-

respond to the lata type of mutants in the Œnotheras in

which an odd somatie chromosome has been determined,

although in these Cinothera mutants no breeding evi-

dence has been available to show that the peculiarities of

mutant lata are due to the presence of an extra chromo-

some in any specific chromosomal set. Our tetraploid

262 THE AMERICAN NATURALIST [Vou. LY

mutant ‘‘New Species’’ corresponds to Œnothera gigas

and is brought about by a doubling of the chromosome

number. The color ratios in our tetraploid daturas indi-

cate that Enothera nanella is a Mendelian segregate and

suggest that other of the (Hnothera mutants which give

monohybrid ratios in crosses may be of the same nature.

Our evidence in regard to O. nanella comes from the oc-

currence of this mutant in cultures of O. gigas. DeVries

(14) reports that certain races of gigas when selfed regu-

larly produce from 1 to 2 per cent. nanella mutants, while

certain pedigrees give monohybrid ratios which, on ac-

count of the lesser vitality of the recessive nanellas, show

a higher proportion of the dominant gigas forms. From

the pedigrees approaching a 3:1 ratio he obtained plants

which bred true, except again for the 1 to 2 per cent. of

nanella mutants in their offspring. A glance at the chart

(Table III) will show that, if our theory of tetraploidy

be correct, the 1 to 2 per cent. of mutant nanellas which

deVries obtained by selfing plants from 3:1 pedigrees

must have been the recessives in a 35:1 ratio since no

dominant plants in a 3:1 pedigree of a tetraploid race

could be expected to breed true. The dominant pheno-

types must either throw 3:1 ratios again or 35:1 ratios.

The deviations of the nanella mutants in this case from a

35:1 ratio is accounted for by a similar proportionate

deviation in the 3:1 ratio. The work of Muller (8) on

balanced lethals strongly suggests that such of the Učno-

thera mutants as are not caused by chromosomal duplica-

tion are due to cross-overs from a balanced lethal con-

dition.

What then is a mutation? I do not feel we need to be

bound by its application to the evening primrose for

reasons of priority, since Waagen (15) had previously

used the term in paleontology in an entirely different

sense. I believe, with the idea that mutations must in-

volve a qualitative change, that we shall ultimately con-

fine the term to mutations of genes, although such muta-

tions may later be shown to be as different from our

No. 638] TYPES OF MUTATIONS 263

present conceptions of them as are mutations in the

(Knotheras from the conceptions in deVries’s classical

publication, ‘‘The Mutation Theory.” It may still be

desirable to employ the word mutation as a collective

term to designate the sudden appearance of any apparent

genetic novelty—whatever its real cause—until we know

better. Strictly speaking I should not call chromosomal

aberrations mutations when the changes are purely quan-

titative. The occurrence of tetraploidy would therefore

be no more a mutation than the doubling of chromosomes

at the origin of the sporophyte from the gametophyte of

ferns.

We have seen that chromosomal duplications and re-

lated phenomena may simulate gene mutations in their

effects upon the individual. What is their possible sig-

nificance in evolution? Let us first consider tetraploidy.

Numerous investigators have called attention to the fact

that the chromosome numbers of plants are more fre-

quently in multiples of two and four than one would ex-

pect from random sampling. Pairs of related species

have been listed for which one member had twice as many

chromosomes as the other. Such species have even been

called tetraploid. We feel strongly the desirability of

confining the term tetraploidy to those cases in which the

4x number is brought about by a doubling of homologous

chromosomes. Doubling by transverse division is a pos-

sible method, but would not be included in the term.

Tetraploidy has been observed in experimental cultures

of cnothera, primula and datura. Do such tetraploid

plants occur in nature, and are they capable of giving

rise to taxonomically new species? It may be mentioned

that the tetraploid datura was called ‘‘ New Species ”’ be-

fore its tetraploid nature was suspected. It satisfied the

requirements of an independent species. The pollen was

relatively good, and the mutant formed a distinct race,

self-fertile and fertile inter se, while practically sterile

with the parent stock. Tetraploid plants, therefore, stand

slight chance of being swamped by hybridization with the

264 THE AMERICAN NATURALIST [Vou. LV

species from which they have sprung. Once arisen, their

chances of survival would depend upon their ability to

complete with other forms in the struggle for existence.

There are no certain cases of tetraploidy known outside

of cultivation. It must be admitted, however, that their

identification would be difficult. I have shown that gigan-

tism is not an invariable diagnostic feature of tetraploid

daturas. As yet no cytological criteria of tetraploidy

have been established. The breeding behavior, which is

the only safe test, might easily be misinterpreted, as it

was apparently by deVries in the case of the tetraploid

gigas and nanella. Moreover, a suspected form: must

show a pair of Mendelizing characters before a breeding

test can be applied.

Despite the paucity of evidence in regard to the oc-

currence of tetraploidy in nature, the speaker believes

that it may have been one of the principal methods in the

evolution of plants. Its occurrence would furnish the bar-

rier between a new species and its parental form that

_ Darwin sought, and it would give a reason for the prev-

alence of even numbers in the counts of chromosome

pairs. I believe that a search for tetraploid forms in

nature will be rewarded. Perhaps they will more likely

be found in horticultural races propagated by vegetative

means. I take this occasion to suggest the desirability

of testing for tetraploidy any gigas-like plant that may

be found in the wild or under cultivation. We are making

a special study of tetraploidy at the Station for Experi-

mental Evolution and should be glad to receive plants

suspected of being tetraploid from any who do not care

to make the necessary tests themselves.

Even if proper tests should show that few forms in

nature were tetraploid in the sense that each chromo-

somal set in somatic cells was composed of 4 homologous

members, tetraploidy might still be a stage in the origin

of species with an even number of pairs of chromosomes.

In the 3 forms in which tetraploid plants have arisen

under observation, the 4 homologous chromosomes in a

No. 638] TYPES OF MUTATIONS 265

set apparently assort at random in the reduction division.

If, instead of acting individually without predilection one

for another, the four should come to assort in pairs, we

should have a different ratio in the F, generation (15:1

instead of 35:1). There would still be duplication of

genes and a 4x number in reference to the parental form,

but independent assortment of the chromosomes would

have been lost. It will not be possible to go into the

details of the argument. It is at least suggestive that

Shull (12) has found 3 cases of duplicate genes in the

shepherd’s purse, which has 32 chromosomes (that is 4

times 8); and Nilson-Ehle (10) has found a ease of tripli-

cate genes in a wheat having 42 chromosomes, which is 3

times the number in another variety of wheat (11). °

If tetraploid plants have been of influence in evolution,

it is probable that the other types of duplication have

also been of influence. A mutant of the Globe type with

a single duplication in one of the 12 sets ordinarily fails

to hand on the duplication through the pollen. Occa-

sionally it might do so, and we should then expect a con-

stant race with 4 homologous chromosomes in one of the

12 sets. If these 4 should cease to assort at random and

pair, we should again have a possible duplication of genes

and an added pair of chromosomes characteristic of the

race.

There is not time at my disposal to discuss mutations

of genes. In a recent paper (2) on a somatic mutation

in portulaea, I have indicated my belief that mutations

of genes may occur at any stage in the development of

the plant. We have found color mutations which affected

only the epidermis, and therefore could not be trans-

mitted through seeds. We have also found similar color

mutations which affected only the sub-epidermal tissue,

and therefore could not show in the petal; but which be-

came evident from the seeds produced from this mutated

tissue. There seems to be no preferred location for the

origin of factorial mutations in flowering plants, although

they are more readily transmitted if they occur in the

266 THE AMERICAN NATURALIST [Vou. LV

gametes or in the embryo. The fact that in vegetatively

propagated Mucors (3) I have found mutations relatively

common where the possibility of sexual reproduction was

ruled out, indicates that sudden genetic changes are not

necessarily associated with sexual processes.

It has not been possible in this brief presentation to

give an extended classification of mutations, nor to dis-

cuss in detail their possible significance in evolution. It

will be sufficient if I have made clear the distinction which

must be kept in mind, in any discussion of the subject,

between mutations in individual genes and those brought

about by chromosomal aberrations.

BIBLIOGRAPHY.

1. Bateson, W., and E. R. Saunder

1902. a Studies in aa Pildbony of Heredity. Datura.

o the Evolution Committee of the Royal Society,

2, acts A

1920 iia Mutation in Portulaca showing Vegetative Reversions.

Genetics, 5: 419-433, fig.

3. Blakeslee, A. F.

1920. Mutations in Mucors. Journa! cf Heredity, 11: 278-284, figs.

28.

4, peo A. F., and B. T. Avery.

917. — ‘ets and Jimson Weeds. Journal of Heredity, 8:

5-131, figs. Po

5. Blakeslee, K F., and B. T. Avery.

1919. Mutations in oe “ima Weed. Journal of Heredity, 10:

111-120, figs.

6. ee A Fg. =p and M. E. Farnham.

192 Chivetommel duplication a epes Phenomena in Datura

Mutants. Science, N. S., 52: 388-390.

i Godrin, D. A.

1873. Des hybrides et des métis de Datura. DaN 1-75.

- 8. Muller, Herman J.

1918. Genetie variability, twin hybrids and constant hybrids, in a

case of balanced lethal factors. Genetics, 3: 422—499, fig. 1.

9. Naudin, C.

E Noenes recherches sur l’hybridité dans les végétaux. Nouv.

Arch. Mus., 1: 41-54.

` aeran Ehle, H.

1909. Kreuzungsuntersuchungen an Hafer und Weizen. Lund’s

Univ. Arsskrif

No. 638] TYPES OF MUTATIONS 267

a1, nomen Tet

1918. Kurze Miteinne iiber die Chromosomenzahlen und die Ver-.

wandtschaftsverhaltnisse der Triticum Arten. Bot.. Mag.

To ous 32: 151-154

12. orn G.

A aes Duplication “i Genetic Factors in yas eared Purse.

Science, N. &.,

13. deVries, H.

909. Das Spaltungsgesetz a Bastarde. Berichte d. Deutschen

Bot. Gesellsch., 18: 83-90.

14, nails E H:

1915. @Œnothera gigas nanea, a Mendelian mutant. Bot. Gaz., 60:

337—345.» ; 2

15. Waagen, W. H.

1868. Die Formenreihe des Ammonites subradiatus. Benecke’s Geog-

nostische Pa'äontologische Beiträge, 2: 185-186.

BOOKS AND LITERATURE

Die Chromosomenzahl von Zea Mays L. Ein Beitrag zur Hy-

pothese der Individualität der Chromosomen und zur Frage

über die Herkunft von Zea Mays Lt By YOSHINARI Kuwapa.

The author of the paper, the title of which appears above, has

well summed up his purpose in the subheading. As this article

reports investigations of considerable cytological importance in

a publication which is not likely to have wide circulation in

America, it was thought advisable to review it at some length.

As Professor Kuwada clearly and concisely states his results

and conclusions in his summary a translation is given below.

1. The chromosome number of Zea Mays L. is 10 (when the diploid

number is 20). In forms either closely related systematically or sup-

posedly ancestral types the chromosome number in the root tips is in

general constantly 20 (seldom does the number approach. the octoploid

number

2. It has been found that in one race of sugar corn which I received

from the Agricultural Faculty of the Imperial University of Tokyo

the chromosome number is different in different individuals. In the

roots tips 21, 21, 22, 23 and 24 chromosomes were found. The number

of tetrads is correspondingly different, namely, 10, 11 and 12. There

is no relation between the chromosome number and the chemical nature

of the endosperm.

3. Through a parallel study of the number and size of the tetrads

and the length of the chromosomes in the root tips it has been shown

that the number of chromosomes is increased through the cross frag-

mentation of certain chromosomes.

e measurement of the chromosome length in the root tips

and the unequal lengths of the component elements of the tetrads

* allow us to draw the important conclusion that Zea Mays is of hybrid

nature, and indeed, as Collins has rightly said, a hybrid between

Euechlena and an unknown plant of the Andropogonee.

e chromosomes supposedly derived from Euchlena are longer

than those coming from the Andropogonex species, so that the tetrads

under certain cireumstanees are made up of elements of different .

lengths. The two chromosomes of the first kind A—B? and C—D have

1 Jour. of the Col. of Science, Imp. Univ. of Tokyo, Vol. 39, Art. 10, 1919.

2 It has been necessary for convenience to take some liberties with the

method used by Kuwada for expressing his idea of the potentiality of frag-

mentation possessed by the various chromosomes. Here a solid dash be-

268

-= No. 638] BOOKS AND LITERATURE 269

an inclination to fragment easily under certain conditions while the

corresponding chromosomes of the latter type a b and e d do not show

cross fragmented into two chromosomes A B and C D and this condition

is morphologically and genetically fixed. We therefore have three kinds

of corresponding chromosomes: the eross-fragmented chromosomes,

those having a tendency to fragment and those in which both of these

TET are lacking

. In the formation o. the tetrads the chromosomes A B and C D

are Trish to A—B and C—D and recessive to a b and ea. The

dominance in the first case is somewhat unstable, so that the number

of tetrads is subjected to fluctuation within certain limits. The dif-

ference in the behavior of the corresponding chromosomes A—B, C—

and a b,e dto A B C D is also a point in favor of Collins’ hypothesis.

7. If the chromosomes A B and A—B form a tetrad in the reduction

division four combinations result: A. B; A—B, A—B and -B.

The corresponding ends of the chromosomes A- and -B fuse latay

easily to reform the chromosome A—B. The possibility of fusion de-

pends absolutely on the proximity of the corresponding ends of the

passive cross-fragmented chromosome A— and -B. In this respect the

parallel arrangement of the homologous chromosomes in the somatic

cells is of great importance. The chromosomes A and -B or A- and

B which would otherwise occasionally fuse to form the chromosome

A and -B or A- and B remain sometimes as A and -B or A- and B:

the result being the variation in the number of the chromosomes. Two

kinds of gametes occur, in one the chromosome number is constant and

in the other it varies. The chromosomes in the first instance have the

formula A B (number of chromosomes above normal) or A—B (nor-

mal number of chromosomes), and in the latter instance A—B or A- B

or occasionally A- —B. When the chromosomes A B and a b form a

tetrad the result is very simple in that only two combinations are pos-

si B and a b. In these cases the number of chromosomes is

constant. The union of A and B is only a phenomenon ascribed to the

presence of a b.

The empirical results agree with those developed from theoretical

considerations based on the laws of chance.

8. The applicability of the laws of chance to the chromosome num-

tween two letters indicates a weak place that may easily break, a eones

underscored single letters have no power of uniting (no free ends), while

the binding together of two letters by underscoring represents a chromosome

which can never fragment.

270 THE AMERICAN NATURALIST - [Vor LV

ber and the constancy of the true length of the chromosomes in the

hybrids is a contribution favoring the individuality hypothesis.

9. The nuclear and cell size is dependent on the chromosome size

and on the other hand the latter is modified by the cell size.

According to Kuwada there are two hypotheses concerning the

origin of Zea Mays L. Iltis (1911) first suggested that this mod-

ern form might have been derived from some unknown tribe of

Andropogoneex, while Collins in 1912 put forward the claim that

Zea Mays L. was a hybrid between an unknown species of the

Andropogonee and Euchlæna.

In his cytological studies Kuwada finds that in species of

Euchlena and Andropogonee the chromosome number is the same

as in Zea Mays L.—namely 20. In only one of the investigated

groups of plants belonging to the Andropogonew was the chro-

mosome number above 20, which places this particular species

beyond consideration. The measurement of the chromosomes in

a Euchlena from south Florida shows that their total length is

greater than is the case in Andropogon Nardus L. var. Georingin

Hack. The respective total chromosome lengths in each case are

given as 188.25 mm. and 111.3 mm

Kuwada gives the results of a large number of measurements

of the chromosomes in various varieties of maize taken at random

or from plants in which the cytological conditions have been

studied in the parental, F, and F, generations. His conclusion

that the figures indicate that two length types of chromosomes

are concerned in the modern plant do not seem to the present

writer to be entirely born out by the facts.

In the measurement of chromosomes previous studies have

shown that complexes from the same individual in the same or

in different parts of the structure may show considerable varia-

- tion in the total length of their component chromosomes. In

general, of course, small cells will have smaller chromosomes and

larger cells, larger chromosomes, but even in similar tissues very

appreciable differences may occur. These variations are ob-

viously due both to internal and to external causes. Fluctuation

in the climatice or nutritive conditions may affect growth and

vigor, while the volume of the cell imposes limitations on the size

of the contained chromosomes. It has been shown by the present

writer? that, be the total length of a complex long or short, the

3 Hance, R. T., 1917, ‘‘The Diploid Chromosome Complexes of the Pig

(Sus scrofa) and their Variations, Jour. Morph., Vol. 30. 1918a, ‘* Varia-

tions in the Number of Somatic Chromosomes in (nothera scintillans De-

No. 638] BOOKS AND LITERATURE 271

individual pairs always bear the same relation to each other, al-

lowing the conclusion that whatever influences the size of the

chromosomes generally affects all similarly. The figures of Ku-

wada bear out these observations very well. As pointed out

above, his interpretation of his work seems somewhat forced. He

recognizes the factors playing rôles in the behavior of the chro-

mosomes, but does not feel that his results can be entirely ex-

plained by them.

To illustrate what is meant by the above criticism let us con-

sider a cross made by Kuwada between sugar corn 22,,,, and Black

Mexican 58,,,), another sugar corn. In both, dividing cells from

adventitious roots were uniformly selected. The former has

chromosome complexes averaging 149.05 mm. in length while the

latter gives a total of 172.17 mm. This to Kuwada indicates a

real and genetic difference in chromosome length, although in

the same Black Mexican plant 58,,,, complexes from side root

tips average only 145 mm. in length. This would signify that

the length 172.11 was no more fundamental in plant 58,,,, than

was 145, and lessens the weight of the evidence that the higher

number betokens genetic chromosome differences with the length

149.05 in plant 22,,,,. When the two plants are crossed the

chromosome lengths in the hybrids are almost exactly one half of

the sum of the lengths of these structures in the parents if 172.17

is accepted as the typical complex length for plant 58,,,,.—

1/2 (149.05 + 172.17) =160.61. It may be pointed out here that

one half the sum of the complex length found in the various roots

of plant 58,,,) also closely approximates the same figure—

1/2(172.17 + 145) =158.58. The F, plants from the above

eross possess sets of chromosomes whose length is very close to

that expected on Kuwada’s assumption of 149.05 and 172.17 as

the basic or typical lengths of the parental chromosomes. The

chromosomes in the F, plants varied from 155.75 mm. to 168.9

mm. and averaged 161.86 mm. This number fits in well with

the anticipated result and at first would seem to justify the con-

sideration of 172.17 mm. as the representative length for plant

58a) However, the chromosomes in the F, offspring were

found in cells in the radicles of seed germinated in moist saw

dust. The chromosomes in this early root tip in many forms are

not infrequently larger than are found in the growing parts of

Vries, Genetics, Vol. 3. 1918b, ‘‘ Variations in Somatic Chromosomes,’’

Biol. Bull., Vol. 35.

272 THE AMERICAN NATURALIST [Vou. LV

the older plant and Kuwada’s figures and statements show that

maize is no exception to the general rule. This tendency for the

chromosomes in the radicle to be larger puts a fictitious value on

their measurements in this organ for comparison with the di-

mensions of chromosomes found elsewhere in the plant. As a

matter of fact, in the number of examples given the average

length of the chromosomes in all the plants is only a trifle more

than one per cent. shorter than the similar data in regard to

the chromosomes of the radicle, which difference would not

greatly affect the end result. In this instance, although the

physiological location of the chromosomes was undoubtedly one

factor in determining their size, objection on this ground alone

to the submission of the records of the F, chromosome lengths in

substantiating the figure 172.17 as the fundamental chromosome

length for plant 58,,,, would not seem to be entirely valid.

However, to base a broad conclusion on the lengths of the chro-

mosomes found in a particular part of a plant, even though com-

paring them with chromosomes from similar parts of other plants,

- is likely to obscure the real condition.

As has been shown in plant 58,,,), lengths of 172.17 mm. and

145 mm. were found. That these are not fixed lengths for the

particular tissues concerned in this variety of corn is shown by

the data given for other plants of the variety Black Mexican, in

which lengths vary (for corresponding tissues) from 132.5 mm.

to 181.25 mm., the average being 159.32 mm. There can be little

question that die variety Black Mexican, as long as it is genet-

ically pure, can have anything but comparable sets of chromo-

some throughout, holding in mind that though the lengths may

vary the inter-pair relationship remains constant. Less varia-

tion in chromosome length is shown for the three plants of the

variety ‘‘ Sugar Corn’’ which were studied. The range of

averages here is from 147.8 mm, to 151.6 mm

Lastly, if real differences between the lengths of the chromo-

somes in plants 58,,,, and 22,,,, exist greater differences be-

tween the members of the pairs that are found in the hybrid

offspring would be expected. Actually these elements mate

up well as to length and if unequal homologous chromosomes have

entered the zygotes union in a common environment has regu-

lated their proportions. As the dimensions of the chromo-

somes are in part a function of their environment the selection

as typical of any one complex or of even the average of com-

No. 638] BOOKS AND LITERATURE 273

plexes from certain tissues only is not justified, considering our

present ignorance of chromosome volume.

In support of the difference in length between the homologues

of chromosome pairs as indicative of the genetic length types

which Kuwada believes he has demonstrated he publishes draw-

ings of tetrads showing the unequal length of the component

elements. Personally I do not think that the figures are nec-

essarily conclusive proof, since the arrangement of the homo-

logues in several cases suggests a possible foreshortening, making

the true length doubtful, and in other instances the drawings

may well represent an entirely different form or stage of the

tetrad. It is not the intention of this criticism to convey the

impression that the investigator’s figures fail absolutely in prov-

ing his point concerning the uneven length of the homologues,

but rather to indicate that the illustrations are not nearly as

satisfactory and as conclusive as those given in the publications

of Wenrichtand Carothers> for somewhat similar conditions in

other forms.

Between the Enoha and Andropogoņeæ studied chromo-

some length differences appear which can scarcely be accounted

for on the basis of environment. Since the characteristics of

Zea Mays L. are intermediate between these forms the hope is

raised that two sets of chromosomes will be found in the modern

species, which hope the reviewer does not think has been realized.

Indeed, though recognizing the evolutionary position of Zea

Mays L. as given by some taxonomists, he offers the suggestion

that in his opinion the present investigation has not, as far as the

chromosomes are concerned, excluded the possibility of the

origin by mutation of Zea Mays L. from either Euchlæna or the

Andropogoneæ. A knowledge of the behavior of the chromo-

somes of these two forms in hybrids would be interesting and

important.

In explanation of the variation in the number of chromosomes

which Kuwada found in certain lines (20 to 24 chromosomes) he

devised an exceedingly ingenious scheme which apparently thor-

oughly accounts for the numbers of chromosomes occurring in

the offspring. It operates on the laws of chance and its theory

4 Wenrich, D. H., 1916, ‘‘The Spermatogenesis of Phrynotettiz magnus

and the Seg of the Chromosomes, Bull. Mus. Comp. Zool., Harvard

College, Vol. 6

ap cee E. E., 1913, ‘‘The Mendelian Ratio in Relation to Certain

Orthopteran Orena,” ’ Jour. Morph., Vol. 24.

274 THE AMERICAN NATURALIST [Vor. LV

seems to be completely justified by the results. As this expla-

- nation is adequately outlined in the translated summary further

space need not be devoted to it.

The investigator’s theory of factors located in each chromo-

mere which govern the form of the chromosome, while con-

venient in explaining the cause of the reunion of the chromo-

some fragments in maize, is scarcely necessary. Chromosomes

are not inherited as are the determiners for adult characteristics

in the form of minute chemical forerunners, but are passed on

complete in all respects. Consequently, factors to determine

their form in the next generation are not needed—the chromo-

some itself is carried over. The actual form of the chromosome

has been shown by MeClung, Wenrich, Carothers and others to

be determined largely by the location of the spindle fiber

attachment.

It is considered that the reviewed report has not clearly

demonstrated the origin of Zea Mays L. by means of chromo-

some measurements for the following reasons:

1. The length of the selected chromosome complexes in the

forms particularly studied are not typical of the plant and

such selection gives a false impression of the actual conditions.

. The figures illustrating the length differences of the homo-

pees composing the tetrads are not entirely convincing or

satisfactory. —

3. If two types of genetically fixed chromosome lengths exist

in maize we would expect to find an expression of this difference

when both types enter into the same individual. As far as the

reviewer’s interpretation of the tables of length is concerned,

this difference does not exist in the F, plants.

Though there are reasons for not considering that Kuwada

has proved his claims of the origin of Zea Mays L. he, never-

theless, is to be sincerely congratulated on an excellent cyto-

logical contribution involving great labor and care. To the re-

viewer the apparent failure of Professor Kuwada to demonstrate

his main thesis dwindles in importance when the value of the

‘‘side issues’’ of the investigation are considered. His work on

Zea Mays L. presents the following data:

1. The chromosome pairs of a complex may be arranged in a

graduated length series and between each pair there is an ap-

proximately equal difference in length.

2. The genetic relation of the chromosomes is shown i in parent

and offspring.

No. 638] BOOKS AND LITERATURE 275

3. When chromosomes fragment in Zea Mays L. it is the

longer ones that are affected. These fragments may also fuse,

causing variability in the total chromosome number.

4. Suggestive methods of studying chromosomes have been

devised.

5. Fragmentation has been accounted for on the basis of

genetic tendencies and the variable number of the chromosomes

in the offspring of certain plants has been ingeniously explained

with the aid of the device described in his summary.

The first four points (with the exception of the latter part

of the third, which has not been observed) agree perfectly with

the reviewer’s earlier work on the (Enotheras and the pig. As

to the fifth point, he has never found fragmentation in the

germ line.

Difficulties of interpretation in metrical studies of chromo-

somes arise from a lack of standards, i.e., knowledge of the limits

of variation that chromosomes of a given form will show under

many conditions and of the uncertainty introduced by the per-

sonal equation involved in drawing and measuring. With the

hope of deriving such standards the present writer is at work

on a plant and an animal possessing very few chromosomes.

The usefulness of the information drawn from such studies has

been elsewhere discussed.

Rosert T. Hance

ZOOLOGICAL LABORATORIES,

UNIVERSITY OF PENNSYLVANIA.

SHORTER ARTICLES AND DISCUSSION

‘““HOMING”’’? BEHAVIOR IN CHITON?

1. A STATEMENT concluding a recent preliminary account of

the ‘‘homing’’ habits of the pulmonate Onchidium floridanum

(Arey and Crozier, 1918) reads as follows:

To the extent that the homing habits of Onchidium may be proved

to involve associative memory, this snail may be placed in a series

comprising such types as Chiton, Fissurella, Onchidium and Octopus,

all four of which, in a sense, exhibit homing behavior, but of increas-

ing degrees of precision and complexity in the order of arrangement

here given.

The observations warranting a contention of this sort, so far

as it involves Chiton, were not fully available when our analysis

of the sensory responses of Chiton tuberculatus (Arey and

Crozier, 1919) was written, and I have therefore considered it

appropriate, as an addition to that report, to indicate the nature

of the facts leading us to ascribe to Chiton a certain degree of

‘‘homing’’ behavior.

2. It was noted by Heath (1899, p. 4), on the Californian

coast, that the adult Ischnochiton magdalenensis is found dur-

_ ing the day under boulders between tides, but that at night the

molluse comes out to feed on seaweeds growing upon the rocks,

retiring to dark situations after sunrise. Species of Mopalia,

and Cryptochiton, were found to ‘‘remain out on their feeding

grounds only when the day is foggy or dark.” Numerous other

species are more or less photonegative (cf. also Crozier, 1919b),

but some nevertheless continuously occupy situations brilliantly

illuminated (cf. Heath, 1899, p. 4; Plate, 1901; Pelseneer,

1906, p. 50). It has been shown elsewhere (Crozier and Arey,

1918; Arey and Crozier, 1919) that young individuals of Chiton

tuberculatus are photonegative to ordinary daylight, the older

ones photopositive. This matter of photic irritability is inti-

mately concerned with certain diurnal movements simulating

‘‘homing’’ behavior.

Heath (1904) was of the opinion that the bilateral larval eyes

of some chitons, persisting as they do well into postlarval life, -

until the shell plates become opaque, might be functionally im-

1 Contributions from the Bermuda Biological Station for Research, No. 127.

276

No. 638] SHORTER ARTICLES AND DISCUSSION 277

portant in determining responses to light. Earlier, he had noted

(1899, p. 4) that the ventral soft parts of Ischnochiton, especially

the proboscis, might be sensitive to light, and it has been stated

in a general way (Pelseneer, 1906, p. 50) that chitons in w

there are no obvious shell ‘‘eyes’’

tive to illumination.

~_—

Lich

seem, nevertheless, to be sensi-

Direct proof has, however, been given by

Arey and myself (1919) that the tegmental esthetes of Chiton

tuberculatus are photosensitive, and that this form of irritability

is important in determining the habitat of an individual chiton.

C. tuberculatus attains a mean age of about 8 years (Crozier

1918a, 19185). As it grows, the periostracum and the surface of

A B c

PIG.

Illustrating the relative lengths,

in three groups of Chit

ages and degrees of shell-erosion

on tuberculatus taken from localities quite near tegether,

u n the different situations indicated. The extent of erosion seen in each

specimen was estimated by comparison with a ‘les ;

chitons: A, signifying no erosion; B, slight erosion; C, a more severe

mild in comparison with D and E, the last representing relatively

extreme

destruction of the tegmental, surfaces.

the tegmentum become eroded, and the superficial photoreceptive

organs destroyed (Arey and Crozier, 1919). There is an almost

perfect correlation between the degree of this erosion and the

relative illumination of the situations frequented by these chitons

The analysis of this state of affairs, and its implications, will be

more fully considered in another place, but as an illustration I

cite the following record, which is quite typical of many others:

July 1, 1918. At the northwest end of Marshall, Idaho, a small cove

278 THE AMERICAN NATURALIST [Vou. LV

bounded on the west side by rocks exposed to the sunlight of a cloud-

less day; the cove covered by loose slabs of rock, piled upon one

another; both situations, the exposed rock surface and the under sides

of thé loose stones, yielded a number of chitons

From the sunny rock surface, 38 individuals were obtained, ranging

in length from 5.5 to 8.9 ems., and in estimated age (Crozier, 1918a)

from 5 to 11 years. These were without exception seen to have the

tegmenta eroded to a greater or less extent.

From several crevices in the rock, having their deeper recesses well

sheltered from the sun, 8 specimens were secured. These were 4.5-6.9

ems. long, 4-6 years old, and but slightly eroded.

Under the stones, 30 chitons were collected. There were 3.4-6.9

ems. long, 2-7 years old, and at the most very mildly eroded.

Fig. 1 exhibits, for these lots, the distribution of: (1) length,

(2) age, and (3) degree of erosion. The ‘‘degree of erosion’’ of

the shell plates was judged empirically by comparison with a

graded series of ‘‘standard’’ examples, a method sufficiently pre-

cise for the purposes of this illustration.

It is obvious from the figure that those chitons with relatively

uneroded shells are younger, smaller, and live in darkened situa-

tions; whereas the older individuals, larger, with much eroded

shells, occur in the bright sunlight; those taken in partially il-

luminated cracks and crevices are of intermediate size and age,

and their shells exhibit an intermediate degree of erosion,? these

characteristics affording, in fact, an ethological definition of a

certain portion of the chiton population. The individuals of this

general ‘‘intermediate’’ class are- frequently so situated that

they exhibit the type of ‘‘nocturnal’’ activity noted by Heath

(1899) for several species—they creep over open rock surfaces,

feeding, at night, and may remain out there on dark and gloomy

days, but return to crevices (or to the partial shelter of large

stones) when the rising sun is of ordinary brilliancy.

The uneroded chitons are photonegative to moonlight, even,

and although moving about actively at night, do not provide data

bearing on the possibility of ‘‘homing.’’ This is largely true be-

cause the photonegative response of the younger, uneroded, chi-

tons becomes altered, depending upon the destruction of the teg-

mental receptors, in the directions of a photopositive reaction to

ordinary daylight; this alteration depends, not on age, but upon

2 There are several methods of estimating rather precisely the exact amount

of erosion in any given case. In a later paper these methods are fully made

use of in analyzing the observed distributional occurrence of the Chitons.

No. 638] SHORTER ARTICLES AND DISCUSSION 279

the degree of intactiveness of the photoreceptive apparatus, and

its importance in the present connection is due to the fact that

the chitons of intermediate age, eroded to a moderate degree,

and less photonegative* than the smaller ones, come to occur in

places where moderate illumination prevails. Moreover, they

move about much more freely than the younger individuals, thus

often getting some distance away from loose rocks, not plentiful

along many stretches of shore line. Crevices of one sort or

another, or shaded depressions, are therefore the one type of

refuge open to them.

3. It was found, by observing a group of marked chitons each

day for a month, that the older ones and those of the ‘‘inter-

mediate’’ group do not wander readily from place to place

(Arey and Crozier, 1919). Another group containing 14 chitons

of the ‘‘intermediate’’ class as above defined was under daily

observation for three months, and the behavior of this group

strongly suggested a feeble kind of ‘‘homing’’ phenomenon. The

animals concerned spent most of the day under a boulder at

half-tide. At night they crept out for distances of not more

than a meter, feeding on Enteromorpha and other algæ. With the

rising sun, they retreated to the rock-shelter. If the tide were

out at sunrise, they remained more or less fixed until again cov-

ered by the sea, then moving toward the rock.

This sort of behavior, regularly and constantly exhibited,

seemed to represent perhaps the incipient stages of a kind of

‘‘homing instinet.’’* The movements of the individuals of this

group were therefore carefully watched. The bit of shore con-

3 As worked ee in a previous report (Arey and Crozier, heh the partial

destruction of the photoreceptive apparatus through erosion is a principa

factor peptone lower stimulating power of light of a given intensity;

in aie Chiton is actively photopositive to weak light, negative to in-

tense light; with advancing age, therefore, the threshold for photonegativity

becomes highe er.

t Data regarding the preety of a single Chiton favorably situated on

the side of a wharf and watched continuously for 9 months are given else-

where (Arey and ond 1919). This animal was quite old, and very

Tms an n

PRERE and remained hidden. Several other cavities were available, but

this particular one was automatically encountered as the iton moved

shoreward on the. wharf-side and away from the more exposed outer edge

of the wharf. One wonders what the anectodotalist would make of a case

like this

280 THE AMERICAN NATURALIST [ Von. LV

cerned faced in general southeastward, and the tiny platform

over which the chitons crept while feeding was so oriented with

reference to the rising sun that the photonegative orientation of

the animals and their subsequent creeping brought them for the

most part automatically back to the shaded hollow under the

rock. But I noticed repeatedly that in some instances the mol-

luses moved at an angle of 30°-40° across the direction of the

sun’s rays, moving more or less directly toward the rock. If

such individuals were suddenly detached (with the aid of a cold-

chisel and hammer, removing the animal still fixed to a bit of

stone), and so placed as to necessitate its approaching the rock

at a different angle, it usually did so without trouble. If re-

moved to a greater distance than 1.5 meters, no return was ef-

fected, the creature taking up a more or less permanent site in

another shaded hollow.

Aside from light, it must be remembered that there are other

possible directive agencies in such a case. The sea was rarely

still, and even a slight tidal current would be sufficient to reflect

pressure waves from the shore,—so that, perhaps by this means

in part, a chiton would be oriented toward shore, and thus, in

the present case, inevitably toward its rock-pocket. Additional

specimens of the general ‘‘intermediate’’ group, brought from

distant islands, were ‘‘planted’’ in this community, and engaged

in the same nocturnal wanderings and early morning returns.

4. Without further analysis, the activities of a group of Chi-

tons such as that described, may seem to involve a sort of be-

havior resembling the well-known ‘‘homing’’ of Fissurella® and

its allies. Yet with Chiton the matter is clearly less definite than

in other instances recognized among molluses, and, so far as I

have seen, the facts may readily be interpreted in terms of im-

mediate directive stimulation. There is nothing necessarily

specific about the Chiton’s ‘‘home.’’ For this very reason such

homing movements as Chiton may exhibit at a certain period of

5 I encountered a curious instance of the ‘‘local habitation’’ affected by

limpets, when examining the chitons e a reef on the south shore of Ber-

muda these chitons bore on its back a small Fissure`la, the margin

of the sali ot which had become so aetna that it fitted nicely one par-

ticular spot on the much curved surface of the third valve of the chiton.

Under water, the Fissurella wandered over the eroded shell of the large

chiton, feeding upon epiphytic growths, but always returned to its ‘‘home.’’

Chiton was 9.3 cms., the Fissurella 0.9 em. lon

No. 638] SHORTER ARTICLES AND DISCUSSION 281

its life-history may be taken to represent one extreme in the

development of such behavior among molluses, seen perhaps in

its highest condition in Octopus (cf., e.g., Cowdry, 1911).

W. J. Crozier

PAPERS CITED

Arey, L. B., and Crozier, W. J.

1918. The jad epi of bay pulmonate mollusk Onchidium.

Pr at. d. Sci., Vol. 4, pp. 319-321.

1919. The patie paies sx Chiton. Jour, Exp. Zool., Vol. 29,

pp. 157-260.

Cowdry, E. V.

1911. The Color Changes ge pe pe vulgaris Lmk. Univ. Toronto

Studies, Biol. eer . 10, 53 pp. (Contrib. Bermuda Biol.

Sta., Vol. 2, No. os

Crozier, W. J.

1918a. Growth and Duration of Life in Chiton tuberculatus Linn.

-Pro . ci., Vol. 4, pp. 322-325.

19186. Growth of Chiton in Different Environments. Ibid., pp. 325-

1919a. On pr Use of at ae in Some Molluses. Jour. Exp. Zool.,

Vol. 27, pp. 359-

1919d. ele a on the see Pence of the Chitons. AMER. NAT.,

Vol. 54, pp. 376-380,

Crozier, W. J., and Arey, L. B.

9 On the Significance of the Reaction to Shading in Chiton. - Amer.

Jour. Physiol., Vol. 46, pp. 487—492.

Heath, H. ;

1899. The 7 of Ischnochiton. Zool. Jahrb., Abt, Anat.,

Bd. 12, pp. 1-90.

1904, ges Tars al a of rE Proc. Acad. Nat. Sci., Philadelphia,

1. 56, pp.

Pelseneer, K.

1906. Mollusca, in: Lankester, Treatise on Zoölogy, Pt. V, 355 pp.,

London.

Plate, L.

1901. Die Anatomie und Phylogenie der Chitonen (Theil C.) Zoot

Jahrb., Suppl. Bd. 5 (Fauna Chilensis, Bd. 2), pp. 508-600

AN F, To CROSS BETWEEN HORDEUM VULGARE

D HORDEUM MURANIUM

Wirnin the last few years the subject of species hybridization

has increasingly occupied the attention of those interested in

the subject of heredity. The possibility of the genetic analysis

of species hybrids depends upon the ability to cross and to

secure offspring from the species in question. During the course

of a plant-breeding investigation which was commenced at the

282 THE AMERICAN NATURALIST [Vou. LV

University of California, an attempt was made to determine if

the common cultivated barley could be crossed with wild species

of Hordeum. The wild species which were used were H. nodosum

and H. muranium. For the sake of convenience the system of

nomenclature proposed by H. V. Harlan (1918) will be adopted

or the common varieties of barley used in the investigation.

One of the crosses which was attempted was between H. vulgare

vulgare pallidum and H. nodosum. This cross was an entire

failure, however, as no seeds were obtained from any of the

flowers which were crossed. The anthers of the male parent

were fully mature, and the plant which was used for the female

parent was perfectly healthy and normal when the cross was

attempted. As a matter of fact a successful cross was made

the same day between a different head of the same plant of

H. vulgare vulgare pallidum and H. vulgare distichon palmella.

At the present time it would be difficult to say whether the

absolute failure of this particular cross was due to the incom-

patibility of the gametes of the two parents or to certain errors

of technique.

The other cross which was attempted was between H. vulgare

trifurcatum typica and H. muranium. The contrast between the

two parents was very marked and distinct. The low and often

recumbent habit of growth of muranium was contrasted with

the relatively tall and erect habit of vulgare. The light green

leaves and stems of muranium were not nearly as stout as the

gray green leaves and stems of vulgare. The spikes of the two

species are also quite distinct. The spikes of muranium are

compressed and composed of a number of rather narrow elon-

gated spikelets which form rather a loose head. The spikes of

vulgare, on the other hand, are generally composed of a number

of relatively short and wide spikelets. Both species are an-

nuals, but without going into, further detail it is evident that

there are a large number of morphological differences between

these two species.

From the second cross two viable hybrid seeds were obtained.

These grains resembled the typical seeds of the maternal variety

in every respect. When they were planted, however, it required

one and three days longer for the seed to germinate than for self-

fertilized seed of the female parent.

The F, seedlings differed markedly from plants of H. vulgare

in the same stage of development. The sheath or coleoptile had

a greater circumference than the blade, thus fitting loosely

No. 638] SHORTER ARTICLES AND DISCUSSION 283

around it instead of adhering closely to the blade as in vulgare.

The sheath was closed along the side and open only at the apex.

The blade of the first leaf was narrow, linear and spirally

twisted with slightly roughened edges. The blade was about

one twelfth of an inch in width and tapered slightly toward the

apex (Fig. 1).

One plant grew to a height of four inches and developed roots

two to three inches long (see Fig. 2). The other plant devel-

oped somewhat more slowly, reaching a height of two inches

Fic. 1. A first generation hybrid between Hordeum vulgare trifurcatum typica

and Hordeum muranium,

with roots of the same length. At this stage the plants ceased

development and gradually started to wither. Only one blade

was present and this extended to the seed. There was no evi-

dence of any nodes being formed.

Due to a change of residence the writer has not been able to

continue the investigation for a short time, but it is hoped that

this cross may be subjected to further breeding tests as well as a

histological examination. The theoretical hypotheses concern-

ing species crosses have been thoroughly reviewed by other

writers (Babeock and Clausen 1918), but it may not be out of

place to briefly state the particular theories which probably

account for the results considered in this paper

It has already been pointed out that H. vulgare and H. mu-

ranium differ in a large number of morphological characters.

284 THE AMERICAN NATURALIST [Vor. LV

H. muranium may be considered as a monotypic species, and

the slight variations which are found in the species are un-

doubtedly due to the effects of the environment and would be

classed as non-heritable variations. H. vulgare, on the other

hand, is a polytypic species consisting of many varieties which

differ in a number of morphological characters. Most of the

factors which condition the characters of vulgare display har-

Fic. 2. F, species hybrids between Hordeum vulgare and Hordeum muranium

at the stage of growth at which development ceased.

monious interrelations with one another and mendelize in a

normal fashion. Several factors involving chlorophyll reduc-

tion have been discovered, however, and these genes have been

found to be incompatible with the normal functioning of the

chromatin system. In these cases after the food material in the

seed has been exhausted the seedlings usually die, for the change

No. 638] SHORTER ARTICLES AND DISCUSSION . 285

in the factors has been too far reaching to give a normal func-

tioning reaction system.

This brings forth the theory of reaction systems which has

been thoroughly reviewed by Goodspeed and Clausen (1917a).

The purpose of the discussion in the preceding paragraph was

to show that both vulgare and muranium possess a normal re-

action system, and second that a normal reaction system may

sometimes be disturbed by lethal factors. When we attempt to

combine two distinct reaction systems, however (and the dis-

tinet morphological characters of the two species as well as the

breeding test would indicate that the two species possessed

different reaction systems) an inharmonious group is often

formed which fails to function in a normal fashion. In the case

of chlorophyll reduction there is one or at most only a few

factors disturbing the reaction system. In the case of species

crosses there are a number of factors, which in all probability

differ qualitatively, coming from two distinct reaction systems

and these often fail to harmonize. The results are often similar,

however, for the differences between the reaction systems of

vulgare and muranium are so profound that the resulting system

is not able to function after the food material in the seed is

exhaust

The oe of species cross described in this paper is quite

similar to the species cross between Crepis capillaris and Crepis

tectorum recently described by Babcock and Collins, 1920. The

two species of Crepis, besides differing in several morphological

characters, were found to differ in chromosome number. Recip-

rocal crosses gave equivalent results, or the dominance of

tectorum cotyledon characters in F, accompanied by hybrid

vigor. The F, seedlings died, however, in every case at the end

of the cotyledon stage. Cytological examination revealed a com-

plete lack of order in the cell systems, and as a result these

systems failed to function and development ceased. The species

cross in barley involves slightly greater contrasts perhaps than

those in Crepis but both give nearly parallel results.

There are, as has been pointed out by others, all degrees of

incompatibility of reaction systems in species crosses. The

range of compatibility includes cases of complete or nearly com-

plete fertility, as in the species crosses in Antirrhinum (Baur

and Lotsy), examples like those found in Nicotiana (Goodspeed

and Clausen 1917b) where the fact of incompatibility does not

become evident until the fertilization of the F, plants, and

286 THE AMERICAN NATURALIST [Vou. LV

finally we have species which exhibit complete incompatibility

by refusing to cross with one another. The range includes

many intermediate conditions like those found in Crepis, which

nearly approach complete incompatibility. The cross between

H. vulgare and H. muranium, then, is well down the scale and

ean be grouped in the class with the two species of Crepis as

showing nearly complete incompatibility.

WALTER Scorr MALLOCH

AGRICULTURAL COLLEGE OF TEXAS

NOTATIONS

Babcock, E. B., and Claus

1918. Genntios in Relation tù Agriculture.

Babcock, E. B., and Collins, J. L.

1920. Yitapiaie ai in Crepis. Univ, of Calif. Pub. in Agri.

paka a ON ES '

Baur, E., and Lotsy, J. P.

1911-12 Pape on Antirrhinum; see review in Babcock, E. B., and

Clausen, R. E., Geneties in Relation to Agriculture

Goodspeed, T. H., ani Clausen, R. E.

19174. Mendelian Tao pieis aen Reaction-System Con-

rasts in Heredity. ER. NAT., b

Goodspeed, T. H., and Clausen, R. g

1917b. pes Nin of F, Species elie between Nicotiana Sylves-

ris and Varieties of Nicotiana Tabacum with Special Refer-

ence to the pha aay > Seca -System Contrasts in He-

redity. Univ. of Calif. Pub, in Bot., Vol. 5, No, 11.

Harlan, H. V.

1918. The Identification of Varieties of Barley. U. S. Dept. of Agr.

Bull., 622.

A NOTE ON UNILATERAL REACTIONS OF THE MELAN-

OPHORES OF THE HEAD IN FISHES

In most discussions of the physiology of the chromatophores

of fishes it is apparently assumed that the reactions are strictly

bilateral, i.e., synchronous on the two sides. The writer, how-

ever, has lately observed a number of cases in which the reaction

was either unilateral or imperfectly bilateral.

Upon death, the melanophores of one side of the head in some

eases become all ‘‘contracted’’ to the extreme, while those of the

other side become widely ‘‘expanded.’’ As a result, one side of

the head becomes very pale, the other side blackish, the two areas

being abruptly opposed along the mediodorsal line. This notable

No. 638] SHORTER ARTICLES AND DISCUSSION 287

color change at death has been observed by the writer in an

adult pike (Esos lucius); in a young-of-the-year of the shiner

(Notropis cornutus); and in embryonie and larval whitefish

(Coregonus clupeaformis) and lake-herring (Leucichthys ontari-

ensis).

Dorsal Aspect of Head of an Embryonic Whitefish ab ag to Illustrate

the Unilateral Reaction of the Melanophor

This phenomenon is not confined to death, however, as the fol-

lowing observations demonstrate. A nuptial male of Pimephales

notatus (a minnow in which the head becomes densely charged

with black pigment during the breeding and nesting activities),

apparently normal in respect to its eyes and other structures and

functions, found guarding its eggs, had one side of the head

abruptly pale. Similarly embryonic and larval coregonine fishes

were repeatedly observed to have the melanophores expanded

only on one side of the head during life. In the case of the male

Pimephales, no change in the pigmentation of the head was noted

while the fish was being observed for several minutes, nor upon

its capture, death or preservation.

288 THE AMERICAN NATURALIST [Vou. LV

In other cases the unilateral reaction of the melanophores was

less permanent, appearing as a transient phenomenon; due per-

haps, to a differential reaction rate of the chromatophores of the

two sides. Two experiments! illustrative of this point may be

cited.

1. A live, normal, apparently healthy embryo of the lake white-

fish (Coregonus clupeaformis), developed approximately to the

hatching stage, was found to have the dorsal melanophores con-

siderably ‘‘expanded’’ on the head, slightly expanded on the

body. Following the removal of the egg envelope, under ap-

proximately unchanged conditions, these color-cells ‘‘contracted’’

in this order: (1) body, (2) left side of head, (3) right side of

head. Still under similar conditions, the cephalic melanophores

again expanded, those of the left side most widely. No further

change could then be induced, even by rather intense light-heat

stimulation, until the left eye was dissected out and the embryo

again held before the light. Reaction occurred at once only on

the left (now the blinded) side, the lateral melanophores con-

tracting more rapidly and more completely than the inner ones;

as in the first instance, reaction followed (some time after the

removal of the stimulus) on the right side, = melanophores

contracting in the same order as on the left si

2. A similar embryo of the same species had P dorsal melano-

phores of the head well expanded when removed from its egg

envelope. The pigment granules of all melanophores on the right

side then rapidly migrated into the center of the cells, under

observation. No reaction occurred on the left side, even follow-

ing stimulation with a bright light, although this caused first a

partial contraction and then a re-expansion of the right chroma-

tophores. Reaction on both sides was finally accomplished. by

sudden transfer of the eggs from water near room-temperature

(about 25°) to water at 1.7° C., but even in this case the contrac-

tion was more complete and rapid on the right than on the left

side.

CARL L. HUBBS

UNIVERSITY OF MICHIGAN

1 These experiments were incidental to other studies which the writer car-

ried on during the winter of 1919-1920 in the bionomics laboratory of the

University of Chicago; he desires to thank Doctors Lillie and Bellamy of

m institution for the opportunity they kindly afforded him to do this

work,

THE

AMERICAN NATURALIST

Vou. LV. July-August, 1921 No. 639

FURTHER DATA ON THE INHERITANCE OF

BLUE IN POULTRY?

PROFESSOR WILLIAM A. LIPPINCOTT

Kansas AGRICULTURAL EXPERIMENT STATION, MANHATTAN, Kansas

I. Previous Work

The principal facts concerning the genetice behavior of

blue in the Andalusian breed of domestic fowl were pre-

sented in an earlier paper (Lippincott, 1918a). Pre-

vious work on the genetics of the blue Andalusian was

reviewed and a limited number of further data were

offered.

The latter showed that blue Andalusians are like black

Andalusians in that they are self-colored. They are, on

the other hand, like the blue-splashed Andalusians in that

homologous pigmented feathers in both sexes have the

same condition with reference to the restriction of pig-

ment in the feather structure. The 1:2:1 ratio obtained

from mating blue Andalusians together may be inter-

preted as the combination of two 3 to 1 ratios. These

relationships are shown in Fig. 1.

The restriction of black pigment in the feather struc-

ture to give the blue appearance found in blue and in

blue-splashed Andalusians was shown to be due to the

action of a dominant factor R. The extension of black

pigment to all feathers of the body as in both black and

blue Andalusians, was found to be due to the action of

another dominant factor E.

1 Contribution from the Sees wR of Genetics, Wisconsin oh, ee tag

Experiment Station, No. 29, and from the Department of Poultry Hus

bandry, Kansas peatas Experiment Station, No. 15,

289

290 THE AMERICAN NATURALIST [Vou. LV

| BLACK

3 BLUE

| NON-SELF

Fic. 1. Showing that the 1: 2:1 ratio is a combination of two 3:1 ratios.

No. 639] INHERITANCE OF BLUE IN POULTRY 291

It was pointed out that while, on the basis of their ex-

pression in the phenotype it appeared more logical to

consider these factors as dominants, each closely linked

to the recessive allelomorph of the other, they may, so

far as the experimental evidence shows, be considered as

true allelomorphs occupying identical loci on homologous

chromosomes, and each expressing itself independently

of the other. :

The finding of crossovers between- R and E would be

conclusive evidence proving the former of the two con-

ditions proposed. It was shown that while no crossovers

had been reported the critical data on the case were very

limited and the likelihood of crossovers being detected

and isolated by breeders is very small. It might well

- have been suggested further that even though crossing-

over does rarely occur, for instance, so that less than one

per cent. of the individuals are the product of crossover

gametes, the chances of detecting them experimentally

are small, considering the limited number of matings

(as determined by the equipment available at most ex-

perimental institutions) which are likely to be devoted to

a search for crossovers.

Though much has been made of the blue Andalusians

as a ‘‘heterozygote phenotypically intermediate between

the parental types’’ it was shown that while all self-blues

so far found had proved to be heterozygous for R and E,

they were not in the strict sense intermediate between the

parental types. The F, progeny of a cross between blue-

splashed Andalusians and white Wyandottes was re-

ported as self-blue and far darker than either parent.

It was further shown in the earlier paper that R not

only restricts black pigment, so as to render pigmented

areas bluish-gray in appearance, but also affects the

shape of the pigment granules, so that instead of appear-

ing as rods as in black individuals, they are quite round.

In this particular R is quite dominant over its allelo-

morph, whether one chooses to assume that the latter

is E or r.

292 THE AMERICAN NATURALIST [Vou. LV

EXPLANATION OF PLATE I

he phe Net oi wn in Plates I and II were taken by James Machir, my

a aa to whom it is a pleasure to acknowledge.

Fic. A. Blue ROSEE an ms = Fia. F. Blue-splashed Andalusian

Fic. B. Blue Orpington fem male.

Fic. C, Blue- splashed icaid > and D—Blue een ın.

female. nd #H—Blue Orpingt

Fig. D. Blue Andalusian female, p kai F—Splashed PESAR

Fic. E. Blue Orpington male.

No. 639] INHERITANCE OF BLUE IN POULTRY 293

It was also shown that both the restricting and the

rounding actions of R were interfered with in certain

regions of both blue-splashed and blue males. In both

color types the pigmented feathers of the neck (hackle),

back, and saddle are black or bluish-black instead of blue

as on the remainder of the body. The black pigment

granules in these regions are for the most part rod-

shaped rather than round. It was suggested that this

interference with the action of R is a secondary sexual

characteristic, presumably due to the presence of tes-

ticular or the absence of ovarian influence.

II. PURPOSE or THE PRESENT PAPER

It is the purpose of this paper to present further data

concerning the inheritance of blue and its relations to

the sex glands, and to draw such conclusions as these data

justify. A report is given of the breeding behavior of

blue as found in the Andalusian, Orpington and Leghorn

breeds, and of certain crosses of these breeds with each

other and with other breeds, which do not possess blue

varieties. The relations of the factors involved to cer-

tain factors present in the non-blue varieties of other

breeds is considered and evidence concerning the relation

of the sex-glands to the action of the factor R presented.

Ill. MATERIAL anp METHODS

The breeding stock used was from several sources,

being in part from the pedigreed flock of the University

of Wisconsin, where the work reported in the earlier

paper was done. It was also in part from the pedigreed

flock of Kansas State Agricultural College where the in-

vestigation was continued under the direction of Dr. Leon

` J. Cole of the University of Wisconsin, my indebtedness

to whom it is a pleasure to acknowledge. The stock was,

however, mostly from unpedigreed lines, though pure-

bred within the meaning of the poultryman. In no case

were individuals used which were not from families show-

294 THE AMERICAN NATURALIST ` [Vou. LV

EXPLANATION OF PLATE II

Fic, A. A black Andalusian male

Fic. B. A blue F, female passe a white Wyandotte x blue- Meen Andalusian

O D = (at a and a sy oh right) chick in the down. Th

are UnA of a whi Se th Roc blue Andalusian 9. The alpini

spots inherited Pt a sire are aus S vis ale

ric. D. A young blue F, male from a blue-splashed x black Langshan cross.

f la ale.

Fic. F. A blue F, male from a white Wyandotte x blue-splashed Andalusian

No. 639] INHERITANCE OF BLUE IN POULTRY 295

ing the characteristics of their respective varieties with

constancy in so far as could be learned. In as much as

only varietal (color), as opposed to breed (shape) char-

acteristics were being studied, less attention was paid to

the latter in selecting material. In no case, however,

were individuals used which showed disqualifying breed

characteristics.

With a single exception no individual was used whose

genotype proved to be inconsistent with the ‘‘breeding

true’’ of the variety to which it belonged, or, in the case

of the blue-splashed Andalusian, the variety from which

it arose. This single exception was a blue-splashed

Andalusian female (2107) purchased from a breeder who

made only blue X blue matings. She proved to be

heterozygous for P, a factor necessary for the production

of black pigment. The family from which she arose

must have been producing occasional whites which were,

in all likelihood, being discarded as extremely light blue-

splashed wasters from the blue X blue matings. This

point was not followed up, however, and the facts ascer-

tained. It has been by taking advantage of situations

similar to this one that white varieties have been estab-

lished in several breeds.

There were several individuals discovered whose fac-

torial composition varied from the normal, or usual, for

the varieties to which they respectively belonged. Owing

to the particular factorial complex of which they were a

part, however, these factors behaved as cryptomeres, not

affecting the adult phenotype of the variety. Specific

reference is made to these individuals in a later section

of this paper.

The matings were, for the most part, made in covered

cross. Indications of a factor or factors for lacing may be seen in the hackle

and saddle feathers.

Fic. G. A blue-barred (at left) and a black-barred (at right) chick partly

feathered. These are offspring of the same mating as the chicks shown in Fig. Q

this plate. The barring was inherited from their white Plymouth Rock sire.

A and E—Black Andalusian.

B and F—Blue F,’s.

296 THE AMERICAN NATURALIST [Vou. LV

o0 90000

. TS a. big mer

> :

i ’ E

“a

j 2 «* Sag

a. Peas "e ef

aa 5900 (E2385 ose z

aes F

i tad

aid

i In

i 4 i

B o

Ses,

nas wam Y >

e Spr to o* PO .: Os Sone

apd xe SN te eos oer

EXPLANATION OF PLATE III

Fic. A. Appearance of pigment granules in a stra ry

dark bide crossbred chick. The sire was sad a white yarra x

blue-splashed Andalusian cross. was a a Andalusian. There

occasional rod-shaped granul patter eld draw

Fig. B. Appear:

of down from a ve

a on ue Fy

aun feather from a blac

s of the same pee are town in the down and definitive

Langshans, black Orpingtons and the hace offspring fr

the several breeds used in this investigation. Camera lucida

Fic. o. Appearance of pigment granules in a strand of down from a blue

No. 639] INHERITANCE OF BLUE IN POULTRY 297

yards and every precaution taken to insure no mixing of

matings and the proper identification of the eggs laid by

each female in each mating. Not only was the assistant

in charge of the trapnesting selected because of his habit-

ual accuracy in details, but the eggs from each individual

hen were kept together, separate from the eggs of other

individuals, and carefully compared one with another

before being put into the incubator. Any off type or off

colored eggs were discarded, so far as these experiments

are concerned. In spite of these precautions it is too

much to hope that some errors have not crept in, though

it is believed they are very few.

Owing to the fact that the original stock was of rela-

tively unknown composition, it was necessary to make

such matings as would not only throw light on the

behavior of the factors under observation, but would also

be likely to bring to light unsuspected factors whose

action might interefere with the action of the genes

being studied. This necessitated introducing test females

in the matings where the males were uncertain, and of

mating many of the females with test males a second

season instead of repeating the mating already made.

In both eases the result was to reduce greatly the numbers

of offspring from some of the crucial types of matings,

and considerable numbers of ‘‘test’’ offspring were

hatched and described, for the reporting of which here

there is no particular object.

The counts of living chicks were made in the down at

hatching time and individual descriptions recorded, each

F, chick from a white Wyandotte x blue-splashed Andalusian cross. Camera

anA

re draw fan

s s mped appearance of pigment granules in a curved barbule from a

aesintive exch of a blue sprang sian,

arance of ent granules in a strand of down from a blue Ad

Fg pe a “binesplas pig rere tei x white Plymouth Rock cross. Cam

lucida drawing

Fic. F. Appe arance of p peers granules in a definitive feather from a blue

PPE Camera lucida draw

Fic. A small area of the Ste of a definitive feather from a black Anda-

lusian ks cell boun a and nuclei may be made out. There is no clumping

of pigmen: ent within the cells.

Fic. H. Appearance of pigment granules in a definitive feather from a blue

Andalusian. Camera lucida drawin

298 THE AMERICAN NATURALIST [Vou. LV

chick being marked with a numbered wingband. The

system of keeping pedigree records in use has been de-

` scribed elsewhere (Lippincott, 1918b) and need not be

repeated. The descriptions were checked when the chicks

were three weeks old, and again at some considerably

later, though not specified, time, when the birds were leg-

banded for the breeding-pen or laying-house, or were sent

to market. The descriptions of all chicks dying were

carefully checked at the time they were found, though a

small number disappeared without their descriptions

being checked. Unless there was reason to suspect that

their classification might be likely to change after the

taking of the first description such individuals were

counted.

Fortunately the different classes of offspring could for

the most part be distinguished in the down, and counts

were accordingly made of chicks which reached an ad-

vanced stage of development but which failed to hatch.

It was the practise to test all eggs for live germs at the

end of the tenth day of incubation and remove all infertile

and dead eggs. A second testing was made on the eight-

eenth day when all the dead eggs found were opened, the

embryos described and their sex recorded. On the

twenty-second day, after the hatch was well over, all the

eggs which failed to hatch were opened and the descrip-

tions of the dead chicks made a matter of record.

In most cases the embryos from crosses among the

three color types of Andalusians which passed the first

test, developed far enough so that the differentiation

between color types could be made with precision after

the down had been carefully washed, and dried with the

aid of an electric fan. In those crosses involving reces-

sive white parents, only those unhatched chicks could be

counted which lived past the eighteenth day.

There were two possible sources of confusion in the

classification of the chicks in the down. These were the

differentiations between blacks and occasional very dark

blues, and between blue-splashed and recessive whites.

No. 639] INHERITANCE OF BLUE IN. POULTRY 299

The blacks and dark blues could quite readily be sepa-

rated by examining the down of each chick microscop-

ically. The blacks carry only rod-shaped pigment

granules while in dark blue down rounded granules pre-

dominate. These are frequently arranged in rows as

reported in my former paper (1918a, pp. 98,99). In the

case of every mating where this method of classification

was brought into use to aid in distinguishing individuals '

which failed to hatch, down samples were saved at hatch-

ing time from the living dark blue and black chicks as

well, and the first description, the record of the micro-

scopic examination, and the later descriptions after

definitive feathers were developed were carefully com-

pared and checked. In nearly all cases the descriptions

in the down and in the definitive feathers agreed.

In certain matings, however, it was found that descrip-

tions in the down were not reliable and could not be

counted. This was particularly true of one family of

Andalusians which. carried considerable red in the

plumage and which has been referred to by Platt (1916)

and Pearl (1917) in other connections. Within this fam-

ily and its crosses the expression of the R factor in the

heterozygote was frequently delayed so that individuals

which were described in the down as blacks and which

showed only rods under the microscope turned out to be

blues when the definitive feathers appeared, and then

showed the characteristic round pigment granules of the

blue. None of the chicks tracing their ancestry to this

family are included in the:counts herein reported.

The possible source of confusion in the classification in

the down of the blue-splashed and the white chicks arises

from the fact that while the adults of the white Plymouth

Rocks and white Wyandottes are pure white, or very

slightly flecked with black, the chicks frequently carry

considerable, though varying amounts of black pigment

in the down, which gives certain regions a bluish appear-

ance. This varies in degree from near black to slightly

300 THE AMERICAN NATURALIST [Vou. LV

smoky white. Fortunately for the problem in hand the

localization of this pigment in the down of certain regions

of the body is quite characteristic and quickly recognized.

While a blue-splashed chick is frequently very light blue,

as noted by Bateson and Punnett (1906, p. 20), the pig-

ment is not localized on the top and back of the head, the

wings in the region of the bow, and on the thighs, as it

is on the potentially white chick, and the impression con-

veyed is very different. In potentially white chicks the

remiges, which may be seen just starting to grow out

from their follicles, are pinkish white and exhibit not the

slightest trace of pigment. In the same feathers of the

blue-splashed chick, on the other hand, there is a very

noticeable bluish cast and usually at least one remex that

is distinctly pigmented.

Though in pure-bred white Plymouth Rock and white

Wyandotte chicks the pigment granules in the down are

typically rod-shaped this fact is not of assistance in

classifying with respect to white and blue-splashed off-

spring from crosses involving the factor R, since under

its influence black pigment granules are round whether

in a potentially white or a blue-splashed chick.

Not all chicks from pure-bred white Plymouth Rock

and white Wyandotte matings exhibit this juvenile pig-

ment. Some can only be recorded as white. It is of

interest that the only chicks, three in number, which were

originally described as ‘‘ white, no pigment” or ‘‘creamy

white’’ and later used in a breeding pen, have all proved

to carry a factor for dominant white, as described in a

later section of this paper. The number of such birds

which have been tested is small and no general conclu-

sions can be drawn, but the results are suggestive. It is

rather interesting to note that a photograph of a group of

white Plymouth Rock chicks in ‘‘The Plymouth Rock

Standard and Breed Book’’ (American Poultry Associa-

tion, 1919, p. 419), which is the official guide for the breed-

ing and judging of all Plymouth Rocks, shows individuals

No. 639] INHERITANCE OF BLUE IN POULTRY 301

which are noticeably pigmented. In response to a letter

of inquiry Professor Arthur Smith of the University of

Minnesota, the editor of this book, tells me that my ob-

servation concerning the presence of pigment in these

chicks is correct and he adds in substance that the pig-

mented chicks develop into the whitest adults.

The fertility and hatching power of the eggs from the

various crosses here reported and the viability of the

chicks hatched was increasingly disappointing from

season to season. While the comparative coefficients of

fertility and hatching power have not been calculated, the

ratio between the eggs set and chicks hatched has un-

doubtedly been lower on the average, than for the pure-

bred unrelated matings of the same and other breeds, set

in the same incubator at the same time, and certainly

lower than would be counted satisfactory in ordinary

poultry husbandry practise.

The foregoing applies as well to the rate of mortality.

As representative of the numbers surviving to grow

definitive feathers in comparison with the counts re-

corded in the various tables, those of the F, from the

blue-splashed Andalusian J X white Wyandotte 2? may be

given. The counts made when the chicks were feathered

were 47 blue, 18 blue-splashed, 37 black, and 42 white.

The total count recorded (see Table IV, group 1) was

100 blue, 46 blue-splashed, 65 black, and 64 white. The

reasons for the low hatchability and high mortality have

not been established.

Until considerably more data than are now available

have been secured it seems best to call attention to the

possibility of crossing-over between the loci of R and E

by indicating their possible recessive allelomorphs. It is

accordingly the practise in this paper to indicate these

factors thus: (Re) and (rE).

TV. Tue RELATION oF PHENOTYPE TO SEX

It is convenient to consider the relation of phenotype

to sex before examining the progenies of the various

302 THE AMERICAN NATURALIST [Vor. LV

matings. In order to secure evidence concerning this

relation, six blue Andalusian males were caponized dur-

ing the summer of 1919. Into the body cavities of three

of them ovarian tissue from nearly related females was

introduced, the other three being kept as checks. The

operating was done on July 24 and the birds turned out

on range with hundreds of other birds one week later.

On September 19 one of them (wingband 1387) was killed

by askunk. At that time it was entirely blue, there being

less contrast between the regions that are dark in the

male (hackle, back and saddle) and the other regions of

the body than frequently appears in blue pullets before

comb development indicates the approach of the first

laying cycle, and indeed in many, mature females.

Although it was over four months old (hatching date,

May 6, 1919) it appeared so much like an immature pullet

that it was mistaken for one by the poultryman in charge

and by the writer, until its record and description were

consulted. Concerning the latter there did not seem to

be any chance of error, since the scar made in opening

the body cavity was plainly visible.

Such a situation indicates a fairly complete molt be-

tween July 24 and September 19. This is not surprising,

however, since Rice, Nixon and Rogers (1908, p. 66) have

shown that ‘‘from the incubator to the laying period the

chicks experienced at least four molts, either partial or

complete,’’ and it is further well known that a close rela-

tion exists between molting and ovarian activity.

The other birds operated on at the same time were at

once looked up and described. One of them (wingband

1855) was found to be somewhat intermediate in condi-

tion, some of the feathers of the neck and saddle being

blue, but somewhat darker in shade than the normally (in

the male) blue regions of the body. There were, how-

ever, a few scattered feathers which were almost black

from the tip halfway down the web toward the fluff.

About midway between the tip of the feather and the be-

No. 639] INHERITANCE OF BLUE IN POULTRY 303

ginning of the fluff there was a distinct line of demarca-

tion where the black or near-black became a distinct blue.

This chick was hatched a little over two weeks later

(May 22) than 1387 and had apparently not gone through

a complete molt, some feathers in process of growth at

the time of the operation and showing ovarian influence

on the last regions to develop still remaining.

On October 26 this bird was killed, apparently by a rat,

at which time all of the feathers of the neck and saddle

regions were distinctly blue, though considerably darker

than other parts of the body. The shape of the feathers

was characteristically female. |

The third male into which ovarian tissue was intro-

duced (wingband 1480) showed no influence of the intro-

duced tissue on September 19. This condition still pre-

vailed when it was sent to market October 26. It

appeared normal for a blue capon of that age, over five

months, the hackle and saddle being very dark and char-

acteristically male in shape. Presumably the ovarian

tissue introduced atrophied without having any effect.

Of the cockerels which were caponized, but had no

ovarian tissue introduced, one (wingband 1859) died soon

after the operation. The other two (wingbands 1415 and

1492) showed and continued to show typical blue capon

characteristics with regard to the color and shape of the

saddle and hackle feathers. The feathers were fully as

dark as in normal males of the same age, and as they

matured were even longer than their homologs in normal

males. This result is precisely the same as that observed

by the writer several times in blue capons, concerning

which no descriptive records were kept.

In this connection it should be observed that in the

family of Andalusians here dealt with, it has been not in-

frequently noticed that certain nearly grown pullets

whose combs have not begun to develop, show only very

dark feathers in the regions of the neck and back. These

same birds after their combs begin to redden, thereby

304 THE AMERICAN NATURALIST [Vou. LV

indicating ovarian activity and the onset of laying, ap-

pear to pass through a molt or partial molt whereby the

dark feathers of the back region particularly, are gradu-

ally replaced by those of a clearer blue. The necks of

such females usually remain dark, showing considerable

contrast with the other regions of the body, though being

by no means as dark as the same region of the blue male.

Although the number of desexed males into which

ovaries were introduced was small, it seems fair to con-

clude in the light of the evidence concerning testicular

(Goodale, 1916) and ovarian (Goodale, 1918; Cole and

Lippincott, 1919) influence in fowls that the failure of the

factor R to express itself as fully in the neck, back and

saddle regions of the blue and blue-splashed males as in

the females is due to the lack of some necessary co-

operative action on the part of the ovary, and not to any.

inhibitive action on the part of the testis.

V. THe BREEDING BEHAVIOR or ANDALUSIANS

New data concerning the breeding behavior of the three

color types of Andalusians, as shown by several types of

matings, are presented in Table I.

TABLE I

- SHOWING THE NUMBERS AND COLOR popes OF PROGENIES FROM VARIOUS

ANDALUSIAN CROSSES2

| Blue-svl. | Blue | Black

Group To 99 (Re) (Re) | (Re)(rE) |(rE) (rE)

1..} Blue X blue Obtained...| 46 104 64

| (Re) (rE) (Re) (rE) | Theoretical.| 53.5 107 53.5

2 ue X black btained...| 00 25 24

(Re) (rE) (rE) (rE) Theoretical. 00 24.5 24.5

s. Black X blue Obtained. ..| 00 13 90

(rE) (rE) (Re) (rE) Theoretical 00 101.5 101.5

4 lue | X blue-splashed| Obtained. . . 1 1 0

e) ( (Re) (Re) Theoretioal 1 1 0

5 Blue-splashed X blue ined... 35 33 0

(Re) (Re) (Re) (rE) Theoretical 34 34 0

6 Black blue-splashed Obtained... 0 138 0,

(rE) (rE) (Re) (Re) Theoretical . 0 138 0

7 ue-splash lack Obtained. .. 0 56 0

(Re) (Re) (rE) (rE) Theoretical . 0 56 0

8 Blue-splashed X blue-splashed| Obtained... 0 12 0

e) (Re) ) (Re ) Theoretical . 0 12 0

ians are normally homozygous for P, a factor necessary for the

aeres pe black pigment.

No. 639] INHERITANCE OF BLUE IN POULTRY 305

These results are in substantial accord with those of

Bateson and Punnett (1906, p. 20). A somewhat marked

departure from the theoretical expectation appears in

group (3) of black dS X blue 22 matings, the agreement

in the reciprocal cross (group 2) being as close as pos-

sible. This departure from expectation is due to the

progeny of a single pair of birds (¢136M and 922005)

which produced 25 blues and 6 blacks. If the latter are

left out of consideration the results are 88 blues and

84 blacks. |

However, even in the case of the progeny of 136M and

92005 the Dev./P.E.= 4.1, which indicates a deviation

of doubtful significance. The results of this mating were

carefully considered from the standpoint of crossing-

over, but there is no indication of its having occurred.

According to these results the genetic compositions of

the three color types of Andalusians used in these ex-

periments were as follows: blue-splashed = (Re) (Re),

blue = (Re) (rE), and blaek = (rE) (rE). There was no

evidence of crossing-over between R and E having

occurred.

VI. DATA FROM Crosses oF ANDALUSIANS WITH

CERTAIN RecesstvE WHITE BREEDS

In the previous paper (1918a, p. 106) the writer re-

ported a small number of data on a cross between a white

Wyandotte ¢ and a blue-splashed Andalusian 2. These

have been considerably increased in amount and the re-

ciprocal cross made. Further, both blue and black Anda-

Iusians have been crossed reciprocally with white Wyan-

dottes and all three Andalusian color types crossed

reciprocally with white Plymouth Rocks. The data from

these several matings are set forth in Table II.

The crosses were made in the twelve possible ways,

from eleven of which offspring were secured, the one type

of mating which failed to produce offspring being the

white Wyandotte # X black Andalusian 9. Inasmuch as

306 THE AMERICAN NATURALIST [Vou. LV

there is no evidence that any of the factors here under

observation are sex-linked and there is considerable evi-

dence that they are not, this omission is not serious.

TABLE II

SHOWING THE RESULTS OF CROSSING THE THREE-COLOR TYPES OF ANDALU-

SIANS WITH WHITE WYANDOTTES AND WHITE PLYMOUTH ROCKS

z gta

3 aF 29 '218

ó | m a

1. Blue-splashed Andalusian X white Wyandotte btained 65 10

_PP(Re ) (Re) pp(rE) (rE) heoretical| 65 | 00

2: te Wyandotte X blue-splashed Andalusian|/Obtained 50 | 00

pp(rE) (rE) PP (Re) (Re) heoretical 50 | 00

3.|Blue-splashed Andalusian X white Plymouth Rock btained (179 | 00

PP (Re) (Re) rE) heoretical 179 | 00

4.|White Plymouth Rock X blue-splashed Andalusian Obtained 87 | 00

~ pp(rE)(r. PP(Re)(Re) heoretical| 87 | 00

5.|Blue Andalusian - X white Wyandotte Obtained | 27 | 24

PP(Re) (rE) pp(rE) (rE) Theoretical) 25.5} 25.5

6. White Wyandotte X blue has ian Obtained 13 | 18

pp(rE) (rE) PP (Re) (rE) Theoretical) 15.5) 15.5

7. Blue Andalusian ` X white Plymouth Rock (Obtained 80 | 5

PP(Re) (rE) pp(rE) Theoretical 67.5) 67.5

8.| White oe Foye Rock X blue ‘Andalusian Obt d 32

rE) PP(Re) (rE) Theoretical, 28 | 28

9.| Black fev horse white Wyandotte Obt 18

(rE) (rE) pp(rE) (r. heoretical 00 | 18

10. Black Andalusian X white Plymouth Rock (Obtained | 00 |132

PP (rE) (rE) pp(rE) (rE) 3 neoretical 00 |132

11. White Plymouth Rock x bikak Andalusian Obtained

pp(rE) (rE) PP(rE)(rE) [Theoretical 00 | 28

The results of these crosses are understandable on the

‘assumption suggested in the earlier paper that the indi-

viduals from the recessive white races are homozygous

for the factors E and p, p being the recessive allelo-

morph of P, a factor necessary for the production of

black pigment in the feathers. Sturtevant (1912) first

suggested that Wyandotte white is recessive, a fact which

was overlooked in my earlier paper. Morgan and Good-

ale (1912, p. 115) have made a similar assumption for the

white Plymouth Rock.

Since in the series of experiments being reported here,

reciprocal crosses of white Wyandottes and white

Plymouth Rocks gave only whites, thereby: showing no

No. 639] INHERITANCE OF BLUE IN POULTRY 307

evidence of recombination, it seems fair to assume that

the white of both breeds is due to the same recessive

factor p in homozygous condition.

The condition of the white Rocks and white Wyandottes

reported in Table II, with reference to E, appears clear,

since in all crosses with blue-splashed Andalusians (and

as will appear later, in the case of the Wyandotte, with

blue-splashed Orpingtons) which are homozygous for P

and R, but do not carry E, all offspring, 381 in number,

were without an exception, blue (see mating groups 1 to

4, Table IT).

On this basis blue Andalusians, PP(Re)(rE), mated

with such recessive whites should produce blues and

blacks in equal numbers. Mating groups 5 to 8, inclusive,

in Table II show the results of such matings, which com-

bined give 144 blues to 129 black (136.5 to 136.5 would be

equality), a fair realization of the expectation.

As would be expected from the foregoing, crosses of

similar recessive whites with black Andalusians (PP(rE)

(rE)) (see Table II, groups 9 to 11, inclusive) gave only

blacks. Of these there were in all 178 individual and no

exceptions.

-The offspring of the crosses reported in Table II fre-

quently gave evidence that the recessive white parents

carried pattern factors as eryptomeres, but for the sake

of clearness these complications, which have nothing

directly to do with the study in hand, have been ignored

in summarizing the data. As was to be expected, the

white Plymouth Rocks carried the sex-linked pattern

factor for barring. All pigmented offspring by a white

Rock sire showed evidences of barring as soon as the

definitive feathers appeared. Two such, the offspring of

a white Plymouth Rock ¢ and blue Andalusian 2 are

shown in Fig. G, Plate II. Even at hatching, the occipital

spot, which may be a juvenile effect of the factor for bar-

ring, gave notice of the presence of the barring factor.

In the work here reported it was found possible to classify

308 THE AMERICAN NATURALIST: [Vou. LV

in the down pigmented offspring of a non-barred

3 X white Plymouth Rock 2 cross accurately with regard

to sex, by the presence or absence of the occipital spot.

Morgan and Goodale (1912) made use of this spot in

classifying barred and non-barred chicks which failed to

hatch and Punnett (1919) also has made use of it in sort-

ing newly hatched cross-bred chicks according to sex.

The progeny of crosses involving white Wyandottes

frequently displayed Wyandotte lacing of. a lesser or

greater degree of perfection, though the appearance of

this pattern was neither as constant nor as distinct as

that of the barred pattern. The appearance of the lacing

was to be expected if, as is generally stated in the litera-

ture on Wyandottes (see McGrew, 1901), the white

variety was derived directly from the silver Wyandotte,

which is laced.

In connection with these recessive white crosses is to

be noted the fact that several white individuals, although

‘* pure-bred ’’ in the terminology of the poultryman, gave

results which differed from the foregoing. Four white

Wyandotte females proved to carry both the R and EF

factors and were of the same composition with respect

to these factors as a pure-bred blue Andalusian, but unlike

the blue Andalusian they carried p in the homozygous

condition. One of these, which has already been reported

on elsewhere (Lippincott, 1919), carried the sex-linked

pattern factor for barring as well. Dryden (1916, p. 67)

has also reported a white Wyandotte carrying a factor

for barring.

One white Plymouth Rock and eight white Wyandottes

proved to be heterozygous for a factor for dominant

white. These were tested and found to be homozygous

for p. In other words they carried both dominant and

recessive white. Bateson and Punnett (1905, p. 117)

appear to have had birds of this type and Dryden (1916,

p. 66) reports a white Wyandotte which produced only

white chicks when mated to a black Minorea, hence must

No. 639] INHERITANCE OF BLUE IN POULTRY 309

have been homozygous for a dominant white factor.

Whether it carried P or p, the evidence does not show.

So far no attempt has been made to ascertain whether

this factor for dominant white is the same as that nor-

mally carried by the white Leghorn and which Hadley

(1913 and 1914) designated as J. For convenience and

to recognize the possibility of its differing from J the

factor here dealt with is referred to in this paper as I?

(inhibitor of pigment) and its allelomorph as 7”.

VII. Back-crossts or F,’s FROM BLUE-SPLASHED

- ANDALUSIAN X RecgEsstvE Waite MATINGS

The results of crossing the F, blues from the blue-

splashed Andalusian X recessive white crosses is shown

in Table III.

While by no means all possible back-crosses have been

made, enough are represented to show clearly that factors

Rand E were appearing in approximately equal numbers,

and that this was also true of P and p, though in some

cases the presence of J? complicated matters somewhat.

It was, unfortunately, not always possible to use the

actual parents in making back-crosses and though indi-

viduals from the same families were employed, this

proved to be no criterion that they would be of the same

genotype as the individuals used in the original cross.

There can be no question as to their factorial composi-

tion, however, as each individual has been either delib-

erately tested or had happened to be so mated for another

purpose as to give dependable evidence on its composi-

tion with respect to J” and p.

So far as it goes, the evidence, which is substantiated

by the results of other crosses to be reported in a later

section of this paper, also shows that the meeting of P

and R was according to chance, thereby indicating no

linkage between these two factors.

It will be noted that the blue F, 9? in group 5 of Table

III had a blue Andalusian mother instead of a blue-

[Vou. LV

THE AMERICAN NATURALIST

310

*SS019 [VULSILO OY} JO UOTJDIIIP PUL PULY 94} O}BoIpUL 0} pasn st soepqey yueNbasqus pue S1} UL UOUIAUOI STILT, ¢

x (A4) OD arardd CH4) (A) gt 144

ogz'te | SLES | SLES 00 [8010109 J, O Pon ae ok IYOPULAM Ə i

eg. L e 00 poureqo p puy qd “Omid X TTA TTR y

; (a4) (24) dd (a4) (q4)ad

0% OL Or 00 TEDNI, PIY TN FP cow qINoULALT oit .

7a 8 8 00 poureyqgQ D pussy “qa TA > eR Ia OPIA ,

(H4) (4) gt [4d ‘ Cd ataiad

0sz'z9 | SLE'OT | SLEOI 00 [#01}01004, 1, ua Ste e A N o

+9 L Z 00 poure3qo eee ah OU XS ‘pussy “ya TOMIE v

(4) CDdd i (A“) apd

00 00 EELA a at sane uvisnypepu suyds-on Ô OU “UM | on a

00 00 #3 ez paure}qo eee prym ee €

OEA )dd A aoad

00 00 S'e os [eo1}0.10904, T, _UvISN[epu sejds-ən o Py S ən ne

00 00 e t poure}qo oo NX p ‘pusim qa TOMA z

(CNCOdd i (14) 0M) dd

00 00 get S'EI Teotorooy L uvisnye sejds-ən O PUM Ye jom di

00 00 ZI cI poureyqo pate me ie 4X ~~) “puy ‘1ds -m I

IM yva əng PENTE sy Pe TD

SUSSOUD AHA, AAISSHOay X NVISATIVANYV CAHSVIAS-AAIQ ANV HAIG WOwd SUATJ E PNISSOYO-MOVE AO SLTASAY AHL DNIMOHG

II yay L

No. 639] INHERITANCE OF BLUE IN POULTRY 311

splashed. From the nature of the behavior of the factors

R and E already described, this would make no difference

with regard to the blue offspring, for the blue progeny of

a blue Andalusian female by a white Wyandotte male

would be of exactly the same composition with respect

to R, E, and P as all the offspring of a blue-splashed

Andalusian mother by the same sire.

It will also be noted in this group (5) that while the

father of the F, blue was a white Wyandotte, the male

used in this cross was a white Plymouth Rock. Since it

has been shown that for the factors being studied, white

Plymouth Rocks and white Wyandottes are identical, this

should not affect the ratios.

VIII. Tue F, Ratios rrom BLUE-SPLASHED ANDALUSIAN

X Recessive Warre MATINGS

The F, ratios from various blue-splashed Andalusian

X recessive white crosses are shown in Table IV.

As will be seen, the four F, classes predicted for

such crosses in the writer’s earlier paper (1918a, p. 113)

on the basis of the F, results, have been obtained.. No

other classes have appeared. This would seem to indicate

that the factorial compositions of the blue-splashed

Andalusians and white Wyandottes then proposed were

correct and that the white Plymouth Rocks used were of

the same composition with respect to the factors R, E

and P as were the white Wyandottes.

Seven F, blue males were used in securing the F,

ratios. The legband numbers of these males may be

found in Table IV, in the column headed ‘‘ Band No.”

The direction of the original cross is indicated for each

male and for the group of females with which he was

mated. ‘The direction of the cross was the same for the

males and the females in all cases but two. Males 296M

and 258M were mated with females which were products

of the same crosses, respectively, as they themselves

(groups 2 and 7), and also with females from the recip-

rocal crosses (groups 3 and 8).

[Vou. LV

THE AMERICAN NATURALIST

312

-WVM pure spog

(q+) (ay)dq sonq "a pue ‘(g4) (gq4)dd soy op

oy ‘(24) (24) qa Wernpepuy pəyserds-ənjq :əre səssorə osəyy ut pəapoaur sod} 10[oa Jo ajnmaoy HI, +

Oog"e2z | SZ9907 | 939907 | Sz'ETF | TonoIooyL

LSZ PEZ - TST og? poureyqQ £980" = d OTF9'S = zX S988019 [|V 10} SOILA [RIO],

` j 899260" = d ; 028 = :X

ese | S2g6'e GLE6'E 0928'2 TEINI, = O PIN I oa WS WOY TA i s

if £ £ FI poeureyqoO D NE an" BoM © ‘puy ‘ds emm .

5 98Z180° = d i 0192'9 = :X

GLIp | SZIele | otete | 9Z0zo TENAIO I, mou M n PU ue, vee

Tg 68 9% 1¢ pourniqg © ‘puy ‘ds oy eee © ‘puy ‘ids en 2

6er = d 6FEL'S = X

GL'6I | SZIS'FI | ezts'Ft | ¢z9°6z | onog, $ PON TA 5s omg x qop —Ê EU TA ay ten 9

SI LI Or Pe pouregqo " p puy ‘ids © 'puy ‘jds

$661: = d 1289} = :X

oosz'ees | SLZerIt | cere | 9L8'Z8 PINANI, 6 Pov 8, qong x qeg — E POV S ma onig -e

Z? bP IP 6 pouresqQ © PUAN “yar © ‘PUAN “pA

SlZI' = d 10e2'¢ = :X

SI S298 9739'S SALT PIY 1 o Poy Ie ly nq X W99 6 pay ‘18 1 ony Kae i

9 L PI 61 poureyqQ P PTL qa O 'PUBAM ‘Ya

zeces’ = d - 6928" = :X

O0SZ'LZ | SLEFOS (geroz | 9Z8°0F | monoioong, a O ee O PUEA UA i ong “+g

1Z oz LI oF pouregqo © “pues a © ‘puy ‘pds

i ordz = d ; 3000°F = +X

oF ore ore 69 TEINA, P te n O POSM e vee

SP 68 +Z eZ poureyqo 2 puy jas 4N n po p puy jas "1 OMe z

6607 = d LLS? = :X

0092'89 | Sz9STS | 9Z9S'19 | OSZT'EOT| TENAIO, D PUIA TA iong x mog S COSSA ia a ad ici

+9 cg oF oor pəureqo 2 ‘puy ‘ds © ‘puy ‘ids

VM OVAL | 'ras-ona ong 86 | © dnoay

‘ON

| pus

PSMOOY HLOAOWATA ALIHM ANV SALLOGNVAM ALIH AA ANV SNVISOIVONY GHHSVIdS-aN1g JO SASSONO WOM Solvy “T ONIMOHG

AT WTAViL

No. 639] INHERITANCE OF BLUE IN POULTRY 313

As may be seen by inspection of Table IV, but one male

(296M) gave a group of offspring (3) which was very

close to expectation. The chances that as great a devia-

tion as this one would appear as a result of random

sampling are four to one. The mothers of this group

were the product of a cross which was the reciprocal of

that which produced their sire. The offspring of 296M

when mated with females which were the product of the

same cross as himself (group 2) gave a deviation so great

that the chances against its appearing as a result of

random sampling are three to one. The chances of the

appearance of deviations as great as those shown by the

offspring groups of the other males were as follows: 86E

(group 1) one chance in a little less than five; 66M (group

4) one chance in about eight; 65E (group 5) one chance

in approximately five; 46E (group 6) one chance in about

two and a quarter; 258M (group 7) once in about twelve

times with females from the same cross as he, and once in

twenty-five when mated with females from a reciprocal

cross (group 8).

It would be unusual, though not impossible, to have so

many comparatively wide deviations from expectation

simply as a result of random sampling.

If the genetic constitution of the F,’s was as has been

previously postulated, and these were in fact all chance

deviations, it would be highly probable that the lumping

of all the data given in Table IV would approximate the

calculated ratio fairly closely.

The lumped data are given at the bottom of Table IV.

It will at once be seen that the goodness of fit as meas-

ured by P is poorer than the poorest constituent group,

and would be probable, on the basis of random sampling,

once in about twenty-eight times. It seems fairly clear

that some disturbing force was operative.

The two possible causes of disturbance which present

themselves are linkage and a differential viability of

classes, or it might be a combination of the two.

314 THE AMERICAN NATURALIST [Vou LV

Linkage between the two principal pairs of factors in-

volved in the crosses, Pp and (Re)(rE), may not be

appealed to because the only possible linkage relation

would produce results diametrically opposed to those

with which we are confronted. Since according to our

hypothesis the recessive white parents were in each case

of the composition pp(rE)(rE) and the blue-splashed

Andalusian parent PP(Re) (Re), it is evident that linkage

would require the production of p(rE) gametes by the F,

blues, more often than P(rE) gametes. And similarly

the combination P(fe) should also appear more often

than p(Re).

A complete linkage between these pairs of allelomorphs

would result in an F, ratio of 1 blue-splashed and 2 blue

to 1 white, the blacks not appearing. The tendency of

even weak linkage would be to reduce the proportional

number of blacks. This should be true irrespective of

the direction of the cross. It would further be true, that

unless crossing-over occurred in both sexes any linkage

whatsoever would inhibit the production of F, blacks

homozygous for P. As will be shown in a later section

of this paper, however, F, blacks homozygous for P have

been identified. Even a casual inspection of Table IV

shows that a relative preponderance of blacks is a quite

constant characteristic.

Crossing-over in the male fowl has been found by Good- ©

ale (1917) and in the male pigeon by Cole and Kelley

(1919). The latter investigators definitely state that

there is no crossing-over of sex-linked factors in the

female pigeon. Goodale states that none had been ob-

served in the female fowl, but that a definite test of the

matter would be made later. So far as the writer is —

aware no further report has been made. It should per-

haps be pointed out that so far only sex-linked factors

have been dealt with, no autosomal linked groups in birds

having so far been reported.

There are no F, data available from crosses where p

and (Re) are found in one parent and P and (rE) in the

No. 639] INHERITANCE OF BLUE IN POULTRY 315

other. The F,’s from such a cross have been secured by

mating an extracted white of the composition pp(Re) (Re)

with a black Andalusian, PP(rE)(rE), which gave all

blues. From these an attempt will be made to secure F,’s

in considerable numbers. Back-crosses to the parental

types will also be made. The F,’s should approximate

the same ratios as appear in Table IV and also give some

evidence on the second possible explanation of the per-

sistent deviations about to be discussed.

The calculation of theoretical expectancies presup- -

poses the equal viability of all phenotypic and genotypic

classes. If for any reason the individuals of one or more

of the obtained classes tend to be less viable than certain

other classes, deviation from expectancy will occur if the

lack of viability expresses itself prior to making the

counts.

As has already been pointed out, the lumping of the

data presented in Table IV brings forth a poorer fit than

is shown in any of the constituent groups. The deficient

classes are the blue-splashed and the white, while the

most preponderant class relatively is the black.

It seems to be a rather tacit assumption among poultry-

men, particularly, it must in truth be said, among those

breeding pigmented varieties, that the recessive white

varieties are less vigorous (and so in all probability less

viable) than the pigmented varieties of the same breeds.

In how far this assumption is based on fact there is no

-critical evidence to call upon.

Regarding the relative viability of splashed and self-

colored races there is no suggestion from any source.

Splashed varieties are, so far as I am aware, nowhere

bred as such, and the experience of practical breeders

may accordingly not be appealed to.

While in the case in hand the assumption of low via-

bility on the part of the individuals of the splashed and

recessive white classes seems to correspond with the

facts, such an assumption, though convenient, is not cor-

w816 THE AMERICAN NATURALIST [Vor. LV

roborated by other evidence. That the splashed classes

are not necessarily always deficient is shown by the

progeny of the blue-splashed X blue mating in Table I,

group 5, and of the F, blue X blue-splashed matings in

Table III, groups 1, 2 and 3.

The latter fact suggests that possibly certain individ-

uals used in these matings carried recessive factors tend-

ing to cause low viability, which were linked to the factor

R. Until the fact of a differential viability is demon-

. strated, however, it is useless to speculate on this possi-

bility. The reason for the deficiencies in the blue-

splashed and also in the white classes, therefore can not

at present be determined.

IX. [DENTIFICATION OF THE F, GENOTYPES

As indicated in my former paper (1918a, p. 113) the

genotypes expected in the several F, phenotypes from the

blue-splashed X recessive white crosses are as follows:

blue, PP(Re)(rE) and Pp(Re)(rE);_ blue-splashed,

PP(Re)(Re) and Pp(Re)(Re); black, PP(rE)(rE)

and Pp(rE)(rE); white, pp(Re)(Re), pp(Re) (rE) and

pp(rE)(rE). Although the limitations of equipment

were such that comparatively few F, individuals could

be tested, fortunately all of the genotypes but one have

been identified by making the appropriate crosses. The

blues mated to individuals homozygous for p and E gave

blues and blacks in equal numbers, or, blues, blacks and

whites in the approximate ratio of 1:1:2, as the case

might be. The blue-splashed mated to individuals of the

same constitution produced all blues, or, equal numbers

of blues and whites, depending upon whether or not they

were homozygous with respect to P. Similarly the blacks

gave all blacks, or, blacks and whites, depending upon

their condition with respect to P.

The whites on the other hand were mated to blacks

known to be homozygous for P and E. The pp(Re) (Re)

whites, as mentioned in an earlier section of this paper,

No. 639] INHERITANCE OF BLUE IN POULTRY 317

gave all blues, just as would blue-splashed Andalusians.

The pp(Re)(rE) whites produced blacks and blues in -

approximately equal numbers, exactly as would blue

Andalusians. The parental white genotype pp(rE) (rE),

which would give all blacks, was curiously enough, the

one of the whites which did not happen to be selected

for testing.

It is important to note that while eight out of the nine

F, genotypes were identified, no genotypes were found

other than those expected.

X. DATA on ANDALUSIAN X BLACK LANGSHAN CROSSES

It appeared desirable, in order to ascertain whether

there was anything inherent in Andalusian black which

made its relation to Andalusian blue different from that

of other black breeds, to make certain matings of Anda-

lusians with black Langshans. The Langshan was chosen

because not only is it a different breed, but it also belongs

to a different group of breeds. The original black Lang-

shans were, according to Brown (1906, p. 63), imported

from China, while the Andalusians, according to the same

authority (p. 107), originated from native stocks along

the borders of the Mediterranean Sea. So far as is

known they have nothing in common in their immediate

ancestry. Davenport (1914) even points to the proba-

bility that the immediate wild ancestors of the Asiatic

breeds differed from those of the Mediterranean breeds.

If blacks differ in their relation to Andalusian blue it

would seem probable that Andalusian black and Lang-

shan black might show this difference.

The results of the Andalusian-Langshan matings are

shown in Table V. As may be seen readily by reference

to this table the results are in every case precisely those

which might be expected if a black Andalusian had been

substituted for the black Langshan. So far as the prin-

cipal factors under discussion are concerned it appears

that the black Langshans used were identical in composi-

318 THE AMERICAN NATURALIST [Vou. LV

TABLE V

SHOWING THE RESULTS OF SEVERAL ANDALUSIAN X BLACK LANGSHAN

+| rg]

a ð|

a sal s|%

Š z eS Ae a | ag

Sai |

1: . Blue Andalusian ` X black ont ay ‘Obtained 0 34

PP (Re) (rE) PP(rE) (rE) Theoretical 0 pe 50 aa eb

; h. Wyand. g|

2.. Blue Andal m ned 21

pesare X biak black Lang. Ẹ ete 0 18 18

PP(Re) (rE) Pp(rE) (rE) |

3.. Blue-splashed Andalusian X black Langshan Obtained 65 0

PP(Re) (Re) PP(rE) (rE) ‘Theoretical 0 65 0

ack Andalusian x black zangshan Obt: ned 0 14

PP(rE)(rE) gt (rE) Theoretical o 0 n

b löi blue And. 3 oe e And. d

= black Lang. @ ka Ti Lang. 9 Obtained ; 4 5 9

fhe m PP(Re)(rE) eoretical|4.6 | 9 | 4.5

6. |Blue Pte And: b blue And. g

lack Lang. 92 X. pisak black Lang. 2 |m Obtai ned s 2 = 5 i 5

iar al : PP(rE) (rE) Ti heoretical v . .

7..|Blue wa g7 _.& black Langshan Obtained 0 12 13

4 ac ng. 9 TheoreticalO 12.5 12.5

lue >

8.. Blue i

vf black Lang. 9 SO ined lb 1 6 | 8

PP(Re) (rE) PP (Re) (rE) Tt tical'3.25 6.5 | 3.25

tion with the black Andalusians, being PP(rE)(rE).

The condition of the Langshan with respect to P was

found by mating individuals with white Wyandottes,

whereby only black, i.e., pigmented, offspring were

produced.

XI. Tae RELATION or ORPINGTON BLUE TO

ANDALUSIAN BLUE

Among the Orpingtons, an English breed, is a blue

variety. Like the blue Andalusian it is an inconstant

breeder with regard to color, segregating into blue-

splashed and blacks as well as blues. Though by no

means as widely bred as the blue Andalusians, it has

numerous admirers, some of whom have claimed verbally

to the writer that the proportion of wasters, t.e., blue-

splashed and blacks, was much smaller than in the Anda-

lusians, though no figures are obtainable by way of sub-

319

No. 639]

6997 =d 9090F = :X

. (4) (0a) (94) (oad |

92 +Z 9% 09 poureyqo 2 PUAN “Yar © 'PUBAM “yA p

0 ez oF eg TEINIT (44) EAR) aea) |

0 GI 8G 6I pəurwgqo uoysutdig anq X uoysUIdIG, ənjg g

0 0 1Z 0 [Bo1}01004,L, CEM CMdd (44) (q4)jdd |

0 0 13 0 poureyqo uo}3urd1o peyseyds-onyq X ƏVOPUBAM FYM we:

0 0 pI 0 VIAN, (24) CY) (D4) |

0 0 tI 0 poureyqo uog}3urd10 poyseyds-onjq X uersngepuy Aoupg e

0 0 is 0 [e9109 J, (A+) (A+) CA) A) |

0 0 18 0 poureyqQ uoyuIdiQ Youlq X uvisnyepuy poyseds-onjg, ee

0 ZI 0g Ai poureyqQ uog3urd10 anjq X , uersnpepuy ong ot

amm | g ong | '18s-onjg ô ô | 2 | dnoip

INHERITANCE OF BLUE IN POULTRY

saaug

TA WTA VL

WHLO HIM ANV SHATASNAH]T, ÐNONWNV SASSOUD NOLONIdYQ MOV IG ANV GUHSVIdS-an TE “ag NIVLYAD JO SLIASIY FHL ÐNIMONQ

320 : THE AMERICAN NATURALIST [Vou. LV

stantiation. It seemed desirable from several standpoints

to ascertain what factors were involved in the production

of Orpington blue, and whether the blue Orpington dif-

fered from the blue Andalusian in its genetic behavior.

A number of matings were accordingly made, the data

from which are shown in Table VI.

These data are consistent with the supposition that the

factors involved in the production of Orpington blue are

identical with those which produce Andalusian blue. The

crossing of blue Andalusians and blue Orpingtons gave

exactly the same sort of result as that obtained by mating

blue Andalusians inter se, as shown by group 1. The

blue-splashed Orpingtons mated with white Wyandottes

gave only blues (group 4) just as did the blue-splashed

Andalusians. And finally the F, ratio from white

Wyandotte X blue-splashed Orpington crosses gave the

“same phenotypic classes as were obtained in the F, from

the white Wyandotte X blue-splashed Andalusian cross,

with a deviation from expectancy as great as would be

probable once in four times. It is interesting to note that

while the white class is deficient in this case, the blue-

splashed class is not.

XII. Data rrom BLUE LEGHORN Crosses

In the spring of 1917 there appeared in the large pure-

bred single comb white Leghorn flock of the Pabst Stock

Farm at Oconomowoc, Wisconsin, two blue females. The

flock was not pedigreed and nothing is known of the indi-

vidual ancestors of these birds. They were of fair Leg-

horn type and were, as far as known, the offspring of

pure-bred white Leghorn parents. Through the courtesy

of Mr. Fred Pabst, and Dr. L. J. Cole of the University

of Wisconsin, these individuals came into the hands of

the writer and were entered on the records of the Depart-

ment of Poultry Husbandry of the Kansas State Agricul-

tural College as numbers 767 and 768,

Number 767 was a fairly even shade of medium to light

No, 639] INHERITANCE OF BLUE IN POULTRY 321

blue when received and showed some evidence of barring,

though this was not very distinct. Number 768 was much

lighter in shade than 767 and showed no evidence of bar-

ring. In contrast with ordinary blue she would, from a

little distance, be mistaken for a white. The pigment

granules in both cases were round.

The results of mating these birds in various ways are

presented in Table VII. The numbers are rather small

TABLE VII

SHOWING THE BREEDING BEHAVIOR OF Two BLUE LEGHORN FEMALES, WHEN

MATED WITH VARIOUS MALES OF KNOWN FACTORIAL COMPOSITION

A a la

F 9 Sa) 2) 3 13

Aa a mie

White Leghorn 117M X TOF 0 0 0 5

IIPP(rE)(rE) iiPP(Re) (rE) Theoretical a o 0 16

White Leghorn 117M 0 0 0 11

IIPP(rE)(rE) iiPP(Re) (rE) Tiei] © 0 16 n

Blue Andalusian 78M X TOT ain 2 4 2 0

(Re) (rE) PP (Re) (rE) Theoretical |2 4.42% 0

Blue Andalusian 78M X 768 Obtained 3 eg 4 0

PP(Re)(rE) PP(Re)(rE) Theoretical |3.5 | 7 3510

White Plymouth Rock 155M x 767 Obtained -J0 (20 (24. 0

(rE) (rE) PP(Re) (rE) Theoretical 0 22 (22 0

White we andotte 192M x Obtained ) 8 8 |12

IPi?pp(rE)(r iPiPPP(Re)(rE) Theoretical yoly i4

white Rocko" 155M Obtained |2 6 1 0

eee” blue Leghorn Q 767 259M X TO Theoretical |2.25 4.50| 2.25) 0

Pp(Re) (rE PP(Re) (rE)

Black Siete 288M 768 btained (0 25 22 0

; PP(rE) (rE) PP(Re) (rE) Theoretical '0 ‘23.5 '23.5 | 0

but two facts seem fairly evident. First, that 767 and

768 are alike with respect to the factors under discussion

in this paper, and second, that they give no indication of

being different in their make-up with respect to the fac-

tors R, E and P from pure-bred blue Andalusians.

The appearance of the blue offspring of 768 (which it

will be recalled was very light) when mated with black

or blue Andalusians, was such as to suggest the possi-

bility that accessory factors, necessary for the produc-

tion of blue of normal shade, were supplied by the Anda-

5 The theoretical expectancies calculated as for blue Andalusians.

322 THE AMERICAN NATURALIST [Vor. LV

lusian males, though no attempt was made to isolate and

identify them.

Since these blue Leghorns arose in an unpedigreed

flock, their origin is conjectural. A plausible explana-

tion seems to be that two individuals heterozygous for I,

the dominant Leghorn factor described by Hadley (1913),

which inhibits the production of pigment shown (also by

Hadley, 1914) to be normally present in the white Leg-

horn, happened to mate and that at least one of them

carried the factor R as a cryptomere. That white Leg-

horns may sometimes carry the factor seems to be shown

by the fact that Dryden (1916, p. 67) secured blue chicks

in an F, generation from a barred Plymouth Rock X white

Leghorn cross. And further, in the course of the breeding

operations reported in this paper, blues appeared in the

progeny of a black Andalusian and a white crossbred, the

latter being the product of a black Andalusian Xx white

Leghorn cross. In both cases it appears that the factor

ER must have been brought in by the white Leghorn. This

suggestion also involves the assumption that the white

Leghorn carries the factor E. That this is the case is

shown by the fact that in the F, from a blue-splashed

Andalusian X white Leghorn cross, the details of which

are reserved for later publication, both blacks and blues

appeared.

XT. Tue PROBLEM or TRUE-BREEDING BLUES

The fact that the blue varieties of both Andalusians

and Orpingtons as now constituted do not breed true is a

matter of considerable importance to their breeders. It

is a heavy handicap to both varieties: While one hun-

dred per cent. of blues may in each case be secured: by

_ mating blue-splashed individuals with black, as a matter

of practical breeding this mating is seldom made. This

is owing to the fact that there are several more or less

variable qualities of color for which rigid selection is

practised which are not apparent in either the blue-

No. 639] INHERITANCE OF BLUE IN POULTRY ous

splashed or blacks. The breeder therefore prefers to use

for breeding purposes only those individuals which show

the desired phenotypic condition, even though so doing

necessitates the discarding of approximately half the

offspring. While this leaves a comparatively small num-

ber of individuals, as compared with other breeds, upon

which to practise selection, the blue Andalusian at least.

is bred in considerable numbers, thereby indicating its

economic desirability and its attractiveness.

As was pointed out in the earlier paper (1918, p. 111)

if R and FE are not at identical loci on homologous chromo-

somes and crossover individuals were found which pro-

duce RE gametes, the problem of the constant-breeding

blue would be solved.

The situation regarding’ black in rats may not be with-

out its bearing in the present case. Black rats which

bred true have been known for some time. Castle (1919)

has, however, reported certain races of blacks which

failed to breed true. This type of black was tested

through several generations by Castle (1919), Ibsen

(1920) and Dunn (1920). Blacks mated to blacks quite

persistently produced whites, blacks and red-eyed yellows

in the ratio of 1 to 2 to 1. Castle (1919) found one pos-

sible cross-over individual which died without being

tested. Ibsen (1920) has so far failed to find any, and

Dunn (1920) reports between one and two per cent. of

cross-overs. These cross-overs, which were longer

sought for and among larger numbers than has yet been

possible with Andalusians, would appear to make it pos-

sible to synthesize a true breeding (?.e., homozygous)

black, from the line which has not been breeding true

through a considerable number of generations.

It is also worth noting in this connection the possible

bearing of Sturtevant’s (1919) finding families of

Drosophila carrying at least two definite factors in the

second chromosome which almost completely inhibit

crossing-over in the region contiguous to their loci.

324 THE AMERICAN NATURALIST von. LV

If, however, after a long-continued search, it becomes

increasingly evident that R and E are indeed allelo-

morphs, as originally suggested by Bateson and Punnett

(1905), it was suggested (p. 113) that hope might be seen

in the progressive selection of the darker, that is, more

fully pigmented, blue-splashed individuals, there being

considerable variation among the latter in this regard.

There is a further possibility which should not be over-

looked, namely, that other factors might be found, per-

haps in other breeds, which would act on black pigment

to give the blue appearance on the one hand, or extend it

to give self-colored individuals on the other. If dupli-

cate factors for Æ or R should be found, a means of

producing the long sought true-breeding blue would seem

to be at hand. The fact that three factors are known

which produce white in fowls lends emphasis to the possi-

bility. It would seemingly make little difference in the

ultimate outcome whether the new factor was linked to

R and £E, or was located on a different chromosome pair.

In either case it would be possible to get a ‘‘self-coloring”’

and a ‘‘bluing’’ factor in the same gamete which, it ap-

pears, has so far not been done.

XIV. SUMMARY

1. It has been shown that the development of black

pigment in the blue-splashed, blue and black races of the

Andalusian and Orpington breeds, and of black Lang-

shans, depends upon the action of a dominant hereditary

factor P, for which they are normally homozygous.

2. The allelomorph of P is p. Individuals homo-

zygous for p are white, as in the white Wyandotte and

white Plymouth Rock breeds.

3. The extension of black pigment to all feathers of the

body, resulting, if no pattern factors are present, in self- `

colored individuals, depends upon a dominant factor E.

This factor has been found in the Andalusian, Orpington,

white Plymouth Rock, white Wyandotte and black Lang-

No. 639] INHERITANCE OF BLUE*IN POULTRY 325

shan breeds. Some evidence is presented which indicates

its presence in white Leghorns.

4. The blue appearance of blue and blue-splashed

Andalusians and Orpingtons, is due to the arrangement

and restriction of black pigment, the result of a dominant

factor R. This factor has also been found in individuals

of the white Wyandotte and white Leghorn breeds,

though its presence is probably not usual in these breeds.

5. No individuals of the Andalusian, Orpington, white

Plymouth Rock, white Wyandotte or black Langshan

breeds have been found which did not carry R, E or both.

6. The mutual relations of R and E are such that they

have never been found together in the same gamete.

This indicates that they are allelomorphie, t.e., occupy

identical loci on homologous chromosomes, or, each is so

closely linked to the recessive allelomorph of the other,

(Re) and (rE), that crossing-over rarely, if ever, occurs.

7. No evidence of crossing-over between R and E has

been found and the tentative conclusion must be in accord

with that previously held, that R and E are allelomorphs.

8. Both R and E are independent of P in their heredi-

tary behavior, though dependent upon its presence for

their manifestation.

9. The cooperative influence of the ovary is necessary

for a full expression of ÈR in the regions of the neck, back

and saddle.

10. On the basis of the evidence presented in the body

of this paper the genetic formule of the breeds and varie-

ties employed, with respect to the factors under observa-

tion, are usually as follows: blue-splashed Andalusians

and Orpingtons PP(Re)(Re); blue Andalusians and

Orpingtons PP(Re) (rE) ; black Andalusians, Orpingtons

and Langshans PP(rE) (rE); and white Plymouth Rocks

and Wyandottes pp(rE) (rE).

11. The possibility of the occurrence of factors which

duplicate the somatic effects of R and E is pointed out,

and the relation of this possibility to the production of

constant-breeding blues briefly discussed.

326 THE AMERICAN NATURALIST Vor. LV

XV. BIBLIOGRAPHY

American Poultry Association,

1919. The Plymouth Rock Standard and Breed Book. 432 pp. Pub-

lished by The American Poultry Association.

Bateson, W., and R. C. Punnett.

1905. Heports to the Evolution Committee of the Royal ee a

pp. 99-119.

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

pp. 11-30.

Brown, page d.

19 Races of Domestic Poultry. 234 pp. Published by Edward Ar-

nold, London.

Castle, W. E. ~

1919. Studies ` Heredity in Rabbits, Rats and Mice. Carnegie In-

stitution of Washington, Publication No. 288, pp. 56. 3 plates.

oT and F J. Kelley.

1919. Studies on Inheritance in Pigeons. III. Description and Link- .

age Relations of Two Sex-linked Characters. Genetics, Vol. 4,

pp. 183—203.

nae and W. A. Lippincott.

The Relation of Plumage to Ovarian Condition in a Barred Ply-

mouth Rock Pullet. Biological Bu`etin, Vol. 36, pp. 167-183,

2 ir ates. ;

eres C. B. ;

4. The Origin of Domestic Fowl. Jour. of Heredity, Vol. 5, pp.

313-315, 2 plates.

Dryden, Jas

1916. ec Breeding ayp Management. 402 pp. Published by Or-

e Judd Co., N. Y.

Dunn, L. C.

1920. eee in Mice and Rats. Genetics, Vol. 5, pp. 325-343.

Goodale, H

1916. EEEE in Relation to the Secondary Sexual Characters of

Some Domestic Birds. Prong 243, Carnegie Institution

of Washington, 52 pp., 7 p

1918. Feminized Male Birds. aac Vol. 3, pp. 276-295, 2 plates.

Hadley, P. B.

1913. Studies on the cepa of Poultry I. The Constitution of the

White Leghorn Breed. Rhode Island Agr. Exper. Sta. Bull.

155, pp. 140-216, 3 plates.

1914. II. The Factor for Black Pigmentation in the White Leghorn

Breed. Rhode Island Agr. Exper. Sta. Bull. 161, pp. 449-459.

Ibsen, H. L.

1920. Linkage in Rats. Amer. NAT., Vol. 54, pp. 61-67.

Lippincott, W. A.

1918a. The Case of the Blue Andalusian. Amer. Nat., Vol. 52, pp.

95-

No. 639] INHERITANCE OF BLUE IN POULTRY 327

19186. Pedigreeing Poultry. Kansas Agr. Exper. Sta. Cir. 67, 16 pp.

1919. The Breed in Poultry and Pure Breeding. Jour. of Heredity,

pp. 71-79.

T P F.

1901. American Breeds of Fowls. II. The Wyandotte. Bureau of

Animal Industry Bull. 31, 30 pp.

Morgan, T. H., and H. D. Gooda

1912. See Hinked fine iaia in Poultry. Annals of the N. Y. Acad-

emy of Sciences, Vol, 22, pp. 113-133, plates 17-19.

Pearl, R.

Ia: The Probable Error of a Mendelian Class Frequency. AMER.

Nart., Vol. 51, pp. 144-156.

Piatt, F. L.

1916. ‘‘Western Notes and Comment.’’ Reliable Poultry Journal,

Vi 6 i

Punnett, R. C.

1919. Sex-linked Inheritance and Its Practical Application in the

Breeding of Poultry. Journal of the Board of Agriculture

(England and Wales), February, 1919. Summarized in Official

Report of the International Poultry Conference, London,

Forain 11 to 15, 1919, pp. 64, 65

ee on UO, and 0. A.B

1908. oi Molting of Fowls. Cornell Unir. Agr. Exper. Sta. Bull. 258,

pp. 17—68

Sturtevant, A. H.

1912 7 Experiment Dealing with Sex Linkage in Fowls. Jour.

Exper. Zool., Vol. 12, pp. 409-518.

1919, Inherited Linkage Variations in the Second Chromosome. Ca

negie Institution of Washington, Pub, 278. Part III, pp.

INHERITANCE IN NICOTIANA TABACUM.

ON THE EXISTENCE OF GENETICALLY

DISTINCT RED-FLOWERING VARIETIES!

DR. R. E. CLAUSEN anv DR. T, H. GOODSPEED

UNIVERSITY OF CALIFORNIA

In our studies of inheritance in Nicotiana Tabacum?

it has been demonstrated that the red flower color of

macrophylla (U. C. B. G. 22/07) is recessive to the light

pink of angustifolia (U. C. B. G. 68/07), and the same

relations are exhibited by the red of calycina (U. C. B. G.

110/05) as contrasted with the light pink of virginica

(Maryland, U. C. B. G. 78/05). In both eases the F, was

pink, F, conformed to the ratio 3 pink: 1 red, and F, and

subsequent generations yielded data consistent with a

single factor difference between these two flower colors.

It was also shown that when macrophylla was crossed

with the white-flowering variety alba (U. C. B. G. 30/06),

F, was pink, F, conformed to the ratio 9 pink:3 red: 4

white, and F, and subsequent generations gave data in

agreement with a two-factor difference for this charac-

ter contrast.

Allard,’ however, had presented evidence, at first sight

contradictory to ours, to the effect that the carmine flower

1 The experimental data cited herein were obtained from cultures made

possible by a portion of the Adams’ Fund allotted to the Department of

Botany by the Department of Agriculture of the es of California.

2 Setchell, W. A., T. H. Goodspeed and R. E. Clausen, ‘‘A Pr oy

Note on the Results of Crossing Certain Varieties of Nicotiana Tabacum,’

Proc. Nat. Acad. Sci., 7: 50-56, 1921. A complete illustrated account of

these experiments is in press under the title, ‘‘Inheritance in Nicotiana '

Tabacum. I. A Report on the Results of Crossing Certain Varieties,’’? Univ.

Calif. Publ, Botany 5, no. 17. For descriptions and illustrations of the

varieties mentioned in this paper ef. Setchell, W. A., ‘* Studies in Nicotiana.

I., Univ. Ca if. Publ. Botany, 5: 1-86, 1912.

, H. A., ‘‘Some Studies in Blossom Color Inheritance in Tobacco,

with Special Retecehe to N. sylvestris and N. tabacum.’’ AMER, NATURAL-

Ist, 53: 79-84, 1919.

328

No. 639] INHERITANCE IN NICOTIANA TABACUM 329

color of his Giant Red-flowering tobacco is dominant to

pink in a simple mono-hybrid relation, F, being carmine

and F, 3 carmine:1 pink. He also crossed this carmine-

flowering variety with a white-flowering form and ob-

tained light carmine in F, and a distribution which might

be taken to conform to the ratio 9 carmine: 3 pink: 4 white.

We took these results to indicate that his carmine, which

must be very similar to our red, was nevertheless genet-

ically distinct from it. This belief was somewhat

strengthened by the fact that our red does not fall upon

the carmine of the Ridgway‘ color scale, but lies slightly

removed from it between rose red and pomegranate

purple, although a difference of this kind might conceiv-

ably be due to the effect of differences in the residual

genotype. We have, however, a variety, purpurea (U. C.

B. G. 25/06), which exhibits a red flower color somewhat

darker and more intense than that of macrophylla, and

which some preliminary crosses indicated was dominant

to pink and white. We accordingly suggested the follow-

ing factor formule for these four colors:

WWRRPP = carmine

WWRRpp = light pink

: WWrrpp = red

wwRR pp = white

In this formulation W W RRPP, represents the basic type,

carmine in color; w, the difference from it which gives

white, irrespective of which members of the pairs occupy

the R or P loci; p, that which gives pink; and r, that

which changes pink to red. Obviously white-flowering

varieties may be of four different genotypes, viz.,

wwRRPP, wwRRpp, wwrrPP, and wwrrpp, but our

white variety alba was clearly wwRRpp. This formula-

tion brings our results into accord with those of Allard

and accounts for the existence of genetically distinct red-

flowering varieties. We have now obtained further evi-

dence in support of the correctness of this formulation.

4 Ridgway, R., ‘‘Color Standards and Color Nomenelature,’’ 1912,

330 THE AMERICAN NATURALIST [Vou. LV

We found it necessary to use ‘‘Cuba’’ (U. C. B. G.

200/14),° another white-flowering variety, in these

studies. Since there is the possibility just indicated of

the existence of genetically distinct white-flowering

varieties, it became necessary to determine the genetic

constitution of ‘‘Cuba’’ with respect to the Rr and the

Pp pairs of allelomorphs. A number of crosses were

made, therefore, between ‘‘Cuba’’ and macrophylla as

the starting point for these determinations. In the ac-

count which follows H,,;,= macrophylla 2? X ‘‘Cuba”’ g

and H,,,—the reciprocal. In the season of 1919, 50

plants of F,H,,, and 100 plants of F,H,;; were grown.

They were all pink-flowering except that one plant pro-

duced a small white-flowering branch in an inflorescence

otherwise pink-flowering. This bud variant, one of the

few which we have observed in tobacco, will be taken up

in a subsequent report. The further data on these re-

ciprocal hybrids are listed in Table I. The F, popula-

TABLE I

F, AND Back-cross Data oF THE CUBA-MACROPHYLLA (WHITE X RED)

SERIES

Flower Color

Pca sl Parentage r Totals

Pink Red White

20.075. . 4 19F:HiPssW | 59 22 19 100

20.076. . «| 19F:H174P23P 54 22 23 99

Totals for F' populations | 3 44 42 199

16F:Hıs2..| 15FıHı4 9 X 200/140 12 pian 12 24

19F Hise. . i 48 — 50 98

16F:Hıs3. . 200/149 X 15FiHinc 12 — 13 25

19F:Hıs3. . i 53 — 47 100

16F Hiss. . 200/149 X 15Fi: His" 11 — 13 24

19F:Hıss. . itto 50 — 49 99

16F:Hıs. .| 15FiHizs9 X 200/140 6 — '19 25

19F:Hıs.. ditto 53 oae 47 100

Totals for back-crosses to white 245 — 250 495

20.059... .| 19FiHin? X 22/079" Jo n 28 = 50

5 For description cf. Goodspeed, T. H. ‘‘ Parthenogenesis, Parthenocarpy

and Phenospermy in Nicotiana,’’ Univ. Calif. Publ, Botany, 5: 249-272,

1915.

No. 639] INHERITANCE IN NICOTIANA TABACUM 331

tions give totals of 113 pink: 44 red: 42 white, whilst the

9:3:4 expectation, disregarding fractions, is 112 pink: 37

red:50 white. In the back-crosses to both the white and

the red parents the data are obviously in satisfactory

agreement with the 1:1 expectations. These figures’ do

not establish conclusively the validity of a bigenic for-

mulation for this case, but taken together with the data

from the Ausa-MacropHyLia series which we have pre-

sented elsewhere® it seems most reasonable to interpret

them in this manner. An alternative mono-hybrid inter-

pretation might be argued, but it would not fit the F,

totals as well as the dihybrid ratio. The growing of F,

populations would, of course, soon settle the question,

but the results so far secured indicate essential genetical

identity of alba and ‘‘Cuba”’ in their flower color factors.

In order to demonstrate the difference in behavior of

red of macrophylla and carmine of purpurea we have

made parallel crosses between them and a number of

other Tabacum varieties. The flower colors of these

varieties and of their F, hybrids with macrophylla and

purpurea are listed in Table II. In each case the F, with

macrophylla was pink but with purpurea it was always a

full, intense carmine. Among two hundred plants of the

Cusa-PuRPUREA series one plant appeared which bore

TABLE II

F, RESULTS OF PARALLEL CROSSES OF MACROPHYLLA AND PURPUREA WITH A

SERIES OF TABACUM VARIETIES

Flower Color Flower Color

Variety Name and Number Flower Color | of F, with | of Fi with

Macrophylla

angustifolia (U. C. B. G. 68/07)....... Light pink Pink | Carmine

“Caria” 0: C. D: Gi 72/05). oa Pinkish Pink | Carmine

POuhe” (U: G: BG. 20A.. i a White Pink | Carmine

carminė flowers on one side and light pink ones on the

other. Further studies on this, the most striking case

6 Setchell, Goodspeed, and Clausen, loc. cit.

332 THE AMERICAN NATURALIST [Vor. LV

of somatic variation we have ever observed in Nicotiana,

are in. progress. The F, results in themselves sufficiently

demonstrate the existence of a genetic difference between

the red of macrophylla and the carmine of purpurea.

We have also secured further data from the Cusa-

PuRPUREA series which demonstrates the mode of inheri-

tance of carmine when crossed with the same white used

in the CUBA-MACROPHYLLA series. These results are set

forth in Table III. The totals from the F, populations,

TABLE III

F, AND Back-cross DATA FOR THE- CUBA-PURPUREA ‘(WHITE X CARMINE)

SERIES

Gard | Flower Color | bea

oe Parentage erate

ehcp | Carmine Pink White

| j

19F2Higo. . 16FiHisPs 58 14 26 98

19F2Hie 16F:HiaPs 28 8 ii 47

sarera 19F:H P:R 48 13 39 100

20.078.... 19Fi Hig: PsP 56 13 31 100

Totals for Fz populations | 190 | 48 | 107 = B45

20.060. . . .200/149 X 19F,\HinPsR 16 6 | 28

20.061. S; 200/14 Q X 19FiHinPsP 12 12 | 25

Totals for back-crosses to white | 28 18 | 53 |

190 carmine: 48 pink:107 white, are to be compared with

a 9:3:4 expectation of 194 carmine:65 pink:86 white.

The results from the back-crosses, 28 carmine : 18 pink: 53

white, are to be compared with an expectation based on

the 1:1:2 ratio of 25 carmine:25 pink:49 white. Pink

is again deficient and white in excess, but not to such an -

extent as to give significance to the figures. Further data

from F, families would be desirable for completion of

the analysis. Thus far the data are in agreement with

those presented by Allard for carmine versus pink and

white, and they support the conclusion that his carmine

variety is identical in its main genetic flower color factors

with ours.

No. 639] INHERITANCE IN NICOTIANA TABACUM 303

The further question now arises as to whether there

are any phenotypic differences between carmine and red.

There is a detectable difference between the flower color

of macrophylla and that of purpurea, for the former has

distinctly more of a purplish tinge and is not quite as

intense in coloration as the latter. But these two varie-

ties differ genetically in a large number of other charac-

ters. It is not possible, therefore, to decide the question

by direct examination, because any distinctions which are

found to exist may depend upon differences in the

residual genotype rather than upon the specific factor

differences which we have studied. Obviously the most

satisfactory material for determining the differences

between the two colors would be two varieties which had

the same residual genotype, but the establishment of

such varieties would entail the expenditure of a consider-

able amount of time and labor. We can, however, obtain

some evidence on this problem by comparing the red F,

segregants of the Cupa-MacropHyLua series with the

carmine ones from the Cusa-Purpursa series. In both

cases there was a certain amount of variation in intensity

of coloration in the F, classes, but it was found that, if

they were mixed together, it was impossible to separate

them again into red and carmine. In cases involving

both classes in the same experiment, they would doubtless

have to be considered as making up a single phenotype.

We have been interested in determining experimentally

whether the morphological similiarities of existing

Tabacum varieties might safely be taken as an index of

phylogenetic affinities. Thus Setchell,” commenting on

the relationships of purpurea, states,

There are combined in this plant characters of our N. angustifolia as

to petiole, N. Tabacum var. brasiliensis as to cucullate tip, tallness, and

perhaps also the wing on the petiole, and N. Tabacum var. macrophylla

as to flowers.

It is very natural to regard the sharply constricted leaf-

7 Loc, oit., p. 11.

334 THE AMERICAN NATURALIST [Vou. LV

base of purpurea as a modified petiolate condition, but as

a matter of fact our studies have shown that its affinities

in this respect lie closer to the sessile leaf type gentically,

to which it is recessive, than to the true petiolate class

which is dominant to the sessile type. In the present

article we show further that the flower colors of macro-

phylla and purpurea are distinctly different genetically

and their similarity in appearance can not be regarded

as an indication of phylogenetic relationship. It is, there-

fore, evident that any taxonomic system which proposes

to portray the phylogenetic affinities of the polymorphic

assemblage of Tabacum varieties must be derived from

genetic studies of character differences.

Allard has suggested the use of these flower color forms

for instructional work in genetics. The demonstration

of these additional relations increases their interest and

value for such purposes. Among other points of interest

a cross between macrophylla and purpurea should give

- a carmine F; and the rather unusual F, segregation ratio

of 13 carmine (and red):3 pink. We have verified the

~ production of carmine F, in this cross, but have not yet

grown the F, progeny. The ease of hybridization, the

readiness with which large quantities of guarded seed

may be secured, and the extremely long period over which

the seed of tobacco retains its viability may be urged as

additional advantages in its utilization. Where green-

house and garden space is available for their growth—

plants may easily be grown to maturity in six-inch pots—

these varieties and their hybrids would provide excellent

material for practise in hybridization and for demonstra-

tions of segregation and unique character interrelations.

While there is a certain amount of variation within the

several phenotypes here considered, viz., carmine, red,

pink, and white, it has not been found to interfere

seriously with segregation into the main color classes.

AN ANALYSIS OF THE RELATION BETWEEN

GROWTH AND NUCLEAR DIVISION IN A.

PARASITIC INFUSORION, OPALINA SP.

R. W. HEGNER, Pu.D.

SCHOOL OF HYGIENE AND Pusiic HEALTH, JOHNS HOPKINS UNIVERSITY

AND |

HSIANG-FONG WU, M.D.

NATIONAL MEDICAL COLLEGE OF PEKING, CHINA

Tuts investigation was undertaken for the purpose of

analyzing the relation between growth and nuclear divi- |

sion in a species of Opalina of the frog during the growth

period in the tadpole. ‘The multinucleate condition of

Opalina and the absence of cell walls render it of par-

ticular value as material for the study of the phenomena

involved in nuclear division and growth. The specimens |

used in our investigations were obtained by Dr. Charles

E. Simon from tadpoles collected at Chester, Nova Scotia,

during the summer of 1920. Unfortunately we are unable

to state either the species name of the Opalina or that of

the host. Dr. Maynard M. Metcalf, who has examined

the slides, thinks the Opalina is probably an undescribed

species. The material was well fixed in Schaudinn’s

solution and beautifully stained with iron-hematoxylin.

A sufficiently large number of specimens (455) were

drawn with a camera lucida so as to furnish reliable re- -

sults when measurements were treated by statistical

methods. The area of the drawings was determined with

a planimeter and the correlation with the nuclear number

determined. Table I is the correlation table for the

nuclear number and area of 341 specimens. The area of

the drawings, which were made at a magnification of 650

1 From the Department of Medical Zoology, School of Hygiene and Publie

Health, Johns Hopkins University.

335

336 THE AMERICAN NATURALIST [Vou. LV

diameters, is given in square millimeters. The coefficient

of correlation is remarkably high, namely, .755 + .016;

this proves that an increase in size is accompanied by a

corresponding increase in nuclear number. ‘The rest of

the specimens that were measured, 114 in number, were

drawn at a magnification of 1400 diameters. The coeff-

cient of correlation of this lot was found to be .875 + .015.

TABLE I

CoRRELATION TABLE FOR NUMBER OF NUCLEI AND AREA OF 341 SPECIMENS

he area is given in sq. mm. and obtained from camera-lucida drawings

made at a magnification of 650 diameters. Coefficient of correlation

.755 + .016,

Number of Nuclei .

Area

3 4 5 6 ra 8 9 110] 11 | 12 113/1415 | 16 | 27 | 28 | 29

300 + 1 1

400+ }/1/1)]2 1 5

500 + s Bes ag a ee i 4

600+ 11/7/7138] 1 19

700+ AO tke | oe 24

800 + 2a 113-16 1 Ss 1 36

900 + 6 125; 816)1 46

1000 + 1 j|13 {14/91/18 45

1100+ 2 3:12 20; 64,1 48

1200+ kha bob See eee 24

1300+ ; 1/1 UA ct Sad abr ek ee al a S as i 2 21

1400 + PEP sy 4) 6} 4:2 20

1500 + 1 Ard We a te 11

1600 + A Vn ON ery a a i 1 9

1700+ 1 1 2431 7 | 4

1800 + 1 212 1 } 6

1900 + | faa ae i Eac buat a | 7

2000 + 1 | 1

2100 + | 5 a ae 2 4

2200+ | 1 1

2400 + 1 be E 3

2800 + 1 1

3200 + 1 | 1

| | |

3 |19 40 75 |54 i650 132 126 116 1915)513],1) 1211 | 1 341

Metcalf? has pointed out that in multinucleate Opalinas

the nuclei within a single specimen may be in different

stages of division at one time. This we have found to

be true also of the nuclei during the growth stages in the

tadpole—a condition that has enabled us to analyze with

Sameer st M., 1909, ‘‘Opalina,’’ Arch. f. Protist., 13: 195-375. Es-

pecially p. 2

No. 639] GROWTH AND NUCLEAR DIVISION 337

considerable accuracy the exact relation between cyto-

plasmic mass and nuclear division. For example, among

the specimens with four nuclei, there were a few with

three ‘‘resting’’ nuclei and one nucleus in division (Fig.

3); obviously one of the four nuclei is undergoing divi-

sion before its three sisters. If the sum of the areas of

a number of specimens in which there are four nuclei of

equal size (Fig. 2) is divided by the total number of

nuclei, a fairly accurate idea may be obtained of the

amount of cytoplasm associated with each nucleus. Ac-

cording to the nucleo-cytoplasmie relation theory? an

increase in the amount of cytoplasm as compared with

the amount of nuclear material furnishes the stimulus

which initiates nuclear division. A comparison between

specimens with four equal nuclei, and specimens with

four nuclei one of which is undergoing division, should

reveal approximately the increase of cytoplasmic sub-

stance necessary to inaugurate nuclear division. A num-

ber of cases of this sort were available in our material

and were studied with the following results.

Table IT shows the relations between area, and number,

volume and surface of the nuclei in the 207 specimens that

could be used for this purpose. The measurements were

made of camera-lucida drawings at a magnification of

650 diameters. Beginning with the group of 15 at the

top of the table we can make the following comparisons.

1. Fifteen specimens, each with 4 equal nuclei (Fig.

2), have an average area per nucleus of 176.1 sq. mm.; 10

specimens, each with 3 equal nuclei and a fourth nucleus

in division (Fig. 3), have an average area per nucleus of

185.0 sq. mm. The specimens in which division has been

initiated have an average area per nucleus 8.9 sq. mm.

greater than those with an equal number of nuclei, none

of which are in division. We have used the area through-

3 For a recent discussion of this theory see Hegner, R. W. 1920, ‘‘ Rela-

lations between Nuclear Number, Chromatin Mass, Cytoplasmic Mass, and

Shell ep aR in Four Species of the Genus Arcella,’’? Jour. Exp.

Zool., 1-95,

338 THE AMERICAN NATURALIST [Vou. LV

TABLE II

TABLE SHOWING THE RELATION BETWEEN AREA AND NUMBER, VOLUME AND

FACE OF THE NUCLEI

The area, volume and surface of the nuclei were oat vias from camera

drawings made at a magnification of 650 diameters

Average Average

Average | Average | volume of — slog Aver Average

Area in | Area per | Nuclei in Surface S Surface per

Number Sq. Mm. Nucleus in Cubic in ‘Cubic cea frag Nucleus in

of Speci-| Number-of Nuclei | at Magni- -Mm. | mm. at . at q. Mm.

ipaa neson |" aton | Mo0F | Magee at Mag “hese.

atea ce | une | eee Stam. | ate

15 4 (equal in 704.2 176.1 | 215.49 53.87 261.32 65.33

size

10...} 4 (one in 739.9 | 185.0

division)

I3.. 5 (two small) 851.5 170.3 257.11 51.42 321.10 64.22

16...| 5 (equal in 860.5 172.1 281.91 56.38 345.01 69.00

size)

17...| 6 (two small) 951.4 158.6 273.69 45.62 363.85 60.64

49...) 6 (equal in 926.7 154.5 264.83 44.14 352.89 58.81

size)

5...| 6 (onein 1,032.5 172.1

division

11...| 7 (two small) 993.5 | 141.9

31...| 7 (equal in 1,066.7 152.4 269.17 38.45 379.68 54.24

size)

10...) 7 (onein $176.7 | 168.1

division)

30.. 8 equal 1137.8 | 141.0 336.53 42.07 458.93 57.37

207

out our work as a measure of cytoplasmic mass, hence it

appears from the results of our measurements that an

increase in mass per nucleus represented by an increase

in area within the limits of 8.9 sq. mm. is the stimulus

that initiates nuclear division. The exact mass of cyto-

plasm represented by this increase in area of 8.9 sq. mm.

might easily be obtained under more favorable cir-

cumstances.

2. When we compare the-measurements of the 10 speci-

mens with 4 nuclei, one of which is dividing (Fig. 3) with

13 specimens of the stage immediately following, with 3

large nuclei and two that have just reorganized after

division (Fig. 4), we find that although the latter average

111.6 sq. mm. larger per specimen the average area per

nucleus is 14.7 sq. mm. less. Thus there has been an

No. 639] GROWTH AND NUCLEAR DIVISION 339

actual increase in size but a decrease in the mass of cyto-

plasm associated with each nucleus.

1 16 17

Figs. 1-17. Outline drawings of stages in the growth of Opalina sp.

made with a camera lucida at a magnification of 650 diameters and reduced

to a magnification of 325 diameters.

3. During the growth of the two small nuclei to their

full size (Figs. 4-5) the size of the organism increases

from an average area of 851.5 sq. mm. per specimen to an

average area of 860.5 sq. mm., or an average area per.

nucleus of from 170.3 sq. mm. to 172.1 sq. mm. Although

an increase in size has taken place, the average area per

nucleus of 172.1 sq. mm. in specimens containing 5 full-

340 THE AMERICAN NATURALIST ` [Von. LV

sized nuclei is less than that of specimens with 4 full-

sized nuclei, t.e., 176.1 sq. mm.

4. A further increase in the average size of the speci-

mens occurs between the stage with 5 nuclei of equal size

(Fig. 5) and that with 6 nuclei, two of which have just

emerged from mitosis. Measurements give for the for-

mer an average area of 860.5 sq. mm. and for the latter

951.4 sq. mm. The specimens with the two small nuclei,

however, possess as before (see (2)) a lower average

area per nucleus, t.e., 158.6 sq. mm. as compared with

172.1 sq. mm. in specimens with 5 full-grown nuclei.

5. The measurements of the next stage, i.e., specimens

with 6 nuclei of equal size (Fig. 7), are more difficult to

explain, since the average area of the specimens (926.7

sq. mm.) is actually less than that of the younger speci-

mens (see (4)) with 4 large and 2 small nuclei, and the

average area per nucleus falls from 158.6 sq. mm. to 154.5

sq. mm. These results may be due to a thickening of the

entire animal which would increase the mass and tend

toward a decrease in area or the nucleo-cytoplasmie rela-

tion may change as the animals become older. That there

is an actual decrease in the average area per nucleus as

growth proceeds is indicated by the measurements of

later stages as given in Table III. This table shows a

decrease per nuclear area from 186 sq. mm. in specimens

with 4 nuclei to 96.8 sq. mm. in specimens with 29 nuclei.

That this decrease is gradual is indicated when averages

are made of three successive groups containing each a

larger number of specimens. Thus the 4, 5, and 6 nu-

cleated groups containing 134 specimens have an average

area per nucleus of 173.8 sq. mm., the 7, 8, and 9 nucleated

groups containing 136 specimens have an average area

per nucleus of 144.4 sq. mm., the 10, 11, and 12 nucleated

groups containing 51 specimens have an average area per

nucleus of 143.5 sq. mm., and the 13, 14, and 15 nucleated

groups containing 13 specimens have an average area

per nucleus of 128.3 sq. mm.

No. 639] GROWTH AND NUCLEAR DIVISION 341

TABLE III

TABLE GIVING THE AVERAGE AREA PER SPECIMEN AND PER NvcLEvs oF 338

SPECIMENS DRAWN WITH A CAMERA LUCIDA AT A MAGNIFICATION

OF 650 DIAMETERS

Average Area Specimen in Average Area Nucleus

user vials ae of Sa. Mm. a Magnineation in sa. Mm; = Maguieation

AES oes + 744 186.0

As 5 891 178.2

FOG aes 6 44 157.3

BAe ee rá 1,069 | 152.7

a EENE 8 1,150 143.7

Oe E A 9 1,231 136.8

2655.3: 10 1,395 139.5

AG ey ea 11 1,635 148.6

Oye ora | 12 1,709 142.4

te 13 1,572 120.9

os ements 14 1,801 128.6

Be oa 15 2,032 135.4

| ee 16 2,452 153.2

I Spa Le 1,806

E ok gi 28 3,277 117.0

i Ree | 29 2,806

6. By the time one of the nuclei of the six-nucleated

stage has been stimulated to division the average area

per nucleus increases again to 172.1 sq. mm. After this

division is completed and seven nuclei are present, two

of them small, the average area per nucleus, as was to be

expected, decreased to 141.9 sq. mm. During the period

necessary for the two small nuclei to reach their full size

(Fig. 8) the area increases again to 152.4 sq. mm. A fur-

ther increase to 168.1 sq. mm. occurs by the time sufficient

growth takes place to stimulate one of these seven nuclei

to divide, and a decrease (to 141.0 sq. mm.) again takes

place when this stage evolves into that with eight nuclei

Fig 9).

The two curves in Fig. 18 show clearly the increase in

area per nucleus up to the point where one nucleus di-

vides, then a conspicuous decrease following nuclear

division, and subsequently an increase during the period

when the nuclei resulting from division regain their full

size, ending in a size at which the area per nucleus is

approximately that present at the beginning.

342 THE AMERICAN NATURALIST [Vou. LV

Figure 19 illustrates the fact that the size of the entire

specimens increases during nuclear multiplication and

growth, but that the area per nucleus remains almost

constant.

180 X

4 Pai

A Ae gae

+ 170

E 160

2 Vs

p s

$ 150 A

140 x

Normal One divid- ro Normal

ing small

Condition of nuclei,

Fig. 18. Curves showing the changes in area per nucleus that accompany

changes in nuclear number and condition, The numbers 4, 5, 6, and 7 indi-

cate the number of nuclei present.

7. The series of measurements of these specimens af-

fords an explanation of the reason why nuclear division

in Opalina is not synchronous. According to the nucleo-

cytoplasmic relation theory, as soon as the mass of cyto-

plasm has increased to a certain point nuclear division

is initiated. The necessary increase to furnish this

stimulus in Opalina may be determined approximately

from our data by comparing measurements of specimens

in which the nuclei are all equal in size with those in

which nuclear division has been inaugurated. Such a

comparison gives the following results.

No. 639] GROWTH AND NUCLEAR DIVISION 343

Average rong per Difference in Area— Amount

ucle Nuc: Necessary to Stimulate Division

4 ( oii eS ake 76.1 sq. mm.

4 (one in ' division) ie 0 sq. mm. 8.9 sq. mm.

G (equal) ei. 154.5 sq. mm.

6 (one in division) 172.1 sq. mm. 17.6 sq. mm.

7 (OGUL) dans 152.4 sq. mm.

7 (one in division) 168.1 sq. mm. 15.7 sq. mm.

These figures, of course, indicate only the relative in-

crease necessary to stimulate nuclear division; the actual

increase could be determined by measuring accurately the

mass of cytoplasm in each case.

5 300

izi

HEF

ev ?

4 5 § 6 7

Equal One dividing fwa small Equal To small Equal One dividing Two smali. Emal Cno dividing Equal -

` Fic. 19. Curves showing the increase in the area of specimens with in-

crease in age and the constancy of the area per nucleus

In looking over our camera drawings it was noticed

that usually only one nucleus was in division in any one

specimen (Figs. 3 and 6) and that in many cases two of

the nuclei in a specimen were smaller than the rest (Fig.

4), indicating that they were daughter nuclei that had

just emerged from mitosis. Of a total of 137 specimens

in which nuclei were found in division, 109 contained one

division figure, 19 contained 2, 8 contained 3, and 1 con-

tained 4. Furthermore, those containing more than

344 THE AMERICAN NATURALIST [Vou. LV

one division figure were usually older than those contain-

ing one only. ‘Thus the average nuclear number of

specimens with one dividing nucleus was 7.2, with two

dividing nuclei, 8.1, with three dividing nuclei, 9.3, and

with 4 dividing nuclei, 10. These data favor the conclu-

sion that the stimulus that initiates nuclear division acts

as a rule on only one nucleus at a time and that the divi-

sion of this nucleus restores the nucleo-cytoplasmic ratio.

When this ratio is again disturbed by an increase of the

cytoplasmic mass another nucleus is stimulated to divide.

Division of two or more nuclei synchronously may be due

to the more rapid growth, the larger specimens in which

this usually occurs, or to the greater chances of two or

more nuclei reacting to the division-stimulus when a large

number of nuclei are present in a single specimen. There

is some evidence that the nucleus that undergoes division

is the one with the greatest amount of cytoplasm sur-

rounding it, but this could not be determined definitely.

No regular distribution of the nuclei was evident. It is

interesting to note in this connection that during the em-

bryonic development of many animals nuclear division

occurs in all cells at nearly the same time. This is espe-

cially interesting in the case of certain insects, in the eggs

of which nuclear division proceeds synchronously without

the intervention of cell walls until thousands of nuclei

are present in a single egg.* An increase of cytoplasm

over nucleus may also, in these insect eggs, stimulate

nuclear division, since after each division the mass of

cytoplasm surrounding each nucleus is increased by the

addition of new material elaborated from the yolk sub-

stance in which it is situated.

8. The average total volume of the nuclei of certain

specimens, average volume per nucleus, average total

area of the surface of the nuclei and average area of the

surface per nucleus were measured in cubic millimeters

and square millimeters from our camera drawings which

4 Hegner, R. W., 1914, ‘‘Studies on Germ Cells,’’ Jour. Morph., 25: 375-

509. Especially Dp. 408—413.

No. 639] GROWTH AND NUCLEAR DIVISION 345

were magnified 650 diameters. These data are shown in

Table II. They indicate one point of considerable inter-

est and that is the fact that as the number of nuclei in-

creases their average volume and surface decrease.

8q. mm.

Volume per Surface per

nucleus in nucleus in

sq. mm.

+ 5 5 6

Equal Tro small Equal Two small Equal

x Number and condition of nuclei,

7 8

Equal Equal

1G. 20. These three curves show that decreases in the volume and surface

of the nuclei are accompanied by decreases in area per nucleus.

Thus in specimens with 4 nuclei of equal size each nucleus

has an average volume of 53.87 cu. mm., in specimens

with 6 nuclei the average volume per nucleus decreases

to 44.14 cu. mm., and in specimens with 8 nuclei the aver-

age volume per nucleus decreases still further to 42.07

cu. mm. Similar results were obtained from measure-

ments of the surface of the nuclei, but the decrease is not

so great since the volume decreases as the cube whereas

the surface decreases only as the square. As the table

(II) shows the average surface per nucleus in sq. mm.

decreased from 65.33 sq. mm. in specimens with 4 nuclei,

to 58.81 sq. mm. in specimens with 6 nuclei, and 57.37

sq. mm. in specimens with 8 nuclei. This decrease in vol-

ume and surface may account for the fact noted pre-

viously (in (5)), that the area per nucleus decreases in

specimens with nuclei all equal in size as the number of

346 THE: AMERICAN NATURALIST [Vou. LV

nuclei becomes greater with advancing age. Since the

volume of the nuclei is less in these older specimens the

amount of cytoplasm associated in normal nucleo-

cytoplasmic relations with them is less and the area of

the specimens per nucleus decreases accordingly.

The curves in Fig. 20 bring out clearly the relation be-

tween area and volume and surface of the nuclei during

the growth period. The average area per nucleus de-

creases as the number of nuclei increases, but at the same

time there is a corresponding decrease in both volume

and surface of the nuclei, thus maintaining approximately

the initial relation between nucleus and cytoplasm.

SuMMARY

(a) A high correlation exists between nuclear number

and cytoplasmic mass (as indicated by area) during the

growth of Opalina sp. The coefficient of correlation in

one lot of 341 specimens was .755 + 016 and in another

lot of 144 specimens was .874 + .015.

(b) By comparing the area of various stages with the

number, size, state of division, volume and surface of the

nuclei the following conclusions were reached. (1) Nu-

clear division is stimulated by an increase of cytoplasm

that may be determined approximately. (2) As the

organisms increase in age the nuclei decrease in volume

and surface; this is accompanied by a corresponding de-

_ crease in the area per nucleus, indicating that the nucleo-

cytoplasmic relation is maintained. (3) Nuclear division

is not synchronous because one nucleus is usually stimu-

lated to divide before the others, and this division is suffi-

cient for the time to reestablish the normal relation be-

tween nuclei and cytoplasm.

AMERICAN FOLLICULINAS: TAXONOMIC

NOTES

E. A. ANDREWS

THE JOHNS HOPKINS UNIVERSITY

Tue ciliated infusorian Folliculina, ‘‘ the bottle ani-

malcule,’’ was first recorded by O. F. Müller in 1781 and

was by him described amongst 75 Vorticellas as one liv-

ing in an ampulla or bottle. The name Folliculina was

suggested by Lamarck in 1816, yet he, having no personal

observations of the creature, placed it among the rotifers.

However, the true affinities of the bottle makers with the

stentors became evident to Claparede in 1858, though

members of the group have been referred by others to

such genera as Cothurnia, Ascobius and Vaginicola from

insufficient knowledge of the animal within the bottle.

Two observers emphasized the nature of the bottle maker

rather than the bottle itself in seeking to establish for it

the generic names Freia and Lagotia; the former given

it by Claparede as having the ‘‘ forme gracieuse et ele-

gante d’une Freia,’’ and the latter by Wright, in the

same year, from the long lobes of the animal that resem-

ble the ears of a hare.

The only careful studies of the animal have been made

by Stein in 1867 and by Mobius in 1887. The latter was

inclined to regard all the then known species as local

varieties of the original Vorticella ampulla O. F. Müller.

But recently Carl Dons in Norway has made very minute

study of the bottles as found in many localities and has

come to the conclusion that these alone may be used as

sufficient basis for establishing species, even without the

animal, which, to be sure, is rarely preserved in museum

material. He would recognize some ten species, most of

which he finds in Norway, but pi of. which are wide-

spread over the world.

These ten species he proposes to distribute amongst

four new genera as follows: The original forms of Müller

347

348 THE AMERICAN NATURALIST [Vou. LV

with simple bottles of very wide diameter and short plain

neck with no lips retain the name Folliculina. The type

species F. ampulla being found in Norway, Denmark, at

Kiel and in the Adriatic, while a peculiar species F.

paguri from the west French coast was described by

Giard in 1880 as Pebrilla paguri.

Forms with more elongated bottles and with longer

necks, often marked with a spiral ridge, as well as pro-

vided with a collar or lip to the mouth of the bottle, he

calls Semifolliculina. The best described is his S. gi-

gantea from Norway, as well as the South Polar Sea,

while S. boecki is the name chosen for an old and widely

distributed form occurring in Norway, Denmark, the

Adriatic, North America, Formosa, as well as both North

and South Polar Seas. SS. similis is an aberrant form

from the South Seas and S. spirorbis is the smallest

form, with very narrow neck, found both in Norway and

upon material from North America, East coast.

Long, straight bottles with the bottom thickened as a

falsification of the actual content as judged from the out-

side are placed in the new genus Pseudofolliculina, repre-

sented by the large P. mellita from the South Polar Seas

and by the similar species P. arctica from the north of

Norway. The remaining bottle animals have been de-

scribed as having some sort of a valve or set of mem-

branes in the neck of the bottle, and these Dons includes

in the genus Parafolliculina. P. amphora he has de-

scribed minutely from Norway and from Iceland, and

P. violacea is a well-known form from west France, the

Adriatic, west Australia, as well as Norway.

All these bottles are minute, less than a millimeter

long, and though made of a chitin-like material that ad-

mits of long preservation in museum jars, they have gen-

erally been overlooked, though so common in many parts

of the world from the surface down to considerable

depths attached to solid objects, such as shells, stones

and plants, either singly or in large aggregates or settle-

ments. Moreover, certain species have been described

No. 639] AMERICAN FOLLICULINAS 349

in England, France and Switzerland as occurring in en-

tirely fresh waters. The above review of the known

species and their distribution does not do justice to Stret-

hill Wright’s careful description of the animals that

make the cases or bottles on the British coasts where he

found species not entirely synonymous with the above-

cited ten.

The first record of the occurrence of any of these bottle

animaleules along the American coasts seems to have

been that of Leidy, who, in July, 1859, at Newport, Rhode

Island, found attached to Anomia-serpula on dead clam

shells dredged by Mr. Powel ‘‘ a singular and beautiful

animal ’’ in a vase-like tube and with the same general

structural appearance as that of the stentors.

He recognized its resemblance to Chetospira mueleri

Lachman and its alliance with the stentors and suggested

the name Freyia Americana for it. Later, according to

Ryder, Leidy considered his species to be the same as the

European Folliculina ampulla. It was not till 1880 that

the bottle animaleule was again observed in American

` waters, and then Ryder on the western shore of the Ches-

apeake Bay, probably at St. Jerome, St. Mary’s County,

found a different form of bottle and animal which he

identified with the Freia producta of Strethill Wright.

The occurrence of this form of bottle animaleule in

other parts of the estuaries of the Chesapeake was

pointed out in 1914, 1915 by Andrews under the name

Folliculina, but without determination of the species de-

scribed. Meantime it was known to workers at Woods

Hole, Mass., that Folliculina occurred there also, though

no published accounts appeared. Dons has recently

mentioned the observation of a Folliculina upon material

from the east coast of America and refers another species

to this coast, probably from the above account of Leidy.

We know merely that the animal has been found at

Woods Hole and Newport and from waters of the Chesa-

peake. Considering that the animal is so very widely

distributed in Arctic, Antarctic and northeast Atlantie

350 THE AMERICAN NATURALIST [Vou. LV

regions, as well as the Adriatic, it is probably common

along many American coasts where as yet overlooked.

In seeking to refer the different forms of the bottle

animaleule found thus far in American waters to known

or new species we are confronted with ignorance of the

anatomy and the life history of the animals and thrown

back chiefly upon the secreted bottle or case, since it is

this alone that is commonly preserved, and since this also

presents preserved characters of form and proportions.

In the life history as known there has been no restric-

tion of the possibilities of form and size change possible

to a single individual. We know that there are free-

swimming forms as seen by Claparéde and more fully

studied by Wright and confirmed by Andrews and by

Penard.

In some cases these swarmers arise from fission of the

parent as Mobius found to be true, but in many instances

the free-swimmers that swarm out are only the old forms

transformed into simpler larval shapes that have later

to make new bottles and then become again complex in

structure. In the former case the result of fission is one

free-swimmer of small size and one remnant individual

left to complete its perfect organization in the old bottle.

There are thus large and half-sized forms: both perfect

sedentary individuals and imperfect swimming larve.

Moreover, we find that not only may each individual

greatly change its shape from muscular contraction, but

may change both shape and bulk under conditions other

than the optimum of good feeding environments. Noth-

ing is known of any conjugation and any influence this

may have upon form and size.

While it is easy to assume that all the known forms of

bottles may prove to be the products of but one and the

same species widespread all through the various oceans

of the world, evidence for this is lacking, and not having

sufficient anatomical basis for classification, we must as

a practical expedient adopt the plan of Carl Dons and

determine the species by the form and size of the bot-

No. 639] AMERICAN FOLLICULINAS 831

tles—ceding the point that these species may well have

but a very temporary and artificial value.

That the species determined from characters of the

temporary dwellings of the animal may, however, prove

to be real species is indicated by the following consid-

erations.

The bottle or case is a secretion from the surface of

the animal, and, as seen by Wright in 1861, the bottom

part, or sac, as we call it, is made first, and then the neck

or tube added, and finally the lip at the mouth of the

bottle. In making the sac the animal flattens its body

and assumes the size and the form that the sac will have

when it is poured out and hardened all round about the

body in this shape, leaving only the blunt anterior end of

the body free from secretion, so that the hardened sac

comes to have a hole in the anterior end and is a bilater-

ally symmetrical product duplicating the form and pro-

portions of the animal as if the latter had been cast in

the sac as in a mold. The form of the sac is the form of

the animal at that period of its life cycle. Subsequently

the size, length, spiral ornamentation, if any, and the per-

fection of the funnel or lip of the tube are all representa-

tions of the habit of the animal in the consecutive phases

of the manufacture of the tube. The angle that the neck

of the bottle, or tube, makes with the body of the bottle,

or sac, is fixed by the degree to which the animal con-

tracts its anterior part to rear it up away from the sur-

face of attachment and general plane of the sac; the

diameter of the tube is that of the head end of the animal;

its circular section is that of the head end; its length is

that of the gradual elongation of the entire animal which

carries the head end gradually ever farther away from

the foot end till the maximum length is attained; the

spiral character of the tube, when present, is determined

by the rotation of the head end and by the localized and

radially differentiated selective secretions and contrac-

tions of parts of the head end; the final lip of the tube is

added by special change of shape of the head end which

352 THE AMERICAN NATURALIST [Vou. LV

assumes a mushroom form and secretes from its under

surface. At any one period of making of tube the tube

expresses the resultant of the two components, contrac-

tion and secretion by the body at that moment.

While the entire animal never has the shape of the

tube, yet each part of the tube exactly fits the head end

of the animal as it progresses away from the foot in co-

ordination with the elongation of the whole animal. The

completed structure represents a solidification of the

form rhythms of the animal as does the shell of a gas-

teropod or the successive exoskeletons of a lobster or the

hard envelope of a rhizopod, and differs from most of

these chiefly in that the animal only temporarily assumes

the forms expressed by the dwelling, and later lives freely

movable in the dwelling and capable of leaving it by

simply detaching the foot-end from the bottom of the sac.

How precisely the bottle represents the animal was

Seen in one instance when camera drawings of two suc-

cessive bottles made by the same animal exactly coin-

cided. If, then, the different shapes and sizes of bottles

do not mean different species, it is because the different

activities and forms of these animals are not specific, but

only varietal or individual differences, or differences due

to changing conditions, such as food, or to different suc-

cessive internal states connected with internal rhythms.

In ignorance of the possible changes of form any animal

may go through, we may, for practical purposes, follow

Carl Dons in describing the bottles as expressions of

‘forms that may be specific in value.

Of the anatomical characters in Folliculina that may

be made use of in classification, the nucleus has been em-

phasized by Dons, who would regard a moniliform nu-

cleus as the attribute of one group of folliculinas as rep-

resented by Mueller’s original species, while all others

have a single lobed simple nucleus. But this is of no

avail, for in the first place, the folliculina thoroughly

studied by Möbius had moniliform nucleus but its dwell-

ing can not at all be confounded with that of Mueller’s

No. 639] AMERICAN FOLLICULINAS 353

Folliculina, and Ryder expressly mentions the long-

beaded nucleus in the form Freia producta, which is most

remote from Mueller’s simple form; and, in the second

place, I find that the commoner Chesapeake Folliculina

has a moniliform nucleus and is by no means close to the

Folliculina of Mueller.

Thus the moniliform nucleus is not ratiioted to ani-

mals in sacs of the Mueller type. Moreover, observa-

tion shows me that in the commoner Chesapeake Follicu-

lina the nucleus may pass from the moniliform shape to

more and more simple shapes, resembling the elliptical

nucleus of so many other species; that is, just as was to

be expected from the findings of Johnson in Stentor, the

form of the nucleus is not constant, but a very long monili-

form nucleus may fuse into a short elliptical shape. The

same change was observed by Sahrlage.

Whether the nucleus is condensed or nodulated can

then be but a poor basis for classification of the Folli-

culinas.

Till recently no micronuclei have been deseribed in

Folliculinas, though known in Stentor, and Carl Dons has

used this as basis for separating the Folliculinas from

the Stentor family ; however, in two forms of Chesapeake

Folliculinas I find minute darkly staining bodies associ-

ated with the macronucleus which may well be micro-

nuclei, though their function has not been observed.

The only other anatomical character available seems to

be the form of the anterior part of the body which is in

some Folliculinas a funnel, and in others a funnel with

two sides, more produced so that they may even form

long arms likened by Ryder to obstetrical forceps, and

again, in another species, by Wright, to the long ears of

a hare.

But here again I find that an animal may have in its’

periods of maximum expansion and feeding activities

exceedingly long arms, which in retracted states during

adverse conditions may be very greatly reduced and mod-

ified in form and proportions, and there are also transi-

354 THE AMERICAN NATURALIST — [Vou. LV

tion stages constantly found in which the arms are either

being regenerated, or ‘‘ redifferentiated,’’ or reduced, to

vanishing point.

It is precisely the great development of the edges of

the funnel that makes the folliculina an advance upon the

simpler state found in the stentors, so that the more

simple folliculinas are those with a peristome readily

referable to the stentor state, while the very highly dif-

ferentiated folliculinas with extremely long ligulate lobes

right and left from the edge of the funnel are the most

remote from the stentors.

While in classification the relative amount of develop-

ment of these lobes is evidently of great importance, it

will require observation of many living specimens, as a

rule, to determine whether a given specimen has short

lobes from its present stage of development in the indi-

vidual life cycle or from its permanent place in the stage

of evolution from the stentor-like ancestor. On the other

hand, the presence of long ligulate lobes will at once de-

termine a high stage of individual and phyletic advance

and place the specimen in the highest group of anatomical

development.

But until the possibilities of change of form in each

individual are known, and in the probable possession of

only poorly preserved specimens, the practical expedient

will be to adopt much of the procedure of Dons in mak-

ing use of the forms of the dwellings in the description

of what may for the present be regarded as species within

the group of Folliculinas.

Relying, then, largely upon the bottles as indicative of

specific differences, in the American Folliculinas, so far

known, we may tentatively adopt the general subdivisions

of Carl Dons, retaining the genus Folliculina for the

- small, very wide sacs with short simple tubes.

No form of this restricted genus has thus far been re-

ported from the American coasts. Whether such forms

are anything more than starved, depauperate or imper-

fectly developed Folliculinas may well be doubted.

No. 639] AMERICAN FOLLICULINAS 355

The genus Semifolliculina he has invented for bottles

of narrower form with longer necks, often spirally orna-

mented and provided with a collar or lip. He says that

one of the most widely distributed, Semifolliculina boecki,

occurs in North America, and probably he had in mind

the specimens described by Leidy as Freia americana as

above related. However, the description given by Leidy

speaks of the ‘‘ convolvulus-like mouth °’ of the tube,

which is plainly shown in an unpublished sketch (Fig.

414, Vol. V, of Leidy’s MS. drawings) made by Leidy,

which, with his notes, were kindly copied for me by Pro-

fessor J. Percy Moore, who succeeded in finding it among

unpublished material left by Leidy. This sketch also

shows not only transverse lines on the tube, but longitudi-

nal lines that may be compared to those of Dons’s Semi-

folliculina gigantea (though Leidy may have drawn some

of the lines to bring out curvature). At all events the

very wide lip of the tube recalls the lip of S. gigantea.

The size of the animal stated in Leidy’s published notice

is almost a fifth of a line. In his manuscript certain

measurements would on this basis mean that: length of

animal is 416 » expanded, but 298 contracted; length of

entire bottle 416 »; width of animal 166 », but at narrow

neck below terminal funnel 83; width of expanded funnel

and lobes, that separate ‘‘ like a labiate flower,’’ is 139 pn.

From the sketch the evident nucleus is a rounded mass

that might be 40 » in diameter.

Dons has given measurements of various parts of bot-

tles of Semifolliculina boecki in contrast with those of

S. gigantea from which it appears that the animal seen

by Leidy was rather larger than the tubes of S. boecki;

thus the combined lengths of sac and tube in S. boecki

are 265-410 and for S. gigantea are 250-1000; the width

of sac, 105-135 in the former and 230-300 in the latter.

Leidy’s animal was apparently 166 wide and its bottle

416 long. Thus both width of lips of tube and dimen-

sions as far as known tend to place Leidy’s animal in

S. gigantea rather than in S. boecki. However, the fig-

356 THE AMERICAN NATURALIST [Vor. LV

ure and the description of ‘‘ vase-like tube ’’ emphasizes

the short neck of the bottle and makes the reference to

S. gigantea doubtful.

A second Semifolliculina is known on the statement of

Dons to have been found in material from the eastern

coast of America, and this is his S. planorbis, character-

ized by very narrow tube and wide sac. This is sug-

gestive of some of the apparently depauperate or dwarf

forms occasionally met with in the Chesapeake and may

prove to be but a transitory condition due to conditions

of food or other factors and not a permanent form.

The folliculina found at Woods Hole, Mass., has not

been described. In 1914 Dr. Elmer J. Lund observed

these Folliculinas on stems of Campanularia, Euden-

drium and Bugula from the wharf of the Bureau of Fish-

eries and in letters to me sketched the sac and tube with

expanded lips and absence of spiral. Apparently the

form and proportions are much as in the sketch of Leidy,

but with longer tube. Professor Lund observed the free-

swimming forms several times during the summer in

various cultures and saw the formation of new bottles

in June.

Some preparations of Obelia mounted many years

since, probably at Woods Hole, Mass., about 1888, and

very likely by Professor Brooks, have yielded me several

specimens of Folliculina that are evidently the same

forms as those seen by Lund and probably the same as

those of Leidy. The animals are exceptionally well pre-

served, in dwellings that give the following measure-

ments:

1. One sae attached its whole length to Obelia is 175 p

long and 50 » deep, while the tube arising from it nearly

at right angles is only 50 » above the top of the sae, is

37 » wide and flares out at the lip to a width of 62 ». The

animal drawn into its sac is 162 » long and 35 mm. in

greatest diameter, with spheroidal macronucleus 15 » in

diameter. The ligulate lobes are 75 » long and 12 wide

and plainly show the characteristic adoral zone proceed-

No. 639] AMERICAN FOLLICULINAS 357

ing from the vestibular spiral out to the tip of the left

arm, thence back to the dorsal curve, out to tip of right

arm, and then along some distance dorsal to the ventral

edge of the right arm, to end abruptly at entrance to

funnel.

. 2, Another specimen seen from above has 250 » length

of tube and sac combined (100 tube, 150 sac), width of

sac 75 », width of tube 40 », width of flared lips 75 m.

The macronucleus is 20 by 15 » and the oral lobes the

same size as in preceding specimen; length of animal

250 », greatest width about 60 », since it was but partly

contracted.

3. A third empty case has sac 125 » along shorter dor-

sal side and tube 125 long. Depth of sac 65 and tube

width 37, width of flaring lips 50 ». The axis of the tube

rises about 135 degrees away from the axis of sac.

4. Another empty sac has the same length, but depth

of only 40 », and the tube was but just begun or else

broken off, with diameter of about 35 m.

5. Another specimen has sac 150 long with great width

of 80 », short tube 88 long and 40 wide with lips 67; it

was seated in the branching angle of the hydroid and

faced toward base of hydroid.

The contained animal was much contracted into sae,

150 by 60, with nucleus elongated 15 by 25, with lobes 37

by 14 », and the left lobe terminated in a papilla 2} by

10 p», recalling the Freia stylifer of Wright that was

stated by Ryder to be probably but a variety and which

represents a temporary state as, we see it in Folliculinas

in the Chesapeake Bay.

6. The sixth specimen measured: sac 125 by 80, tube

150 by 40 with lips 55; inhabited by animal remarkably

well expanded, having main body 166 by 40 with nucleus

17 by 22 and lobes stretched out to 92 » and 10 to 12 wide.

This animal contained a large diatom as food, while some

of the others contained masses of detritus as if bacteria

in digestion. Within the sac next the animal are several

nucleated masses, either foreign protozoa or possibly

358 THE AMERICAN NATURALIST [Vou LV

fragments of disintegrated Folliculina arising from divi-

sion and dying.

Comparing these measurements with those given te

Dons for Semifolliculina boecki, we see that they are

somewhat smaller except for the length of tube; but in

the present state of ignorance of limits of species it would

be folly to separate this from the Semifolliculina boecki,

which includes the so similar Lagotia viridis of Wright,

which Dons has separated from the simpler Folliculina `

ampulla of Mueller.

Not making a new species of the Woods Hole form, we

may tentatively refer it to Semifolliculina boecki as ex-

hibited in the above evidence; moreover, the form from

the adjacent region of Newport described by Leidy,

though much like some of Dons’s smaller S. gigantea, may

be the same as S. boecki, and we thus have probably this

one species along the New England coast, together with

the narrow-necked form F alaulina spirorbis as quoted

by Dons.

Yet along with the above six specimens is one that is

aberrant. Its sac and tube in one line stand out freely

from the hydroid, attached only at the base. The tube

enlarges where joining the sac and envelopes it as a-

swelling within which the edges of the sac end as free lip

or inturned: shelf, producing the appearance of a circular

valve standing inward from the wall. This empty case

is thus much like the Folliculina telesto of Laachmann

as figured by Dons from Drobak, page 88, Fig. 2, and

later called Parafolliculina violacea by Dons.

The dimensions of this single empty case are: length,

250; greatest width, 63; length of tube, 82; width of tube

beneath collar, 30; width of collar, or lips, 42; width of

swollen tube where embracing mouth of sac, 51; width of

sac just below this swelling of tube, 40; diagonal width

of inner projecting flange that might be called a valve,

74. Compared with Laachmann’s specimens 200-260 by

60 and with Dons’s 260-310 by 55-90, it closely resembles

the former in size and the latter in appearance and pro-

No. 639] AMERICAN FOLLICULINAS 359

portions, as seen in the above Fig. 2. This single speci-

men agrees very well in size with the six others on same

material, but differs in the narrow lips and the swollen

base of tube embracing the sac, as well as in the attach-

ment of a sac by end only. .

Knowing the normal mode of secretion of sac and of

tube, one is tempted to suppose that a single exceptional

specimen like this may have arisen by some fault in at-

tachment of a swimmer followed by lack of proper

rhythm of secretion, so that sac was attached by end and

not along whole side, and that later the animal abnor-

mally started to secrete a second collar within the old

junction of sae and tube, bulging out the tube while still

soft by pressure of its persisting mushroom and making

an inner rim or shelf to represent an imperfect collar

after it had already made an insufficient one at the mouth

of the abnormally narrow tube. The whole structure

would thus be an abnormal product resulting from slight

abnormalities in secretional activities of the animal after

unusual attitude in attachment. Such an hypothesis for

explaining the telesto shape might be extended to all the

telesto forms seen hitherto by Laachmann and by Dons,

and these are significantly few; thus Laachmann found

but one specimen on material from Sumatra and Dons

found 20 after much search on Eudendrium from the

Adriatie and but about a dozen amidst very many F.

ampulla (i.e., S. boecki) from Dröbak, and none at all

from North Norway. A form occurring but rarrly and

found in Sumatra, West Australia, Norway and Woods

Hole, Mass., may well prove to be but an abnormality

rather than a real species.

The same suspicion of abnormality attaches to four

additional sacs found empty and clustered together on

the above hydroid material from Woods Hole. Appar-

ently unfinished, they are characterized by great breadth

and shortness and by narrow openings where the tube

had not yet been added. With length of 110-113 +, these

sacs were 90 » wide. They thus recall proportions of

360 THE AMERICAN NATURALIST [Vou. LV

Mueller’s Folliculina, but the form is evidently so dif-

ferent in each of the four that they may be thought of as

greatly shortened Semifolliculina boecki in which a pre-

mature change of axis in secreting has made the sac very

deep where passing over into the tube, which is then

expressed as part of the sac and left unfinished. The

whole is like a short club foot with swollen ankle.

For the present all the known normal material from

the coast of New England (except that referred to by

Dons as S. planorbis) may be regarded as belonging to

those forms described by the name Semifolliculina boeckt

and closely akin to the original Folliculina ampulla of

most authors.

Turning now to the southern coast, the bottle animal-

cules first seen by Ryder in the Chesapeake were referred

by him to Freia producta of Strethill Wright, September

3, 1880. This animal 1,000 » long extended and 100 when

contracted has a dwelling compared to a stocking with

spiral ribbon of four to twenty-four turns to the right.

Found in vast numbers on oyster shell with Bryozoa, its

occurrence agrees with that later reported up the bay by

the present author. What is evidently the same form

has been seen by me in the Severn and other parts of the

Chesapeake Bay in 1912-15 and described without specific

identification.

In Dons’s classification these Chesapeake forms are

evidently Semifolliculina and might be included in the

widely variant S. boecki, but that the tube is so much

longer and the collar so relatively narrow.

Moreover, it has the gregarious habit described by

Wright, the free-swimmers being stuck together in a

secreted ‘‘ colletoderm,’’ and may be identified as the

same as his Freia (Lagotia) producta, if we grant that in

his sketch of the animal and bottle he overemphasized the

‘‘immensely prolonged ’’ tube in proportion to the sac,

which he figures as relatively too short for his compari-

son to a jack boot.

If, then, we may retain the name producta as signify-

No. 639] AMERICAN FOLLICULINAS. 361

ing a long spiral tube with narrow collar, we will separate

it from the other four species S. gigantea, S. planorbis,

S. similis and S. boecki recognized by Dons and may add

the characteristic feature of the animal that its nucleus

is generally moniliform and the lobes very long and ligu-

late so as to be compared to the ears of a hare in length by

Wright and from their curvature of surface to blades of

obstetrical forceps by Ryder—far removed from the fun-

nel appearance of many of the smaller and earlier forms

that had been described.

Tubes that have had second additions added to them,

described by Wright, were also sometimes observed both

by Ryder and by Andrews.

While the ordinary form of the Chesapeake is thus a

much longer tube and markedly spiral as compared to the

simple New England form, so that it may be provision-

ally regarded as specifically distinct and referred to

Semifolliculina producta, there are also other forms of

bottles occurring sparingly with the long tubes that are

more simple and short than the New England forms,

though they may have an added complexity regarded by

Dons as a sort of valve and relegating them to the genus

Parafolliculina.

Of the two species recognized by Dons the Chesapeake

form is evidently Parafolliculina amphora. Character-

ized by the wide flat sac attached along its lower face to

substratum and joined to short tube which swells out

around mouth of sac and then rapidly diminishes to end

with upward turn and narrow mouth with little or no

collar. The whole enveloped for the most part in a halo

of soft secretion and the junction of sac and tube char-

acterized in many specimens by an internal valve-like set

of membranes or modifications of the edges of the sac

where jutting into the swollen tube. The animal is sim-

ple with single nucleus and short arms or funnel and

nearly colorless.

The measurements given by Dons for specimens from

Norway and Iceland are: length over all 110 to 150, of

362 THE AMERICAN NATURALIST [Von. LV

which sac is 73, 82, 100, 112, when tube is 27, 27, 50, 37.

Width 90-120, narrowest width 40-70. Diameter of

swelling of tube 50-80, width of tube about 30-50. Diam-

eter of nucleus from 15-32.

The Chesapeake forms have a much flattened sac, some

130-140 long by 195-110 wide, but only 30 deep. The

exact point of passing to tube is various; the tube may

be regarded as 25-55 wide and swelling to 57 wide, though

it may be but 25 at actual mouth, which rarely has a col-

lar, but may have a flaring of 6 m. The length of tube

may be 30, making the length over all 175. The tube may,

turn upward from attachment nearly 50 » when sac is but

30 deep. The nucleus is from 15 by 17 to 15 by 27.

The halo of secretion about sac is some 7 » or more in

thickness. Characteristic of some specimens is the valve,

so-called, relied upon by Dons as of generic value. This

appears as a dorsal and ventral flap of membrane within

the sac at its continuation as the tube and projects for-

ward. These two membranes converge to meet below

the center of the cavity, which they close off more or less

completely. They vary in number and position and often

seem to be lacking.

A description of these amphora forms of the Chesa-

peake will be published elsewhere and we will here merely

add a summary of the foregoing consideration of the

probable position of the known American folliculinas in

the tentative scheme of classification of the family.

Following Dons we may separate the Folliculinas from

the Stentors on account of the more or less marked de-

velopment of the body as lateral lobes which make a

funnel leading toward the mouth and which may be re-

garded as a specialization of the more primitive feeding

apparatus found in Stentor.

The Folliculinide are thus Heterotrichia with spindle-

shaped body prolonged as lateral lobes to form a funnel

leading toward the mouth and marked ability to secrete

dwellings composed of a sac-like part more or less pro-

longed as a tube which may reach great length and ex-

hibit spiral structure.

No. 639] AMERICAN FOLLICULINAS 363

The sac is fastened by secretion to some foreign body

and the animal lives sedentary till such times as it may

break loose from attachment of foot end to base of inside

of sac and then swim free to soon secrete another sac and

tube. In free-swimming individuals, lobes and mouth

may be absent and later reconstructed. Contractile vacn-

ole absent.

The nucleus is round, oval or moniliform. The micro-

nuclei may be many and minute. Cross division results

in halves, of which one may remain and the other escape

from the dwelling; the size differs much in different indi-

viduals. Complete life cycle, when known, may show

- that some of the apparently specific forms are but stages

in life cycle of others. —

Species based so largely upon the forms of the secreted

dwellings may eventually prove to be but results of di-

verse secretional activities within one species.

Conjugation unknown; reproduction by transverse fis-

sion follows nuclear condensation and dedifferentiation

of peristome; the posterior half grows a new mouth and

peristome. and soon occupies the old dwelling, while the

anterior half swims free with no mouth and simple spiral

membranelle zone, secretes a new dwelling, and differenti-

ates new mouth and peristomal apparatus.

The old genus Folliculina may be conveniently divided

into tentative subdivisions as suggested by Dons; based

chiefly upon shape and proportion of the secreted cases.

These groups may be spoken of as genera, namely,

Folliculina, Semifoliculina, Parafolliculina and Pseudo-

felliculina.

In Folliculina Lamarck, as restricted by Dons, the sac

is commonly as wide as long, the tube is short and with-

out collar and there is no spiral nor valves.

The only species are the original F. ampulla of Mueller

and the F. paguri, which was Giard’s Pebrilla paguri,

and seems to be an abnormality. The fresh-water F.

boltoni seems to be the same as F. ampulla, and is re-

corded from England, Switzerland and from France (as °

the Ascobius of Henneguy).

In Semifolliculina the sac. is longer and the tube may

364 THE AMERICAN NATURALIST [Yon LV

be very long, with a collar and more or less spiral mark-

ing, but no valves.

Some six species may be recognized. The widespread

S. boecki, being separated by Dons from the old F. am-

pulla and bearing the specific name of Claparede’s Co-

thurnia, includes the Freia ampulla and Freia aculeata

of Claparede (which are only stages of growth transfor-

mations), as well as the Lagotia viridis, L. atropurpurea

and L. hyalina of Wright. Similar but much more

evolved in specialization of tube is the S. producta of

Wright.

S. gigantea of Dons is the same as Laachmann’s ant-

arctic F. ampulla and as Stein’s F. ampulla. :

The S. elegans of Claparede may belong here as tenta-

tive species with mouth of tube incised on one side, but

this feature may well be accidental form.

S. spirorbis of Dons is well marked by absurdly narrow

tube and is a very minute form that suggests depauper-

ization.

S. similis recently described by Dons from south polar

material possesses a very wide tube without spiral. The

specimen figured shows tentacular projections of lobe

such as we see in transformation stages.

Parafolliculina has a short tube which is swollen just

above a narrow connection with the sac and may present

internally membranes regarded as valves. There are

two known species. The typical species is P. amphora,

one of the smallest of the bottle animaleules found by —

Dons to remain the year through in some localities in

Norway, while also known from Iceland and, we find, in

the Chesapeake Bay. The other species, P. violacea, has

the case attached only at the base of the sac and not along

its entire ventral face as in all above-named species. If

this feature is incident to some unusual behavior of the

free-swimming stage when about to settle down and con-

struct the sac, the two species may prove to be but one.

It was found on the French coast by Giard and is known

from south Norway, the Adriatic and West Australia.

Pseudofolliculina has no enlargement of the tube, but,

No. 639] AMERICAN FOLLICULINAS 365

in fact, the tube and sac grade into one another and stand

straight up from the attachment. This is by means of a

long cylinder of cementing material, bringing the base of

the animal well above the substratum. There may be a

simple membraneous valve.

Of the two species, P. mellita was taken in the Antare-

tic in 1902-3 by Laachmann at depths of 350-385 meters

and its case has a length .6—.7 mm., but this great length

is partly due to the mode of attachment of the sac by

means of a stalk of secretion. As this stalk is hollow and

filled by a tenuous prolongation of the body of the animal,

it may be that this genus also is founded upon individual

idiosynerasies of secretional activity.

The other species, P. arctica, has been formed by Dons

to include the smaller but similar forms he found in Nor-

way and finally separated from P. mellita as smaller,

with narrower stalk apparently not perforated.

As the above eleven species have been most studied in

England, Germany and Norway, it is natural that Folli-

culinas are known chiefly from those coasts; yet the

known distribution of Folliculinas has already been ex-

tended to the Antarctic, the Mediterranean, the White

Sea, as well as to the east and west coasts of the North

Atlantic. On the east coast of the United States we seem

to have four or five species in two of Dons’s subdivisions

of the old genus Folliculina, namely : the most specialized

and best known Semifolliculina producta of the shores of

the Chesapeake; the accompanying small and simpler

form Parafolliculina amphora; the less well-marked form

Semifolliculina boecki, first found at Newport, R. I., and

later at Woods Hole, Mass.; and finally, as recorded by

Dons, from material from the North American Atlantic

coast the Semifolliculina spirorbis; and if Dons’s Para-

folliculina violacea be a real species, it also is to be cred-

ited to Woods Hole.

That some or all of the smaller and simpler forms arise

from larger and more complex forms under changing

conditions of nutrition in the successive phases of seden-

tary and free life is a tempting working hypothesis.

366 THE AMERICAN NATURALIST [Vor. LV

Whether these widely distributed marine protozoa

which in some cases are able to live in fresh water may

not be found in all parts of the world remains to be

found out; with attention turned to their discovery, it

may be hoped that knowledge of both the life histories

and the taxonomy may soon be placed upon a firmer basis.

Andrews, E. A.

1914. The bottle animaleule, Folliculina. Biological Bulletin, XXVI.

1915. Distribution of Folticulina in 1914. Biologica pais XXIX.

1920, Alternation of Phases in Folliculina. Biological Bulletin,

Barrett, C. A.

1870. On a new tube- Pipo Stentor. Monthly Mic. Journal, III,

p. 188-91, Pl. ;

Claparède and Lachm

1858. Etudes sur ie Infusories et les Rhizopods, I. Genéve, p. 204.

Daday, E.

1886. Ein ‘kleiner Beitrag zur Kenntniss der Infusorien fauna des

von Neapl. Mittheil d. Zool. Sta. (ee ite ene.

SEEE A

1841. Hist. Nat. des zoophytes infusories 1841—atlas with 22 pl.

Dons, Carl

1912. Folliculina-studien, I-III. Archiv fiir Protistenkunde, Bd. 27.

ee 1913. Follie aiiad-Sendion, IV. Vorlaufige Bemerkungen über

jo ystematik der Folliculiniden, nebst beschreibung neuer

eh nee e Arten. Thromso Museums Aarshefter 35-36,

s. and 2 pl., 90 pps.

1915. Kaiafeligrene Naturen’s maihefte, pp. 143-152, 10 figs.

Entz, G.

1884. Ueber Infusorien des Golfes von Neapel. Mitth. Zool. Sta.

Bu 1884, p. 294.

Fabre-Domer

1885. Tnfusories ciliés de la baie de concarneau. Jour. de l'anatomie

t de la physiologie, An, 21, p. 567.

Giard, A

1883. Sur les Infusories du genre Freya. Bull. scient. de Dép. du

Nord., i i . 26 .

1888. Bull. Scientifique de la France et de le Belgique, III Ser.,

An., pp. 310-17, Pl. 20.

Gruber, A.

1884. Die Protozoen des Hafens von Genoa. Verhand. d. Lep. Caro.

Henneguy, L. F.

1884. Note sur un nouvel infusorie cilié (Asċcbius lentus). Archiv

d. zool. exper., (2), I, II, pp. 412-10, Pl. 21.

Laachmann, H.

1901-1903. Zur. Kenutals der heterotrichen Infusorien der Gattung

No. 639] AMERICAN FOLLICULINAS 367

Folliculina Lam. Deutsche Siidpolar-Expedition, Bd. XII,

Zool. IV

Lamarck

1816. Histoire nat. des animaux sans Vertébres. II. Paris, pp. 29-30.

a È

1859. Proc. Acad. Nat. Sci. Philadelphia, p. 194.

Levander, K. M.

1894. Materialien zur Kenntniss der Wasserfauna Helsingfors:

Proto cta z pro Fauna et Flora Fennica, XII, pp.

82-84, P]. III, fig.

Mereschkowski, C.

1 Studien über Protozoen des Nördlichen Russlands. Archiv. f.

mic Anat. Bd. Folliculina ampula. O. F. Müller

n Wiessen Meer.) pp. 153-244, PI. X-XI

Meyer and Mbius,

1865. Fauna der Kieler Bucht. Abhandlungen Berlin Akad., I, 8.

XLV

Mobius, K

1887. us Flaschenthierchen: Folliculina ampulla. Abh. Nat. Ver.

Hamburg, 10 Bd., 14 pp., I Taf.

Miiller, O. F.

1786. Animalcula infusoria fluviatilia et marina. Havoniae, p.

83 ; PD. Er.

Penard, E.

1919. On Folliculina boltoni (S. Kent). J. Royal Mie. Soe., Dec.,

pp. 305-319, 2 Pl.

Rees, J. van

1883. Protozoen der Oos terschelde. Tijdschrift d. neder. Dier-

kundige Vereeniging, p.

Ryder, J. A.

1889. On the occurrence of Freia producta in the Chesapeake Bay.

AMER. Nar., 14, p. 133. ;

Sahrlage, H.

`1916. Uber die Organization und den Teilungs-vorgang des Flas-

chentierchens (Fol iculina ampulla). Arch. f. Protistenkunde,

Bd. 37, Heft 2, pp. 139-171, 2 pl.

Schröeder, O.

1907. Die Infusorien der Deutschen Südpolar- PED CR

Ergebn. d. Deutschen Siidpolar-Exped. us. von E. v.

Drygasliski, Bd. IX, pp. 351-360, Tf. Toe.

Wright, S. i

1857. Edinb. New. Phil. Journal.

1858. Description of New Protozoa. T New Phil. Journal, N.

: , 7. pp. 276-278, Pl. VI, Figs.

1859. Deseription of New ee ab New Phil. foaral X,

N. S., pp. 97-100, Pl. VII, 1

1859. Lagotia producta. Edinb. ‘a Phil. Journal, X, pp. 97-101,

Pl. VII, figs. 1-6.

1861. Pritchard, Hist. of Infusoria. 4th Ed., p. 605, Pl. XXVI.

1862. Observations on British Protozoa, Q. J. Mie. Sci., II, pp. 219-

20, Pl. IX, Fig. 2, 3.

SHORTER ARTICLES AND DISCUSSION

THE DUALITY OF EGG-SECRETION

THERE are two’ properties of egg-secretion which have attracted

particular attention. One of these is the power of agglutinating

spermatozoa (Lillie’); the other is the power of activating the

egg (auto-parthenogenesis, Glaser’). These effects have been

studied separately, but under normal conditions both must bear,

in close affiliation, upon the events of fertilization.

We are still in the stage of hypothesis. We know that without

exudates fertilization does not take place. This was first demon-

strated by F. R. Lillie (loc. cit.), who removed the exudates.

These, as I found later, can be rendered ineffective without

removal, by the presence of charcoal.’ In neither case is fertiliza-

tion possible. We also know, from the investigations of Lillie,

that the agglutinating agent can be neutralized in two ways:

either by spermatozoa, or, by some derivative from the egg itself.

Inasmuch as fertilization is prevented by this derivative, Lillie

has called it ‘‘anti-fertilizin.”’

Lillie also discovered Da the blood of the sea-urchin (Arbacia)

-has preventative effects, but these are different from those of

‘‘anti-fertilizin.’’ There is in this case no interference with the

agglutination tin yet the egg remains unactivated.

iss Woodward‘ and I (loc. cit.) have extended the list of

inhibitors to include such diverse things as the pigmented sub-

stances derived from testicular tissues; fatty and aqueous ex-

tracts of the eggs; and now, oleic acid and olive oil. However

I shall consider further only the inhibitors discovered by Lillie.

These, together with the fact that the exudates are necessary

in normal fertilization, are the foundations of the fertilizin

1 Science, Vol. 38, pp. 524-528.

2 Biological Bulletin, Vol. 26, pp. 387-409.

3 The use of charcoal was suggested to me by Dr. G. H. A. Clowes.

4 Journal of Experimental Zoology, Vol, 26, pp. 459-501,

368

No. 639] SHORTER ARTICLES AND DISCUSSION 269

hypothesis. According to this view, egg-secretion contains a |

body with two bonds; one normally unites with a sperm-borne

valence; the other, with a valence borne by the egg. Agglutina-

tion is a symptom of the first unjon; activation of the egg, a

symptom of the second. ‘‘Anti-fertilizin’’ occupies the aggluti-

nating valence and hence the normal union of sperm and fer-

tilizin is rendered impossible. The inhibitor in the blood, on the

other hand, is effective because it binds the second valence of the

fertilizin which in consequence cannot unite with its normal

receptor in the egg. In the language of Ehrlich, fertilizin is an

amboceptor with ovophile and spermophile side-chains; and,

normal fertilization involves the formation of a kbeaninal com-

pound which, written in linear fashion, can be thought of as

sperm-receptor—spermophile side-chain-ovophile side-chain—ege-

receptor.

The keystone of the hypothesis is the amboceptor. It symbo-

lyzes the effects of the secretion on spermatozoa and the effects

on eggs; it symbolizes also the information gotten from the two

inhibitors—the one that prevents agglutination and fertilization

and the other which prevents merely fertilization. The ambo-

ceptor stands vaguely for the recognized duality of the secretion.

The question is, can we precipitate the amboceptor from the realm

of. the symbolic and bring it within the sphere of the commoner

conceptions of physics and chemistry ?

III ;

First of all, are we compelled to think in terms of the ambo-

ceptor? The alternative, of course, is a two-body view. Lillie’s

book,® (p. 231) decides against this because there is ‘‘a parallel

between absence or loss of agglutinating substance and the

capacity of the egg for being activated. The same results would

be attained if there were two substances concerne :

since the two effects ... appear and disappear together...

the writer assumed that they may be regarded as due to a single

complex substance.’’ Again (p. 232), in discussing the effects

of the inhibitors, Lillie writes:

This still does not prove that sperm agglutination and egg activation

are due to the action of a single substance, but it shows again by a

different method that the capacity for producing both effects is present

5 ‘‘ Problems of Fertilization,’’ University of Chicago Press, 1919,

370 THE AMERICAN NATURALIST [Von LV

simultaneously in the egg secretion and the assumption of a single

substance is the simplest hypothesis.

A much stronger argument for the one-body view can be ad-

vanced. ‘‘ Anti-fertilizin,’’ by hypothesis, prevents normal fer-

tilization by occupying the spermophile side-chain; yet, ‘‘anti-

ertilizin’’ also prevents auto-parthenogenesis (Woodward).

This, we might suppose, would indicate a relationship between

the two side-chains, such that removal from the reaction system

of the one also results in the removal of the other. This state

of affairs is actually realized under certain conditions. For ex-

ample, I have been able to isolate from charcoal materials with

agglutinating and activating properties. However, if we infer

from these observations, a single body, the reaction of ‘‘anti-

fertilizin’’ with fertilizin can hardly serve as evidence that the

latter has two side-chains.

In 1918, Miss Woodward‘ reported experiments that bear in-

timately on the problem. By saturating Arbacia secretion with

(NH,).SO, she secured a white flocculent precipitate which after

purification by dialysis proved to have very intense agglutinative

powers but no capacity for initiating the development of the egg.

She called this precipitate, agglutinin.

An initiatory agent, also, was precipitated. BaCl, was added

to the secretion and after removal of the sea-salts, the primary

deposit was treated with N/10 HCl. As soon as the acid had

been freed from BaCl,, acetone was used in excess to bring down

a second precipitate, heavy and flocculent. This, after purifica-

tion with absolute alcohol and ether, dried as a white powder,

soluble in both sea-water and distilled.

This second precipitate had marked parthenogenetic effects,

but no power to agglutinate sperm. Miss Woodward called the

substance, lipolysin, a name which, as I shall show elsewhere, is

justified since the material accelerates the hydrolysis of fats.

What bearing have these precipitations on the issue? Accord-

ing to Lillie’s book, p. 240, ‘‘separation under the conditions of

chemical analysis may possibly denote a splitting of a single sub-

stance ot the normal egg.’

No. 639] SHORTER ARTICLES AND DISCUSSION Sil

A. THe AMMONIUM-BaRIUM PRECIPITATIONS

1. The agglutinin can be salted out with (NH,).SO,; the

lipolysin can be brought down with BaCl,. If we divide a given

exudate into two portions, we can precipitate in one, first the

agglutinin and later the lipolysin, whereas with the other frac-

tion we can proceed in a manner exactly the reverse.

2. Precipitation in the two cases differs inasmuch as the effect-

ive concentration of the BaCl, is N/7.5; that of the (NH,).S0O,

in the neighborhood of 5N.

If precipitation in the two instances is uncomplicated by chem-

ical unions between precipitate and the reagents used to produce

them, the case is decisive for the two-body view since it is un-

likely that a molecule can be split chemically, merely because one

part has solubilities different from those of another part. How-

ever, the case is probably not so simple. Very possibly we are

dealing with ammonium agglutinate and barium lipolysinate.

If this is true, then all we can say is that the amboceptor, present

by hypothesis at the outset, breaks down between the agglutinat-

ing and the ee valence no matter which of the two groups

is bound firs

B. Tue CHarcoaL METHOD

Charcoal removes practically the entire organic reaction sys-

tem. This can be recovered in its essential parts, by subsequently

treating the charcoal with N/10 HCI. From the clear solution so

gotten acetone throws down a voluminous precipitate in two well-

marked stages: the first fraction, without agglutinating powers,

is lipolytic; the second strongly agglutinates the sperm.*

Very likely factors are involved in precipitation by charcoal

which are not present when (NH,).SO, and BaCl, are used.

Very possibly the amboceptor is split by the charcoal; or, it may

not be split until the acetone in the HCl reaches a certain concen-

tration. In any case the cleavage of the molecule gives results

identical with those gotten by Miss Woodward’s methods.

C. Reactions to HEAT

Lillie has shown that the agglutinating material is extremely

resistant to heat. Exudate which has been boiled agglutinates

spermatozoa perfectly well. It has, however, lost its capacity

as a parthenogenetic agent (Woodward).

6 It is very important to guard against impurities in the charcoal. These

after extraction with HCl give a voluminous precipitate with acetone.

312 THE AMERICAN NATURALIST [Vou. LV

On the one-body view we must reckon with the following pos-

sibilities: (1) a rupture of the molecule; (2) the destruction or

alteration of the ovophile side-chain.

Looked at conservatively, the methods of precipitation alone

do not enable us to decide the one-body—two-body issue. The

same thing may be said of the results of boiling. One conelu-

sion, however, is certainly warranted: there is a constitutional

weakness in the amboceptor so pronounced that this molecule

breaks down with the greatest ease and, under very diverse con-

ditions, always cleaves in a manner that separates the spermophile

from the ovophile side-chain. Indeed one doubts whether the

amboceptor can hold its ovophile side-chain after the agglutinat-

ing group has united with the receptors of the sperm. This, if

true, would be awkward for the theory.

VII

D. FILTERABILITY

Lillie has shown that exudates which have passed through

Berkefeld filters no longer agglutinate spermatozoa. The agglu-

tinating material, in this case, can be recovered, as Miss Sampson

found last summer, by washing the filter-cone in sea-water. In

the original filtrate I was able to demonstrate lipolysin.

It is conceivable that the agglutinating material is not filter-

able because of ‘‘chemical adsorption.’’ If this is true, then

filtration becomes merely another method for the chemical de-

composition of the amboceptor. Yet, we can account in this

manner for only a portion of the agglutinin held back. It seems

very unlikely that the fraction which can be recovered by merely

washing the filter-cone, was held chemically bound. Moreover,

on account of the metal band which holds the filter-cone in posi-

tion, a remnant of the secretion invariably fails to pass through.

This remnant has a higher agglutinating value than the original

exudate. It seems safe to conclude that the agglutinating ma-

terial is held back mechanically. If correct, these considerations

based on filtration are conclusive, for we know of no cases in

which a substance is chemically decomposed merely because the

whole molecule is unable to get through the pores of a filter.

The results of filtration seem to me to necessitate the two-

No. 639] SHORTER ARTICLES AND DISCUSSION 373

body view. None of the facts explicable by the amboceptor

appear to become inexplicable when this body is analyzed into

an agglutinin and a lipolysin. As for the difficulties of constant

association, these are no greater in this case than the difficulties

which arise because serum albumin and serum globulin always

occur together in the blood.

OTTO GLASER.

AMHERST COLLEGE,

February 7, 1921

DESCRIPTION OF A PECULIAR YOLK MASS IN THE

OVIDUCT OF A HEN

A DESCRIPTION of this specimen seems desirable for two chief

reasons: First, because of its unique nature; second, because it

supplies the data with which to answer the question whether

reverse movement, possibly Eee occurs in the forma-

tion of double eggs and similar anoma

The specimen was presented to the locii] laboratory by

Ashton Barbour, of Charlottesville, Va., six hours after it had

been removed from an apparently normai year-old hen. He des-

cribed it as having been taken from the ‘‘egg-bag.’’ When

questioned, he was positive that he had noticed a number of

developing eggs, ‘‘little yellow balls,’’ attached to the dorsal

surface of the abdominal cavity. His anatomical observations

stopped at this point.

The specimen was roughly egg-shaped, and of a yolk or yellow-

ish-orange color. Between ends it measured 914 em. Its

diameter at the point of greatest width (about one third the

distance from the wider end) was 8 em. When opened the mass

was found to contain an egg of average size, with a shell of

normal hardness and thickness. The egg was not exactly in the

center, but was placed slightly to one side and towards the larger

end, causing a variation in the thickness of the lateral walls of

the enveloping mass (Fig. 1). At the thickest point the lateral

wall was 2 em. thick, at the thinnest point 1 em. At the larger

end the wall measured 34 em. in thickness, at the smaller end

it measured 114 em. in thickness. The weight of the enveloping

portion of this yolk mass, after the enclosed egg had been re-

moved, was about 190 gms. The general ovoid shape of the mass

was presumably determined by the enclosed egg.

The mass was made up of layers of yellow, yolk-like material

Sit THE AMERICAN NATURALIST [Von. LV

between which were scattered irregular lamine of a glairy,

mucus-like substance (Fig. 2). In places these lamine had ap-

parently hardened to form clear, firm, gelatinous areas. The

outermost yellow layer was about 1144 mm. thick, and completely

encircled the mass.

On one side, at the point of greatest diameter, there was a

1G. 1. Frc. 2.

Fic. 1. Diagram of yolk mass from acini hen. The inner broken line

represents a outline of the inclosed normal egg. The peripheral aici

area represents the laminated envelope of site and albumen. Natur athe

‘1G. 2. Diagram of transverse section of the yolk mass. The central capita

incl rmal e æ repre-

sent layers of yolk, the clear lamellæ, layeri of albumen. Natura

shallow depression, about 2 mm. deep, almost the size of a dime

in circumference. This was due to a thinning of the two external

enveloping layers at that point. A dark, reddish discoloration

partly surrounded this depression in the form of a crescent.

There were a number of small granule-like hillocks, about the

size of a pinhead, on the surface of the smaller end. These

elevations probably represent casts of the mouths of the oviducal

glands, produced under pressure of the enlarging mass against

the constricted confines of the oviducal walls.

When the egg which the mass inclosed was removed and

opened, it was found to be filled with a yellowish liquid, in which

there were bits of a translucent and whitish mucus-like substance,

the remains most probably of the disintegrated chalaze. The

odor of this liquid was not offensive. It may be best described

as musty.

Portions of the yolk mass were imbedded in celloidin, sec-

No. 639] SHORTER ARTICLES AND DISCUSSION 375

_ tioned, and the sections stained with hematoxylin and eosin.

Transverse sections through this laminated cortical material re-

vealed layers of yolk granules and spherules intermingled with

layers of clear, hardened, egg-white. There were no indications

of the presence of shell, or any unequivocal evidence of shell

membrane, in any of the sections.

Abnormal eggs have been observed and discussed by biologists

for many years. In the AMERICAN Narturauist for January,

1906, G. H. Parker (4) has treated the subject of double hens’-

eggs, ‘‘ovum in ovo,” very fully. He reviews much of the pre-

vious literature on the subject and describes several specimens

of his own, similar to a specimen of a large, double egg which

belongs to the laboratory of histology here. Parker supports

the theory of Davaine (2) and others concerning the formation

of double eggs. _

Briefly the theory is this: The egg is moved by peristalsis

from the ovary to the distal end of the oviduct. As it passes

down the oviduct it receives the usual coverings of albumen, shell

membrane, and shell. The egg is now a normal egg, ready to be

laid. But for some reason, instead of the egg being laid nor-

mally, antiperistalsis occurs and the egg is carried back up the

oviduct. In the upper portion of the duct it meets another de-

veloping egg coming down. The two pass down together. Albu-

` men is laid on and a common shell covers the whole mass. We

now have a giant egg, approximating the size of an ostrich egg,

which contains a second complete normal egg along with its own

yolk and albumen.

Curtis (1) has described a number of interesting anomalies

in hens’ eggs, including double eggs and other anomalous speci-

mens, either with a membrane only or with both shell and mem-

brane. She reports finding eggs in the body cavity of. fowls

whose oviduct had been ligated in the isthmus, or shell gland.

She does not venture to commit herself, however, as to whether

antiperistalsis is the means by which the egg is carried back up the

duct. Patterson (5) describes a specimen which has two shell

membranes. He explains this condition on the assumption that

antiperistalsis had occurred twice before the egg was laid.

Hargitt (3) describes an interesting gourd-shaped egg. None of

these authors, however, mention an anomaly similar to our

specimen.

What may be assumed to have happened in the formation of

376 THE AMERICAN NATURALIST [Von. LV

our specimen was this: The first egg which left the ovary of

the young hen passed down the oviduct normally and had albu-

men and shell laid on in the usual manner. The egg passed on

into the lower part of the uterus (shell gland), but, due to in-

jury, congenital occlusion of the vagina, or some obstruction, the

egg could not be laid. Such interference with normal oviposition,

either congenital or acquired, stimulated a reyersed movement.

(probably antiperistalsis) and the normal egg was carried back

up into the oviduct and lodged there. More eggs left the ovary,

took on albumen as they passed down the duct, but coming into

contact with the preceding egg which occluded the duct, where

broken by pressure, and the soft yolk and albumen collected about

the obstructing egg. In this way the yolk mass about the egg

acquired its large dimensions.

We can thus locate very closely the exact position of the

anomalous yolk mass in the oviduct of the hen. Since it con-

tained layers of albumen it must have lodged below, or in the

lower part of the portion of the oviduct where albumen is laid

onto the yolk; and since there was no shell whatever within the

cortex of the mass, it must have lodged above the point in the

oviduct where. shell is formed. Again, antiperistalsis, or at least

reversal of normal movement, must have occurred because the in-

closed egg comprised a shell, and so. must itself have gone the

full length of the oviduct into the uterus, while at the time the

enveloping yolk mass was formed the original egg must have

been above the shell-forming level.

The above evidence, combined with the evidence of Davaine

(cited by Parker) and Curtis, who report finding soft-shelled

eggs in the body cavity of fowls, seems to prove conclusively that

something of the nature of antiperistalsis in the oviduct does

occur. The inference seems warranted with regard to our speci-

men, that if the included normal egg could have retraced its

course down the oviduct in company with the next following egg,

instead of lodging permanently in the preuterine portion of

the oviduct, a common shell would have been laid onto the two

eggs. This shell would have included the two together, and the

result would have been an ordinary ‘‘ ovum in ovo,’’ similar to

the ones described by Davaine, Parker, Patterson, Hargitt, Curtis

and many others.

No. 639] SHORTER ARTICLES AND DISCUSSION 377

LITERATURE CITED

1. Curtis,

ney "States on the Physiology of Reproduction in the Domestic

Fow XVI. Double Eggs. Biological Bulletin, Vol.

i pp. 181-213

2. Davaine, C.

1861. Mémoire sur les anomolies de l’oeuf. Compt. Rend et Mém.

Soc. Biol., Paris, Sér. 3, T. 2, pp. 183-266. (Cited from

Parker)

3. Hargitt, C. W.

1912. Double Eggs. AMERICAN NATURALIST, Vol, XLVI, pp. 556-

560

4. TE G

1906. Trouble Hens’ Eggs. AMERICAN NATURALIST, Vol. XL, pp.

3-25

5. Patterson, J. T.

1911, A Double Hen’s Egg. Awznrcan NATURALIST, Vol, XLV,

pp. 54-59,

ROBERT BATTAILE HIpeN

THE DEPARTMENT OF HISTOLOGY

AND EMBRYOLOGY

UNIVERSITY OF Yran

THE HEREDITY OF ORANGE EYE COLOR IN

DROSOPHILA MELANOGASTER!

THERE are three points of special interest in the heredity of

orange eye color. First, the eye color is due to the presence of

two sex-linked genes; second, these two genes may separate in

the F, female when orange is crossed to the wild stock, producing

in F,, in addition to orange and wild type males, a third eye

color called salmon? ; third, when an orange male is crossed to the

parent stock, reduced, only orange and wild type males appear

in By

Orange first appeared in the sixth generation of the plus

selected line of the mutant strain reduced. Eleven males ap-

peared from a single pair of parents. Several of these males

(orange reduced) were mated to wild type females. All F,

flies had red eyes. Twenty-seven F, pairs were mated (Table

I). Of the F, males 850 were wild type (red-eyed), 785 were

orange reduced, 585 were reduced, 586 salmon, 15 orange, and 6

salmon reduced. This behavior led us to suspect that orange

1 Contribution No. 182.

2 Since this paper went to press Professor Morgan kindly sent us some

stock òf garnet. The crosses show salmon and garnet to be the same.

378 THE AMERICAN NATURALIST (Von. LV

TABLE I

F, 9 Wip Tyre X F, ¢ WILD TYPE (FROM ORANGE REDUCED & X WILD

TYPE 9

Female Male

Bottle

Number Wild Wild Orange

Type Type Reduced | Reduced Salmon Orange | Salmon

789 194 65 42 61 46

798 127 31 22 33 34

804 15 19 30 11

808 109 26 24 29 30 2

814 72 17 9 14 13

815 140 33 23 30 16 2

816 123 22 16 25 14

831 65 17 8 12 T

854 77 30 11 14 9

859 69 18 11 13 15 1

860 66 18 13 22 13

1,431 30 13 25 Fi 1

1,372 174 48 39 31 31

1,494 28 12 7 8 5

1,495 173 45 34 37 30 1

1,557 149 31 25 42 22 1

1,569 112 87 25 29 19 2 2

1,570 106 20 20 17 17 1

1,571 66 22 9 10 15

1,674 111 26 27 33

Tor. 175 62 41 55 42 1 1

t730.. 207 64 27 72 26

1,724.. 174 55 21 36 24 1

1,720... 58 35 30 40 1

1,441.. 111 39 27 32 21 1

2,431. 22 22 19 1

2, Aare 1 20 22 19 23 2

Totals. ..| 3,133 850 585 785 586 15 6

TABLE II

F, X F, [From (SALMON X 9 WILD TYPE (FROM STOCK) |

Female Male -

Bottle Number

Wild Type Wild Type Salmon

ElSe o 146 65 54

,665.. 123 51 34

1,668... 8: 175 68

1619. oa, 121 69 54

4,082) ics 130 68 54

5,680.5... 161 56

Im Ss, 117 54 62

Totals 963 439 382

No. 639] SHORTER ARTICLES AND DISCUSSION 379

eye color was due to the presence of two sex-linked genes, one

of which by itself produced no visible effect, but when the two

were brought together orange was the result. With this hypoth-

esis in mind we attempted to prove or disprove it and believe

we have demonstrated that the suggested explanation is the cor-

rect one.

First, some of the F, salmon males were mated to wild and

the eye color shown to be due to a single sex-linked gene. An

examination of the F, males when an orange reduced male is

mated to a wild female (Table I), shows that practically all

salmon males (586 out of 592) are wild type with respect to

bristle number, and that practically all orange males (585 out of

600) are reduced. These results showed that the modifying

gene which changes salmon into orange is closely linked to re-

duced. Most of the F, reduced red-eyed males then should carry

this modifier and we should be able to produce orange flies in the

second generation by mating these reduced F, males to salmon

females. Such crosses have been made.and the results show that

the F, reduced males do carry such a gene which when added to

salmon produces orange. The two genes for salmon and salmon

modifier are brought together in the same chromosome by cross-

ing over in the F, female (Table III).

TABLE IIT

(a) REDUCED ¢ (F, FROM ORANGE REDUCED ¢ X WILD 2) X SALMON 9?

WH POF oarenien Are Aek Salmon ¢

O06 ery es Sch ee es Ok ee ee ea te 823

(b) F,9 [FROM (a) ] X ORANGE ĝ FROM STOCK

Bottle Female Male

Number i ‘ praca eae

Wild Wild

Type Salmon Orange on Salmon Orange

Ts a 233 06 134 248 117 116

Bt 5 ace 102 49 62 85 47

SOS. i 157 52 94 184 74 61

Pcs 23 38 63 42

266.433, 34 54 23 24

Totals 606 210 362 634 300 ns

That the modifying gene lies in the X chromosome rather

than in one of the autosomes is clearly shown by the F, ratios

when orange is mated to a wild female. If the gene were reces-

380 THE AMERICAN NATURALIST [Vou. LV

TABLE IV

F, 9 WILD TYPE X F, ¢ REDUCED (FROM SALMON ¢ X REDUCED È)

Bottle Female Male

Number | s ` j

Ta | Reduced Ty oe Reduced | Salmon en Ms pain Orange

8,347....| 183 | 112 60 76 | % 43 1

8,348....| 124 | 119 | 46 57 67 | 40 3

8,354... 88 | 105 | 47 43 ee RS ae ae

Toue i Ms 1 A eel ie | sie) et eee

sive and in an autosome all the F, females should be wild type.

Such is the case. Of the F, males, four should be red-eyed,

three salmon, and one orange. If the gene were dominant, the

ratios among the males should be four red, one salmon, and three

orange. The actual results are given in Table I. We find neither

of the former ratios, but 1435 red-eyed, 800 orange, and 592

salmon. The fact that there is a linkage between the genes for

salmon modifier and reduced is further evidence that the gene

for salmon modifier is in the X chromosome.

Orange eye color, then, is due to two sex-linked genes, one of ©

which when alone.produces salmon; the other called salmon

modifier, when alone produces no visible effect, but when added

to salmon produces orange. The discovery of such a modifying

gene is not new as Bridges has demonstrated seven of them which

modify eosin eye color. It does lend support, however, to the

presence and behavior of such genes.

These same orange reduced males when mated to reduced

females give a very different result from the cross to the wild.

When crossed to reduced, orange behaves as a single sex-linked

character (Table II). Why this difference? At first it was our

impression that the reduced strain carried a non-crossover factor.

This was disproved, however, by mating reduced to other sex-

linked characters. The other alternative and the correct inter-

pretation we think, is that the reduced strain is homozygous for

the gene for salmon modifier. This is easily demonstrated by

mating salmon to the reduced line. If the reduced line carries

the gene for salmon modifier, then a cross of this line to salmon

should give some orange males in F,. Such is the case (Table

IV). The reduced strain, then, carried the gene for salmon

modifier before the gene for salmon appeared.

No attempt has been made to locate accurately the genes for

No. 639] SHORTER ARTICLES AND DISCUSSION 381

reduced, salmon, and salmon modifier. Sufficient crosses have

been made, however, to approximate their positions, which are as

follows: reduced, 5.24; salmon modifier, 5.94; salmon, 41.33.

DESCRIPTION OF TABLES

Table I gives the results of mating the F, flies produced by

crossing an orange reduced male to a wild type female. All the

F, females are wild type. Of the F, males 850 are wild type,

585 reduced, 785 orange reduced, 586 salmon, 15 orange, and 6

salmon reduced. i

Table II gives the F, results of mating a salmon male to a

wild type female. Of the males 439 were wild type and 382

salmon.

Table III (a) and (b), gives (a) the results of crossing an F,

reduced male (from an orange reduced male X a wild female)

to a salmon female. The females are wild type and the males

salmon. In (b) the F, wild type females were mated to orange

males. The orange males were used instead of the salmon to

bring out orange in the females. Orange males and females

appear. Hence the F, reduced males must have carried the gene

‘for salmon modifier.

Table IV gives the F, results of crossing a salmon male to a

reduced female. Orange males appear. Hence the reduced line

must be homozygous for salmon modifier.

FERNANDUS PAYNE,

MARTHA Denny.

ZOOLOGICAL LABORATORY,

INDIANA UNIVERSITY

AN APPARATUS FOR MICRODISSECTION

To secure the greatest accuracy in the control of needles for

cell dissection an apparatus has been devised on the following

plan. The needle is attached to a right angled triangular plate

(Fig. A), each corner of which is moved by a milled headed

screw. The working of any screw causes the movement of the

plate about a line through the points of the other two screws

as an axis, thus producing motion of the needle point in the three

planes of space by manipulation of the screw heads. The nature

of the bearings of the screw points on the plate eliminates all

side play. The conical end of one screw works into a circular

382 ` THE AMERICAN NATURALIST [Vou. LV

` > £

TE, P

a UE | (BRS

: : J :

Fie, A. Detail of movable plate which bears the hirii. illustrating the

meckanical 3 recat involved. See tex

Fic. B. Diagram of one type of the apparatus clamped to the asept

pe foc

1 of t

needle point are secured by turning one screw each (a and c RA ; the

second horizontal movement, by turning two adjacent screws together (a and b

with two fingers of one hand). Coarse sive stint are made by moving the

needle at the ball-and-socket joint d.

hole through a corner of the plate (b,-Fig. A) ; the conical point

of the second screw works into a slit in its corresponding corner

(c), the third screw point (at a) bears on the plane surface of

the plate; these constitute what may be termed a cone-slot-point

‘support. The cone-cirele bearing prevents side slipping of the

plate in any direction; the cone-slot bearing prevents revolution

of the plate about the cone-circle bearing. A single spring (e)

No. 639] SHORTER ARTICLES AND DISCUSSION 383

against the center of the plate holds its bearings firmly against

the three screw points, and takes up all lost motion perpen-

dicular to the plate, in the screw threads. |

This mechanism has been employed in two styles of dissection

apparatus. In one type (Fig. B) the aim has been to secure a

short needle length, and close proximity of the needle’s support

to the axis of the microscope lens, with consequent freedom from

vibration of the needle and a long leverage in the control of its

point. In the other type (Fig. C) the apparatus is set away

from the microscope axis, to allow the greatest possible freedom

of manipulation of objects on the microscope stage. In the

former type, the plates bearing the needles cross at right angles,

under the microscope nosepiece; in the latter type, they are set

beyond the.edge of the microscope stage, nearly parallel. Each

instrument bears two pipettes or needles operated by three screws

apiece. The needles may be attached to the plates either by wax,

which when warmed makes possible a coarse adjustment of the

needles, or by a universal ball and socket joint, clamped loosely

enough to allow gross manipulation under the low power.

C. Diagram of second type of the apparatus, with parts arranged to

leave the microscope stage free for more convenient manipulation of objects upon

it. The ball-and-socket joint for coarse adjustment (at d) is here made to serve

as a pivot in place of the corresponding screw point (b); this screw then

works through this joint to produce one motion of the needle, thus avoiding the

ecessity of turning two screws together to secure this movement, as in Fig. B.

384 THE AMERICAN NATURALIST [Von. LV

The advantage of this type of needle control over the sliding

motions of the Barber pipette holder lies in greater ease and

convenience of manipulation as well as in greater refinement of

needle control. The simplicity of the apparatus is such that its

manipulation requires no great degree of technica] skill, and in

its construction have been avoided as far as possible all elaborate

and complicated adjustments. The screw heads turn easily

enough to be rolled under pressure of one finger each. The com-

plete apparatus for holding two needles may be made to occupy

a three-inch cube, and clamped close under the nosepiece; five

of its six screws can be operated with one hand, leaving to the

other, the mechanical stage ratchets and the sixth needle screw

adjacent to these. The needles may be held so close to their

points that no vibration is noticeable, especially since no moving

part is handled during manipulation. The needle points move

in ares of circles, but since the ratio of length of are to radius

is small, these are virtually straight lines, in the plane of, and

perpendicular to the plane of the focus of the objective.

A more detailed account of the working of this instrument, and

a description of various devices that may be added to facilitate

needle manipulation, will be included in a subsequent paper deal-

ing with the results of its application to the study of proto-

plasmic activity.

Gero. H. BISHOP,

C. E. THARALDSEN

ZOOLOGICAL LABORATORIES,

NORTHWESTERN UNIVERSITY

THE

AMERICAN NATURALIST

Vou. LV. September—October, 1921 No. 640

THE RELATION BETWEEN BODY SIZE AND

ORGAN SIZE IN PLANTS*

DR. EDMUND W. SINNOTT

CONNECTICUT AGRICULTURAL COLLEGE

' In the animal kingdom, particularly among its more

highly specialized members where the primitive condition

of indefinite multiplication of similar organs has given

way to a high degree of differentiation, there is neces-

sarily a close correlation between the size of a given

organ and the size of the organism of which it forms a

part. A particularly large individual will tend to have

proportionally large bodily structures, and vice versa. In

the case of the higher plants, however, with their multi-

plication of similar organs and their notably lower

degree of organization and individuation, an interde-

pendence between body size and organ size is certainly

much less obvious. We need only to call to mind the

general similarity between leaves or fruits from small

and from large trees to realize that in these larger, woody

plants, at least, there is no very striking correlation be-

tween the size of the body and the size of its parts. In

certain herbaceous forms, however, evidence has from

time to time been brought forward that such a correla-

tion does in fact exist and that the largest plants

(whether measured by dry weight, height, number of

stalks or yield) are those which bear the largest fruits

and seeds. The importance of such a conclusion on the

theory and practice of seed selection is evident; and the

1 This investigation was carried on by the aid of a grant from the

American Association for the Advancement of Science.

385

386 THE AMERICAN NATURALIST [Vou. LV

question is also of considerable theoretical significance in

that it bears directly on the perplexing problems of in-

dividuality and organization. The aim of the present

paper is to contribute to the solution of this problem by |

undertaking a careful analysis and interpretation of size

relations in a series of bean plants.

HISTORICAL

The problem, at least in certain of its aspects, has re-

ceived attention at the hands of workers in several fields.

Students of the cereal grains, in particular, have been

interested in determining whether those plants which are

large in the sense of having tall or many stems or a high

yield are plants which produce large heads and seeds.

. This question is of importance in seed selection, since

if high yield and large seed size are correlated, it will be

comparatively easy to pick out from a mixture those

seeds which have been produced by high-yielding plants.

Scattered papers on other crops than the cereals also

provide facts of interest.

Most work has been done with the small grains, par-

ticularly wheat and oats. Lyon (11) in 1906, although

not using biometrical methods, observed that in wheat

the weight of the average kernel is not correlated with

the number of kernels per head or with the number of

kernels per plant. He states that the highest yielding

plants have medium-sized spikes and medium-sized

kernels.

Waldron (20) in 1910, working with oats, reported

substantial negative correlations (— .4 to —.6 approxi-

mately) between average weight of seed per plant and

(1) number of seeds, (2) length of head, and (3) length

of culm, thus indicating that the larger the plant, the

smaller were its seeds.

` Results at variance with those of Waldron were

recorded in 1911 by Love (9), Roberts (14) and Myers

(12), working with wheat, and by Leighty (7), with oats.

No. 640] _ BODY SIZE AND ORGAN SIZE © 387

These authors generally agree that there is a positive,

though small, correlation, usually of about the magnitude

of .2 to .4, between the size of the plant, as measured by

height or by yield, and the average weight of the seeds

it produces. Larger plants also tend rather consistently

to have larger heads.

In two extensive memoirs on oats in 1914 Love and

Leighty (10) and Leighty (8) presented an abundance

of data on this problem. In the former paper the authors

find positive and fairly large correlations between the

size of the plant, as measured by yield, and the size of

each head, as measured by its individual yield or by the

number of spikelets or number of kernels which it pro-

duces. The average weight of kernel per plant is not

very consistently correlated with any character repre-

sentative of plant size, however, although most of the

correlation coefficients are positive and many are, in cer-

tain seasons, significantly large. In the second paper

the author, working with another variety, finds consist-

ent, positive and significant correlations between plant

height and average weight of kernels. He points out

that the degree of correlation in all characters studied

increases considerably as the plants are reduced in size

through crowding.

Arny and Garber (1) in 1918 found that, in wheat,

plant height and plant yield are positively correlated

with spike length; and that average kernel weight is

positively and consistently correlated with yield (total

kernel weight) and with number of kernels. The authors

mention unpublished work of Atkinson and Hutchinson

who found substantially the same results.

With corn, the reports are somewhat conflicting.

Ewing (2) found a positive but small correlation between

yield and leaf length and breadth. Hutcheson and Wolfe

(6) found a significant correlation between yield and both

length and circumference of ear. Olson, Bull and Hayes

(13) and others, however, find no significant correlation

between yield and any other character studied.

388 THE AMERICAN NATURALIST [Vou. LV

With apples, Shaw (15, 16, 18) and Stewart (19) have

pointed out that there is little relation between the yield

of fruit on a tree and the average size of the fruit except

in very heavy yields, when fruit size decreases. Small,

young trees may have slightly larger fruits than large,

old trees. Many investigators maintain, however, that

under most conditions thinning will increase the size of

the fruit and thus imply that there is usually a negative

correlation between yield and fruit size.

In tobacco, Hayes (5) found that there is probably no

significant correlation between number of leaves and

average leaf area.

In beans, Harris (3, 4) found a small positive correla-

tion between number of pods per plant and (1) number

of ovules per pod and (2) average weight of seed per

plant, the higher yielding plants thus having somewhat

larger pods and somewhat heavier seeds.

In peas, Shaw (17) found that a positive but small cor-

relation exists between length of vine and average weight

of seed produced, and that this correlation is much

greater in small plants than in large ones.

The total evidence is therefore conflicting. A majority

of the workers report positive correlations between plant

size and the size of the various organs produced. These

correlations, however, even when significant, are in most

cases so small, and there are so many instances where the

coefficients are clearly not significant or are even nega-

tive, that no general conclusion, supported by the whole

body of evidence, can well be drawn.

MATERIALS AND METHODS

The present paper is the result of a study of a group of

562 bean plants grown during the summer of 1918 as a

part of a larger investigation. The beans were Red Kid-

neys, and although they were not members of a pure line,

they were so similar in all characters studied as to in-

dicate that no wide genetic differences existed among

No. 640] BODY SIZE AND ORGAN SIZE 389

them. Seeds of uniform size were selected and were

planted June first. The soil of the plot varied decidedly

in fertility, with the result that some of the plants grew

luxuriantly, many of them reaching a total dry weight of

150 grams, whereas others were much dwarfed, reaching

only 4 or 5 grams at maturity. The bulk of the popula-

tion were intermediate in size.

Over two hundred of the plants were harvested from

time to time during the summer, representatives of all

stages from the young seedling to the appearance of

flowers being obtained. The rest, 344 in number, grew to

maturity and were harvested then. The number of

leaves,’ pods and seeds were counted and dry weight de-

terminations made of the total bulk of the leaves, of the

stem system and of the yield of fruit, separate determi-

nations being recorded for total pods (without seeds)

and total seeds. From these data, the dry weight of the

entire plant, of the vegetative shoot and of the repro-

ductive structures could easily be determined, as well as

the average weight of leaf, pod and seed for each plant.

Correlations’ were then made between the, average

weight of each of these organs, respectively, and the size

of the plant. The latter was represented either by the

weight of the shoot (stem plus leaves), the weight of the

fruit (yield), the number of leaves, the number of pods

or the number of seeds.

RESULTS

The coefficients of correlation thus determined are set

forth in Table I. Eight of these involve the 344 mature

plants studied and one (the first) involves the 218 im-

mature plants. Three of the correlation tables on which

these constants are based—those showing the relation

between shoot weight and (1) average leaf weight, (2)

average pod weight and (3) average seed weight in ma-

ture plants—are also presented in Tables II, III and IV.

It will be noted that in every case there is a positive

2 Very small leaves, less than one third the average size for the plant,

were not counted.

390 THE AMERICAN NATURALIST [Vor. LV

TABLE I

CORRELATIONS BETWEEN BODY SIZE AND AVERAGE ORGAN SIZE

Dry weight of shoot: Average dry weight of leaf5...... r = + .891 + ..009

Dry weight of shoot: Average dry weight of leaf........ r= + 169 + .015-

Number of leaves: Average dry weight of leaf......... r= + .607 + .023

Dry weight of shoot: Average dry weight of pod......... r= + .801 + .033

Total weight of fruit: Average dry weight of pod...... r= + .460 + .029

Number ods age dry weight of pods.......... = -+ .219 + .035

Dry weight of shoot: Average dry weight of seed....... r= + .229 + .035

Total weight of fruit: Average dry weight of seed...... r= + .390 + .031

Number of seeds: Average dry weight of seed.......... r= + .180 + .035

correlation between the average weight of the organ

studied and the size of the entire plant, however meas-

ured. This correlation is in most cases rather small in

amount, but in every instance but one the coefficient is

more ihk six times as large as its probable error and

may therefore be regarded as significant. The coefficients

are least in the case of the seed, somewhat higher for

the pod and of considerable magnitude for the leaf, being

particularly high in the case of the immature plants.

From these figures, therefore, it might reasonably be

inferred that there is a significant relation, though a

small one, between organ size and size of plant in beans,

TABLE II

CORRELATION BETWEEN Dry WEIGHT OF SHOOT AND AVERAGE

Dry WEIGHT OF LEAP

Dry weight of shoot in grams (Class centers)

£ 2 6 10 14 18 22 26 30 34 38 4246505458 62 66707478 82 869094

e +

"is 1

= p iji 1

Z = 1.05

<o 6 a: + Bae seas Ue 1

Se 8 O° 2 447 6655436 61 2 a 1

SS .75 2 6 3237666 5 54 6 34 1

3% .65 Iiri bws i Se f 1

FS -55 1 2 4441 03 Serie 3 1

Po .45 2.838 3° 1i 6 1 1 1

= Si GE Ie Oe

& 2513 10 3

5 15:6 6

aod fe |

r+ 769+ 015

(For plants below 40 grams in prse weight, r= + .842 + .013, for

plants above 40 grams, r= + .129 + .067.)

5 The group of immature plants ae

No. 640] BODY SIZE AND ORGAN SIZE '* 391

TABLE III

CORRELATION BETWEEN Dry WEIGHT oF SHOOT AND AVERAGE

Dry WEIGHT oF Pop

Dry weight of shoot in grams (Class centers)

2 6 10 14 18 22 26 30 34 38 42 46 50 54 58 62 66 70 74 78 82 86 90 94

2.75 1

2.65 |

g 255 1 1 1 |

a 2.45 1 1

— 228/1 ER 1 1 |

S Be Ii ti 1|

SE 2.05 so iai 2 1 |

S = 1.95 SR 143 1 24 2 |

88 1.85 oe ae ey oo at 2 23 i |

FLU li 4 3/145244 42 111 |

Ealo 44 3(/3655145362321 |

pO155| 48 3)/356314332141 $ 2 |

S Ll Gl i2 281821 iS 8 1|

E S bollii Toa |

s Io Sni gS 1 2i

e 116115 3 |

< 1.05|1 |

96} 21 |

.85| 2 |

75121 |

65 1 |

r = + 301 + .033

(For plants below 16 grams in shoot weight, r= + .591 + .043; for

plants above 16 grams, r= + .050 + )

the larger plants producing larger leaves, pods and seeds

and vice versa. This conclusion agrees with that of most

previous workers.

A more intensive study of the correlation tables, how-

ever, reveals certain facts which do not harmonize with

this conclusion, and which suggest that the whole problem

is somewhat too complicated to be solved merely by de-

termining such a series of correlation coefficients as

appear in Table I. A study of the curves connecting the

- means for organ size in the correlation tables from which

these constants have been derived shows that regression

is far from linear and that the character of the curve is

essentially the same in every case. The eight curves for

the means of organ size on body size in the mature plants

are shown in Fig. 1. In every case it is clear that as we

392 THE AMERICAN NATURALIST (Yor LV

—

Orean Size

Bopy Size

. 1. Curves of the means for organ size on body size in mature plants.

(1) Leaf weight on shoot weight; (2) leaf weight on number of leaves per

plant; (3) weight on shoot weight; (4) pod weight on yield of fruit;

(5) pod weight on number o er plant; (6) seed weight on shoot weight;

(T) seed weight on yield of. fruit; seed weig numb: f se per plan

Vertical lines indicate points where division was made into “large”

“small” plants. 3

No. 640] BODY SIZE AND ORGAN SIZE 393

progress from the smaller plants to the larger ones there

is up to a certain point a marked increase in the average

organ size for the plant; but that beyond this point the

curve flattens and thence onward an increase in plant

size has essentially no effect. on the average size of leaf,

pod or seed.

TABLE IV l

CORRELATION BETWEEN Dry WEIGHT OF SHOOT AND AVERAGE

y WEIGHT OF SEED

Dry weight a aon in grams (Class centers)

2 61014 18 22 26 30 34 38 42 46 50 54 58 62 66 70 74 78 82 86 90 94

pms

WNNRN KR O

w whe

(Class centers)

Ke NN wWWwWsI Oe bo

bo NNWwWAH e C9

Average dry weight of seed in grams

mä

T+ 229+ .035

(For plants below 16 grams in shoot weight, r= + .555 + .046; for

plants above 16 grams, r= — .030 + .043.)

There is thus evidently a much higher correlation be-

tween body size and organ size in small plants than in

larger ones, and a single correlation coefficient covering

both groups clearly fails to give an accurate picture of

the facts. Each of the eight correlations involving the

mature plants was therefore divided into two parts, one

including the small plants and one the large ones, the

line of division coming at approximately the point where

394 THE AMERICAN NATURALIST [Vor. LV

the curve of means stops ascending and begins to flatten

out. This point is marked by a vertical line on each of

the curves in Fig. 1 and in the same way in Tables IT, III

and IV. The coefficient of correlation was now deter-

mined both for the group below this point and for the

group above it—for the small plants and for the large

ones. The constants thus derived are shown in Table V.

A study of these figures makes it clear that in the smaller

plants there is a decided correlation between body size

and organ size (always exceeding ten times its probable

error), but in the larger plants practically none at all.’

This emphasizes the conclusion drawn from the character

of the curve of means, namely, that up to a certain point

increased body size is followed by increased size of organs

produced, but that beyond this point there is no relation

between the two.

Lear Dize

btb eh We ees ts ees Wer Maen eee foe Meer? Bd L

SuHoot S1zeE

Fic. 2. Curve of means for leaf weight on shoot weight in immature plants.

It is of interest to note that in the case of the immature

plants the curve of means for leaf size rises steadily as

8It will be noted that very similar results were obtained by Shaw (17)

in correlating seed size with vine length in peas.

No. 640] BODY SIZE AND ORGAN SIZE 395

plant size increases (Fig. 2), showing no sign of the

flattening characteristic of the mature plants.

What reason may we assign in the case of the mature

plants for this radical difference between large individ-

uals and small ones? And why, in immature plants,

should no such difference exist? The suggestion at once

comes to mind that there may really be no relation be-

tween body size and organ size in any case, but that

organ size may be determined, instead, by the size of the

particular axial growing point from which the organ has

TABLE V

CORRELATIONS BETWEEN Bopy SIZE AND AVERAGE ORGAN SIZE IN THE

ENTIRE GROUP (OF MATURE PLANTS); IN THE SMALLER PLANTS (THOSE

BELOW THE VERTICAL LINE IN Fic. 1); AND IN THE LARGER PLANTS (THOSE

ABOVE THE VERTICAL LINE IN FIG. 1).

Entire Smaller | Larger

Group Plants | Plants

Shoot: Average leaf.............. | 4.769.015 | +.842+.013 | +.129+.067

Number of leaves: Average leaf. +.607 +.023 | +.665+.029 | oa peo eee

Shoot: Average | sajad pee IE +.301 +.033 | +.591+.043 | 050+

Total fruit: Average pod......... +.460+.029 | +.652+.040 | +370 4.040

Number of pods: pe erage pod..... +.219+.035 | +.486+.048 | —.083 +.045

Shoot: “Average m T N +.229+.035 | +.555+.046 | —1030+.043

Total fruit: Average seed......... +.390+.031 | + Pte 047 | +.206+.041

e mber ie seeds: ds: Average seed.. +.180+.035 | +.509+.045 | —.059 +.045

developed. It is a matter of common observation that in

most herbaceous plants the diameter of the newly formed

stem internodes (and therefore presumably the diameter

of the terminal growing point which gives rise to the

primary tissues of the stem) is comparatively small in

the seedling, but increases slowly as the plant grows

larger until a presumably optimum diameter is attained

which is rarely exceeded except in the case of very rank

and luxuriant shoots.* Further growth of the plant as -

a whole results in an increase in the length and number

of its stems, but in no increase in their primary diameter.

Stem diameter is of course not uniform, many lateral

4 The thickness of these primary tissues of the young stem is of course

increased later by secondary growth, with which we are not concerned.

396 THE AMERICAN NATURALIST [Vor. LV

branches and even the main ones under unfavorable con-

ditions, or as vegetative growth slackens, being compara-

tively small; and we know that organ size also varies

considerably in the individual. The point to be empha-

sized here is that there tends to be a maximum for the

primary diameter of the stem which is attained while the

plant is still fairly small, and which thereafter is nor-

mally not exceeded, no matter how great the total size of

the plant body may eventually become. The same rule

applies of course to woody plants, the twigs of a large

tree being no thicker, other things being equal, than those

of a small tree, although both are usually thicker than

the early axis of the seedling.

Now if the organs (leaves and fruits) developed by

the primary meristem owe the size which they finally

attain to the size of the growing point from which they

arise; or if, to put it another way, all the structures

developed at a given growing point—the stem axis and

the lateral organs—are correlated with one another in

size, then the biometrical results which we have set forth

in our bean plants are readily explicable. The compara-

tively small plants are, on this supposition, the ones _

which did not attain at maturity sufficient size to have

arrived at the maximum stem (growing point) diameter;

and the smaller the plants, the smaller is their stem diam-

eter, down to depauperate individuals whose mature

primary axes are no stouter than those of the seedling.

In these smaller plants, therefore, the significant cor-

relation which we observed would naturally be expected

between organ size (dependent on growing point size)

and body size (definitely related to growing point size).

In the case of the larger plants, however (those above

the point at which the curve of means flattened out),

where the maximum stem diameter or growing point size

has already been attained and where, therefore, there is

no relation between body size and growing point size, it is

only natural that there should be (as we observed) no

correlation at all. Furthermore, in the group of imma-

No. 640] BODY SIZE AND ORGAN SIZE 397

ture plants studied, which included everything from seed-

lings to plants coming into flower, there is naturally a

very close relation between body size and growing point

size, since these individuals all belong to that portion of

the plant’s life history where its primary growing points

are progressively increasing in diameter, the maximum

being attained in beans just before the blooming period.

It would be only natural that in this group of plants in

which both body size and growing point size are pro-

gressively rising, there should be a high correlation be-

tween body size and organ size.

This hypothesis of a direct relation between the size

of an organ and the size of the growing point from

which it arose will evidently explain the facts which we

have observed in the case of our bean plants. The

problem now remains to discover a means whereby we

may determine more directly the soundness of the hy-

pothesis. The size of an organ can be measured fairly

accurately either by weighing it or by determining its

dimensions and computing its volume. To get a meas-

urement which shall represent at all adequately the size

of the growing point, however, is a more difficult matter.

The cross-sectional area of the stem axis which is pro-

duced from the growing point might be used; but since

the terminal growing point is of course a primary meri-

stem entirely and since the early activity of a secondary

growing point or cambium often affects almost imme-

diately the diameter of the stem, it is evident that stem

cross section, particularly in regions very far removed

from the growing point, can not be counted on as an ac-

curate indication of growing point size. We can confine

ourselves to a tissue which is entirely primary in its

origin, however, and which is not affected by subsequent

secondary growth if we measure simply the pith. The

magnitude of the correlation between organ bulk and

cross-sectional area of the pith of the internode below

the attachment of the organ might be expected to give

us a fairly good idea as to whether organ size and grow-

398 THE AMERICAN NATURALIST [Von LV

ing point size are definitely related to one another or

not. The organ most readily studied and most clearly

significant in such a problem is of course the leaf.

The bean plant is evidently not well suited for such a

study, since its pith is rather irregular in outline and not

sharply delimited. The twigs and leaves of the rock’

maple (Acer saccharum), however, on which the writer

is carrying on some other work, have proven very satis-

` factory for an investigation of this kind. The pith in

this species is approximately circular in cross section

and is sharply delimited, and the leaves are of uniform

and fairly considerable thickness.

A series of twigs collected during the summer from a

single tree were studied. The area of the blade was de-

termined by tracing its outline on standard weight paper,

cutting this out and weighing the cut-out. Blade thickness

was measured by a micrometer caliper at two points away _

from the main veins and situated symmetrically on oppo-

site sides of the midrib, the average of the two measure-

ments being taken. The product of area X thickness of

course gives us the blade volume. To determine pith area

a cross section was cut at the middle of the internode..

the pith diameter measured in two directions at right

angles to each other by a micrometer stage, the results

averaged and the area computed therefrom.

Total blade volume (the sum of the volumes of the two

blades borne at a given node) was correlated with the

cross-sectional area of the pith of the internode below for

over 100 leaf pairs from this tree, taken from all parts

of its crown. The results are shown in Table VI. It is

quite evident from the size of the correlation coefficient:

(+ .807+ .024) that there is an unquestionable relation-

ship between leaf size and pith area, the size of the leaf

being governed pretty largely by the stoutness of that

portion of the twig from which it springs. It would seem,

therefore, that the size relationships between the struc-

tures laid down by the terminal meristem persist as these

structures develop to maturity, these relative sizes re-

No. 640] BODY SIZE AND ORGAN SIZE 399

maining constant regardless of the actual size which is at-

tained. The terminal growing point, like the animal

embryo, develops as asymmetrical and interrelated whole.

Although the size of the leaf thus seems to be closely

dependent upon that of the growing point, the size of the

reproductive organs is evidently much less so. We have

TABLE VI

CORRELATION BETWEEN THE BLADE VOLUME (IN CUBIC CENTIMETERS) OF

THE Two LEAVES AT A NODE, IN ACER SACCHARUM, AND THE Cross SEC-

TIONAL AREA (IN SQUARE MILLIMETERS) OF THE PITH OF THE INTERNODE

BELOW

Blade volume in cubic centimeters (Class centers)

3.5.7 9 1.11.3 1.5 1.7 1.9 2.1 2.3 2.5 2.7 2.9 3.1 3.3 3.53.7 3.9 4.1 4.34.54.7

7.05 | 1

Pith area in square millimeters

(Class centers)

AAAAAA HH

pæ

Eaa

MRS w&

SRaRRBSRARE RSS

r == + 807 + .024

seen that leaves do not reach their maximum size except

in plants with shoots of about forty grams dry weight or

more. The maximum pod and seed size, however, is at-

tained in much smaller plants, usually those in the vicin-

ity of sixteen grams. In other words, a reduction in the

size of the growing point is felt much more quickly by

the leaf than by the fruit. Itis only in plants which are

really depauperate that the fruit and seed size is reduced

400 THE AMERICAN NATURALIST [Vou. LV.

below the normal. The reason for this may lie in the fact

that the flower is an independent axis rather than a lat-

eral organ like the leaf and may therefore be less affected —

by the size of the main axis from which it springs. It is

well known that flowers are more constant in size. and

other characters and less affected by environmental con-

ditions than are vegetative organs.

It should be recognized that in the grasses, where most

of the. work along this line has been done, conditions

are somewhat different from those in dicotyledons such

as the bean; since the plant body, or at least each shoot

or culm, is essentially determinate in growth, with a lim- `

ited number of parts. This may sometimes affect the

statistical results obtained, but we believe that conditions

are fundamentally the same in the two groups. It is of

interest to note the conclusions of Leighty and others, for

small grains, viz., that the correlations between organ

size and plant size are considerably higher in small,

poorly developed plants than in large ones,—a situation

precisely similar to that which we have reported in beans.

The whole problem can perhaps be solved best, how-

ever, by studying such an indeterminate type of plant

body as is characteristic of the ordinary dicotyledon.

CoNCLUSIONS

We may conclude, therefore, that the size of the plant

body is not the direct causative factor in determining the

size of the leaves, fruits or seeds which it produces, as

has been suggested or implied by many investigators, but

that the size of any given organ depends rather upon

the size of the growing point out of which it has been

developed. Any factor, be it age, moisture or food sup-

ply, which alters the size of the meristem, will thus alter

the size of the organs produced by this meristem. There

seems to be nothing in these higher plants closely cor-

responding to the definite organization with which we

are familiar in the animal individual, where size of body

is definitely related to size of organs.

No. 640] BODY SIZE AND ORGAN SIZE 401

The present study emphasizes the difficulties lying in

the path of the investigator who attempts to solve such

a problem as this merely by determining a single series

of biometrical constants without taking into account the

various morphological and papsiplogtca] factors which

may be involved.

SUMMARY

1. The problem of the relationship between the size

of the plant body and the size of the organs it produces

has been studied by various workers, who find that in

most cases there is a small but significant correlation

between these characters.

2. In a series of bean plants, the coefficients of cor-

relation were determined between plant size, as meas-

ured by dry weight of shoot, dry weight of fruit, number

of leaves, number of pods and number of seeds; and the

average dry weight per plant of leaf, pod and seed.. A .

positive and significant correlation, though usually a

small one, was found in each case.

3. An examination of the curve of means for organ

size on plant size shows that in each case the curve rises

steeply at first and then flattens out. In other words,

an increase in the size of the plant is accompanied by

an increase in the size of its organs if we consider com-

paratively small plants only; but after a certain size

is reached, any further increase in plant size is not

followed by increase in organ size. Separate correla-

tions between plant size and organ size made for small

plants (those below the point where the curve flattens)

and for large ones (those above it) showed a very de-

cided correlation in the former and practically none at

all in the latter.

4. These facts suggested that the size of an organ may

not be correlated with body size at all, but rather with

the size of the axial growing point from which it develops,

In support of this idea attention is called to the fact

that during the early stages of a plant’s growth there is `

402 THE AMERICAN NATURALIST [Von. LV

up to a certain point a progressive and parallel increase

in the size of the plant and in the size of the primary.

meristems of its axes; but that after this point is reached

meristem size remains constant and further increase in

body size is not accompanied by any increase at all in

that of the growing point.

5. Favorable material to test this hypothesis directly

was afforded by twigs and leaves of Acer saccharum.

The correlation between the blade volume of a given leaf

pair and the cross-sectional area of the pith of the in-

ternode below (used as an index to the size of the grow-

ing point from which the leaves had developed) was

found to be high (+ .807 + .024).

6. It is therefore concluded that the size of a plant

organ (leaf, fruit or seed) is dependent not upon the:

body size of the plant on which it is borne, but rather

upon the size of the growing point from which it devel-

oped. :

BIBLIOGRAPHY

ei Amy, A. C., and R. J. Garber. Variation and Correlation in Wheat,

with Special Reference to Weight of one Planted. Jour. Ag

pa 14: 359-392, 1918

2. Ewing, E. C. Correlation of Chariicters in Corn. Cornell Erp. Sta.

Bwl., 287: 67-100,

3. Harris, J. Arthur. On i Correlation between Somatie Characters and

Fertility. II. greva from Phaseolus vulgaris. Amer. Jour.

Bot., 1: aes 1, 1914.

_ papi dena Study of the Factors Influencing the Weight

of the Bean Seed. II. Correlation between Number of Pods per

Plant and Seed Weight. Bull. Torrey Bot. Clu, 43: 485-494, 1916.

5. Hayes, H. K. Correlation and Inheritance in xp icotiana tabacum, Conn.

Exp. Sta. Bull. (New Haven), 171: i

6. Hutcheson, T. B., and T. K. Wolfe. Relation panei Yield and Ear

Cha sa

7. Leighty, Correlation of Characters in Oats, with Special Refer-

ence k Pata ing. Amer. Breeders Assoc. Rept., 7: ary 1911.

8. ————. Variation and Correlation of Oats. II. Effect of sai

in Environment, Variety and Methods on Biometrical en Cor-

nell Exp. Sta, Mem., 4, 1914.

9. Love, H. H. A Study S the Large and Small Grain Question. Amer.

Breeders Assoc. Rept., 7: 109-118, 1911

10. Love, H. H., and C. E. Leighty. Variation aú Correlation of Oats. I.

: Studies Showing the Effect of Seasonal Changes on Biometrical Con-

stants. Cornell Exp. Sta. Mem., 3, 1914.

No. 640] - BODY SIZE AND ORGAN SIZE “£03

14.

Lyon, T. L. Some Correlated nieh in Wheat and their Transmis-

sion, Amer, Breeders Assoc. Rept., 2: 29-39, 1906.

Myers, C. H. Effect of Fertility upon stig ay Correlation in

Wheat. Amer. Breeders Assoc. Rept., 7: 61-74, 1

i pres vs 1 P. Bull and H. K. Hayes. Ear eds Selection and

irg e in Gom. Minnesota Exp. Sta. Bull., 174, 1918.

. Roberts, H. F. Variation and Correlation in Wheat. Amer. Breeders-

Assoc. rade 7: 80-109, 1911.

. Shaw, és — Variation in Apples. Mass, ies, Sta. Rept., 1910:

194-2

16. ————_ er matic Phe we of Apple Varieties. Mass. Exp. Sta.

Rept., 1911: 177-

17. ————— Her ih Correlation and Variation in Garden Peas. Mass,

Exp. Sta. Rept., 101.

18. —— A Sa on pation in Apples. Mass, Sta. Bull., 149:

21-36, 1914.

19. Stewart, J. P. Factors ee Yield, A Size and Growth in

Apples. a ba Rept., 1911: 401-511

5 ~ L. R. A Suggestion Baoe Light and Heavy Grain.

AM , 1910.

. NAT., 44: 48-56

DECREASE IN SEXUAL DIMORPHISM OF BAR-

EYE DROSOPHILA DURING THE COURSE

OF SELECTION FOR LOW AND HIGH

FACET NUMBER}

PROFESSOR CHARLES ZELENY

UNIVERSITY OF ILLINOIS

In the bar-eye race of Drosophila the mean eye facet

number of the males is higher than that of the females,

though there is an overlapping in range as shown in

Table I. In the unselected white bar stock used as the

starting point of a series of selection experiments this

difference amounted to 6.12 factorial units, a factorial

unit being one that produces a ten per cent. change in

facet number. If this value were fairly constant it

would be possible in treating the selection data statis-

tically to reduce the facet values of the two sexes to a

common basis in much the same way that Galton obtained

a mid-parental value in his studies of the inheritance of

human stature. Such a procedure was followed by

Zeleny and Mattoon (1915)? in their paper on selection

in red bar-eye. In attempting to apply the same method

to the white bar series it was discovered that the differ-

ence between the two sexes is not constant but decreases

during the course of selection. It is therefore not prac-

ticable to apply a constant coefficient for the reduction of |

the value of one sex to that of the other.

The selection experiment in question started with a

white bar stock which had been obtained by crossing

white full-eye to bar-eye. Single pair, brother and sister,

matings were adhered to with a few breaks due to steril-

ity. In the low line the lowest available virgin female

1 Contribution from the ia et Laboratory of the University of Illi-

nois, No. 187.

2 Zeleny, C., and Mattoon, E. W., 1915, ‘‘The Effect of Selection upon

the ‘Bar Eye’ Mutant of Drosophila,’’ J. Exp. Zool., 19: 515-529.

404

No. 640] DIMORPHISM OF BAR-EYE DROSOPHILA 405

a 6.12

fo 4.12 z.

$ 3.12 bz

Faas Ra

twb

& 2,12 38

F 5°

4 1.12 z

; i

0.00 u

P 10 30 40 42

generations

Fig. 1. Sexual dimorphism > g = the apoio sua Direct m

ings only. The difference betw males and females is expressed in hearer

units, a factor yait value baie one capable a pe ne a ten per cen

I t

12 units and in the Apie such a value is represented by the upper hori-

departur y

and and the scale of departure from the original value in the unselected

population at the right hand.

was mated to her lowest available brother in each genera-

tion with duplicate matings to insure the continuation of

the line. In the same way in the high line the female

with the highest facet count was mated to her highest

brother.

The data as given in the present paper are grouped

under two heads. In Table II and Fig. 1 are included

only those individuals in the direct line. In Table III

and Fig. 2 there are included, in addition, the sib matings

in each generation. The general results obtained from

the two groupings are alike, but the data based on all the

matings give a smoother curve because of the larger

numbers of individuals involved.

Table I shows the distribution of frequencies of fac-

torial values for eye facet number in the unselected

population of white bar, number 127, which served as the

departures from original

dimorphiam

EN SORE

Fic. 2. Sexual dimorphism. All matings.

406 THE AMERICAN NATURALIST [Von. LV

TABLE I

DISTRIBUTION OF FREQUENCIES FOR FEMALES AND MALES IN THE UNSE-

LECTED WHITE BAR POPULATION AND IN THE FORTIETH GENERATION

or Low AND HIGH SELECTION ror Eye FACET NUMBER

Each class has a range equal to ten per cent. of its mean

Classes in Factorial Unselected SERA, raa palaces

Facets in Each Units Population Low High

Class Se Re reas PA ise eee

9 é 9 a 9 ? 9 J

359-396....... 419.07 | +12.95

325-358. ...... 418.07 | +11.95

294-324....... 417.07 | +10.95 1

266-293....... 416.07 | + 9.95 1

AL +15.07 | + 8.95 3 2

218-240....... | 414.07 | + 7.95 5

197-217.......| +13.07 | + 6.95 15 17

178-196.......| +12.07 | + 5.95 16 1| 20

161-177....... | +11.07 | + 4.95 26 gs | 21

ee +10.07 | + 3.95 19 2 | 14

132145... | + 9.07 | + 2.95 oe 23 | 20

dies 8.07 | + 1.95 35 1 2

108-118....... 4+ 7.07 | + 0.95 7| 33 9 2

ee + 6.07 0.0 ni u 4

woro o iene + $07 -10 27 | 55

BBR a, + 4.07 | = 2:05 | 25 | 37

B 7: 3.05 | — 3.05 | 29

pS oe + 207 | +405 | | 3

Gee Sy 3 A RE e

r ec + 0.07 | — 6.05 18 1

19-65 A 0. —~ 7.05 | 58 8 3

Meh a ie TOE OL ee 71 19 | 34

a ~ 903} 2908| 47 2 | 33 | 24

ret os ae — 3.93 | —10.05 | 40 il 21 9

a8 Ooo Laal- n 15 1 2

96683 oc: i 6.08 | 1908 te 1

9-00" 2s. — 6.93 | —13.05 | 3

‘cag 2S — 7.93 | —14.05 1

PE — 893 | -1605

mn Ea | = 9.93 | -16.05 1

Tos | a89 | 441 | 85 | 77 | 69 | 104

basis of the present selection series, and of the fortieth

generation of low and of high selection for eye facet

number. The column at the extreme left gives the facet

values of the classes. The second and third columns give

the factorial values of these classes, a unit of value being

a difference in germinal or environmental factors which

produces a ten per cent. change in facet value. The mean

values of the unselected population are taken as zero and

in the second column there are given the values on this

scale for the females and in the third column for the

No. 640] DIMORPHISM OF BAR-EYE DROSOPHILA 407

males. It will be noticed that there is considerable varia-

tion in both females and males, but the mean values as

represented by the figures in heavy type are a consider-

TABLE II

SEXUAL DIMORPHISM. Direct LINE

Sexual dimorphism for each of the 42 generations of selection for low

and high facet number. All values are in ten per cent, factorial units. Such

a unit is any difference in germinal or environmental factors which produces

a ten per cent. change in facet number.

Low Line High Line Difference in

Generation p Dimorphism

Dimorphism ia aren Dimorphism Ps pirisee and tow Ties

Pica 6.12 6.12 00 0.00

i eee 3.34 —2.78 2.74 —3.38 -9

ics ie 2.25 3.87 2.36 —3.76 0.11

Far ee 0.95 —5.17 1.54 —4.58 0.59

r TAE 1.12 —5.00 1.61 4.51 0.49

oes 2.20 3.92 4.89 —1.23 69

payee 2.61 —3.51 2:13 —3.99 —0.48

NEU 3.26 —2.86 3.82 — 2.30 56

Bon 0.92 —5.20 4.17 1.95 8.25

t BORSE 0.69 5.43 4.74 —1.38 4.05

5 E E ee 0.48 - 4.17 —1.95 3.69

eA ESE pe 217 —3.95 338" 2.74 1.21

Lag 1.49 —4.63 5.28 —0.84 3.79

pe Bares 20 —4.92 3.20 —2.92 00

ct ene er 1.67 i —4.45 3.22 —2.90 1.55

Bs ee 1.26 —4.86 2.41 3.71 1.15

i 1i PERETE, 0.21 —5.91 2.16 — 1.95

jy ANE 1.13 —4.99 3.33 —2.79 2.20

Bal Se 1.51 —4.61 4.29 ee 2.78

yt Benes 0.78 5.34 2.31 —3.81 1.53

i KREE 0.55 —5.57 2.45 —3.67 90

9 FARAS 0.59 — 3.58 — 2.54 2.99

e o ANTA 1.08 —5.04 0.79 —5.33 —0.29

PS Rag 1.38 —4.74 — —

ES 3.17 —2.95 1.87 —4,25 —1.30

eee 1.05 5.07 2.18 —3 1.13

Beles: 0.61 —5.51 0.57 —5.55 —0.04

of Ba a 1.14 —5.98 2.19 —3.93 1.05

Bes eas s —4.77 1.24 —4.88 —0.11

4 Sie 0.75 5.37. 1.46 —4.66 0.71

ms i R —5.08 1.40 —4.72 0.36

DE sois 1.07 —5.05 —3.83 1.22

n cic: 1.01 —5.11 1.46 —4.66 0.45

Er E 1.97 —4.15 2.09 — 4.03 0.12

Be ces 1.03 —5.09 2.67 —3.45 1.64

eoi P 2.75 —3.37 4.11 —2.( 1.36

Pee 0.83 —5.29 3.22 —2.90 2,39

PaE 2.10 4.02 2.1 —3.99 -03

Bs 0.71 —5.41 1.79 — 4.33 1.08

Mo EPEE 3.59 2.08 —4.04 —0.45

w. 0.26 —5.86 2.20 —3.92 1.94

2 aaa 0.95 —5.17 1.93 —4.19 0.98

ts 4.75 3.24 —2.88 1.87

408 THE AMERICAN NATURALIST [Vou. LV

able distance apart, six class units or, to be more exact,

6.12 as stated in the first paragraph. In both low and

high selection lines it is to be noticed that the variability

has decreased and at the same time the mean values of

females and males have approached each other. This

approach is greater in the low than in the high line.

Tables II and LI give the sexual dimorphism for the

parental unselected generation and for each of the 42

generations of selection. Each table includes, for both

low and high lines, the value of the dimorphism for each

generation and the decrease since the beginning of selec-

tion. In the last column of each table there is a com-

parison of the dimorphisms in the two selection lines.

Table II includes only the offspring in the direct line.

Table III gives the data for the same series, but includes

sib matings as well as those in the direct line. :

Figures 1 and 2 show the decrease in graphic form.

The vertical scale is in factorial units as described above.

The upper horizontal line is at the level of the dimor-

phism value of the unselected population, which is 6.12

units in favor of the males. The lower horizontal line

indicates the position of zero difference between the

sexes. The scale at the left gives the actual dimorphism

values and the scale at the right the departures from the

original value. The continuous zigzag. line gives the

values for each generation in the low selection series and

the dotted line those in the high selection series. Special

emphasis is to be laid on the fact that selection was not

for low or high sexual dimorphism, but for low or high

facet number regardless of dimorphism. In no sense can

it be considered as a direct selection for degree of

dimorphism.

The result obtained when the offspring of all matings

are taken (Fig. 2) is not essentially different from that

obtained in the direct line. Since it gives the smoother

curve because of the larger number of individuals it will

be used as the basis of the following discussion.

In the low selection line the dimorphism drops very

o. 640] DIMORPHISM OF BAR-EYE DROSOPHILA 409

rapidly from its original value of 6.12 units to 1.12 in the

fourth generation. From that point on it shows some

increase and fluctuates irregularly for a number of gen-

TABLE III

SEXUAL DIMORPHISM. ALL MaTINGS |

Sexual dimorphism for each of the 42 generations of selection for low

and high facet number. All values are in ten per cent, factorial units. Such

a unit is any difference in germinal or environmental peat which produces

a ten per cent. change in facet number.

tae Line High Line | Difference in

Generation Phen gol

Dimorphism big semen wk Dimorphism ia Dienstes Pe ie ie tins

Py oes 6.12 00 6.12 00 0.00

h ee 3.34 —2.78 2.76 | —3.36 —0.58

AD aa 1.95 —4.17 1.78 — 4.34 —0.17

2o 1.63 —4.49 16 T -4er —0.18

Boose 1.12 —5.00 1.72 4.40

Oe cas 2.20 —3.92 3.47 —2 127

6.2.5 2.06 —4.06 2.44 — 0.38

7 EA ONER 3.36 —2.76 3.70 —2.42 0.34

aa 84 5.2 3.87 2.25 3.03

p” AES 0.68 — 5.44 3.35 —2.77 2.67

p i 1 E 0.98 —5.14 3.64 —2.48 2.66

| a KATSE 99 —4 3.29 —2.83 1.30

| A NAA 1.28 —4.84 4.74 —1.38 3.46

5 MAE AS 88 —4.24 3.20 —2.92 1.32

Ioe 0.43 —5.69 3.42 2.70 2.99

aE EEE 1.42 —4.70 2.44 —3.68 1.02

AG i208 0.31 —5.81 2.80 —3.32 2.49

i a ee 0.61 —5.5] 3.19 —2.93 2.58

ISo, 1.07 —5.05 4.29 —1.83 3.22

IDni 0.81 —5.31 2.31 —3.81 50

KEA 0.33 —5.79 2.66 —3.46 2.33

rA S 0.33 5.79 3.70 —2.42 3.37

yO 0.92 —5.20 3.72 —2.40 80

ja, PADA 1.42 4.70 3 —2.79 1.91

i GP ee 2.23 —3.89 2.97 —3.15 0.74

BOs ea 0.83 —5.29 2.71 —3.41 88

es 1 ae ee 1.26 — 4.86 1 —4.41 0.45

Siti 0.75 —5.37 2.15 —3.97 40

DB. ae 1.30 — 4.82 1.07 —5.05 —0.23

bt Bor! 0.94 —5.18 1 —4.90 0.

BOL ee 0.89 —5.23 1.32 —4.80 0.43

Blo 1.23 —4.89 1.79 —4.33 0.56

SF i 0.69 —5.43 1.60 —4.52 0.91

88... 1.55 —4.57 1.78 —4.34 0.23

RET 1.00 —5.12 2.51 —3.61 1.51

bs Plas 1.84 —4.28 2.68 —3.44 0.84

woes 0.86 —5.26 2.58 —3.54 1.72

Bisa 1.61 —4.51 2.44 —3.68 0.83

eet 0.71 —5.41 2.01 —4.11 1.31

Ki NES! 1.76 —4,37 2.03 —4.09 0.28

PES, 0.56 —5.56 2.32 —3.80 1.76

A EEES 0.77 —5.35 1.90 —4,22 1.13

7 ie RAEE 1.37 —4.75 3.24 — 2,88 1.87

410 THE AMERICAN NATURALIST [Vor. LV

erations, but eventually settles down to a-value not far

from 1.00. This last value may therefore be taken as

the normal dimorphism in a homozygous low bar popu-

lation.

In the high selection line there is a similar rapid de-

cline, in this case from 6.12 in the parental generation to

1.45 in the third selection generation. There is then a

pronounced increase and fluctuation between 2.00 and

4.00 from the fifth to the twenty-third generation. After

a second general decline the value remains for most of-

the time between 1.00 and 2.50.

While the general course of the curves is the same in

the two cases, the low selection line is the more consistent

and reaches the lower level. .

The probable explanation of the decrease in dimor- |

phism in both selection lines is to be sought in the fact

that some of the accessory factors affecting facet number

are sex-linked. The unselected population is a mixed one

and accordingly has a considerable degree of hetero-

zygosis for these factors in the females. One of the

results of selection for facet number with inbreeding is

a decrease in this heterozygosis, leaving the females

homozygous for low facet factors in the low line and for

high facet factors in the high line. As long as the popu-

lation is homozygous it is probable that the degree of

dimorphism does not change to any great extent because `

there is no disturbance of the factorial proportions. An

increase on one side is accompanied by a proportionate

increase on the other. The same would be true as far

as the heterozygous females are concerned if the domi-

nance of low and high factors were alike. If, however,

the factors for low facet number have a higher dominance

coefficient than those for high facet number it follows

that selection as practised will produce a decrease in

dimorphism. That this will be true in both low and high

lines is made clear by the consideration that hetero-

zygous females under the stated condition must be lower

than the average of the homozygous ones in the same

population. When these heterozygotes are eliminated

No. 640] DIMORPHISM OF BAR-EYE DROSOPHILA 411

by selection it follows that the mean value of the females

approaches that of the males.

The explanation for the general decrease seems clear

enough, but the reasons for some of the details are not

so evident. The dimorphism in the high selection line

is higher than that in the low selection line in nearly

every generation. In part this difference may be due

to a faulty method of determination of the dimorphism

value. If degree of dimorphism were expressed in

facet numbers the high line would of course have a

much greater difference between the sexes than the

low line. The factorial unit method with its loga-

rithmic scale based primarily upon the observed ef-

fects of temperature is undoubtedly a much better one

than that based upon the unreduced facet numbers,

though a simple ratio between the facet numbers of the

two sexes might have served equally well in the present

cease.* But the correction is only an approximation and

may not be great enough to obtain the most rational

value. -

In part, however, the difference between the two lines

may be due to the greater frequency of appearance of

mutant accessory factors in the high than in the low line.

Such factors introduce a new heterozygosis and therefore

increased dimorphism which is eliminated only by later

selection. That such mutants appear is evidenced by in-

crease in standard deviation.

The appearance of mutants and their prompt elimina-

tion by selection may account also for some of the irregu-

lar fluctuations as observed in both lines. It is possible,

however, that environmental factors may exercise a dif-

ferential effect and thereby change the dimorphism. In

favor of this view it may be pointed out that there is some

tendency for the values in the low and high lines to fluctu-

ate together and this may be explained by the fact that

the food cultures of the two lines in any generation were

usually made up from the same food mass.

3 Zeleny, C., 1920, ‘‘ The. Tabulation of Factorial Values,’’ Amer, Nat.,

54: 358-362.

DATA CONCERNING LINKAGE IN MICE

W. L. WACHTER

Bussey INSTITUTION

In exploiting the genetics of any plant or animal it is

of great importance to breed intensively and extensively

until the genetic material is exhausted. Even after the

dominance or recessiveness of a certain gene has been

established, there yet remain to be determined the rela-

tionships between this and other genes known to consti-

tute the hereditary complex of that particular plant or

animal. These possible relationships are allelomorphic,

independent, or linked. Mendel was able to formulate

certain laws of heredity from the data he obtained in his

breeding experiments; but had he continued his work to

include other factors he would undoubtedly have been

led to predict or formulate a theory of linkage. The

chromosome theory of heredity involves independent in-

heritance of genes which lie in different chromosomes no

less than linked inheritance of those which lie in the

same chromosome. It is equally important in relation

to this theory to know whether two genes are linked or

not linked.

A very meager amount of linkage has so far been found

in mammals. Cases have thus far been observed only in

rabbits, rats, and mice. In rabbits Castle (1920) has

recently shown that there is a probable linkage between

the factors for dilution and the English type of spotting.

In rats Castle and Wright (1915), and Castle (1919),

had previously shown albinism, red-eye, and pink-eye to

be included in the same linkage system. In mice the

genes for pink-eye and albinism were first shown to be

linked by the work of Darbishire (1904), as was pointed

out by Haldane, Sprunt, and Haldane (1915), who con-

firmed the idea by observations of their own.

412

No. 640] DATA CONCERNING LINKAGE IN MICE 413

A great amount of data has been obtained by breeding

the house mouse. Yet a further intensive study is desir-

able of the mutant characters discovered, especially as

regards linkage, for from such study alone can be de-

duced the probable localization of the genes in the chro-

mosomes. According to Harvey (1920), the number of

chromosomes in the gametes of the house mouse has been

estimated by different observers at from 8 to 30, the

more recent observers placing the number at from 20 to

24. One of these is probably a sex chromosome which is

heterozygous in the male and homozygous in the female.

In this sex chromosome is apparently located the lethal

gene of the Japanese waltzing mouse reported by Little

(1920). In one autosome are located the linked genes

for pink-eye and albinism. In another autosome must

be located the gene for agouti and its allelomorphs, non-

agouti, light-bellied agouti, and yellow; to a third auto-

some is referred piebald spotting; to a fourth, black-

eyed-white spotting; while dilution and its allelomorph,

intensity, will probably have to be referred to a sixth

autosomal group. Thus seven of the twenty or more

chromosomes ‘are genetically identified, leaving thirteen

or more yet to be identified with visible characters or

mutations yet to be discovered.

Among modifying genes in mice, Little (1915) and

Dunn (1920) have shown that the ‘‘blaze’’ or ‘‘white-

face” variation which modifies the piebald pattern is

heritable. Dunn (1920) also reports a ‘‘belly-spot’’ te

segregate independently of self and piebald, though ap-

parent only in animals heterozygous for self and piebald.

General modifying factors, increasing or decreasing the

area of pigment in the black-eyed-white and piebald pat-

terns, are also heritable. Very little is known at the

present time concerning these modifying factors, and

nothing about their linkage relations.

414 THE AMERICAN NATURALIST [Vour. LV

Tue RELATION oF AGOUTI To PIEBALD

Dunn’s (1920) data on the cross, agouti X piebald, in-

dicated a cross-over percentage, between the genes for

agouti and piebald, of only 46.23 + 1.20, in a total of 783

young produced by a back cross of F, animals to the

double recessive, non-agouti piebald. This shows a de-

viation of 3.77 per cent. from the normal 50 per cent.

value expected if the genes for agouti and piebald assort

independently. This deviation is barely more than three

times the probable error. The data were also taken

incidentally from crosses in which the primary object

was to ascertain the relationship between the genes for

pink-eye and piebald. This may have influenced the final

result in some way, such as less attention being placed on

the discrimination of agouti. Therefore further data

seemed especially desirable on this relationship.

All later crosses were made with reference only to the

agouti and piebald factors. The original matings were

all for coupling, the cross being agouti self X non-agouti

piebald (AASS X aass). The F, animals were hetero-

zygous for the two factors (AaSs). Such animals should

form four classes of gametes (1) AS, (2) As, (8) aS,

should be equal in number; if linked, classes (1) and (4)

should be in excess. When such F, animals are back

crossed to the double recessive, non-agouti piebald, the

distribution is obtained which is shown in Table I oppo-

site ‘‘new data.”’

TABLE I

RESULTS OF A BACK CROSS BETWEEN F, AGOUTI SELF MICE (FROM AGOUTI

SELF, AS, X Non-AGoutr PIEBALD, as) AND NON-AGOUTI PIEBALD

as | as ela ma oo

| | f

Dunn's data....... Obs. |221 |179 |183 | 200 783 | 46.23 + 1.20

Exp. | 195.75 | 195.75 | 195.75 | 195.75

New data......... Obs. ites |102 |110 | 108 432 | 49.07 + 1.61

Exp. |108 |108 108 108

1,215 | 47.25 + 0.96.

tass i381 293 |308

Exp. | 303.75 | 303.75 303.75 | 303.75

No. 640] DATA CONCERNING LINKAGE IN MICE 415

The distribution of the new and more critical data is

excellent and excludes any interpretation of linkage be-

tween the two genes A and s. The cross-over classes,

consisting of agouti piebald and non-agouti self, together

number 212. This is 49.07 + 1.61 per cent. of the total

number of young raised, viz., 432, and the deviation is

less than one per cent. from the cross-over value ex-

pected in independent assortment, which result is well

within the probable error, + 1.61.

Combining Dunn’s data with the new data gives a

larger number of animals on which to base conclusions.

Also, any slight. deviation from the normal independent

segregation due to random sampling will tend to disap-

pear when larger numbers are involved. On the other

hand, a small excess of the non-cross-over class, if it con-

sistently appears in both sets of data, must be considered

significant, should it be more than three times the prob-

able error. The combined data are shown in the lower

lines of Table I. Here the cross-over classes consist of

agouti piebald and non-agouti self, and include 574 indi-

viduals. This is 47.25 per cent. of the 1,215 animals

raised, and has a probable error of + 0.96.. The devia-

tion of the cross-over value from the 50 per cent. value

expected in independent assortment is 2.75 a figure within

the range of three times the probable error (+ 2.88).

The new data and the combined data agree in showing

that the two genes are independent.

Tur RELATION OF BLACK-EYED-WHITE SPOTTING TO AGoUTI

In this section will be presented further data on the

relation of black-eyed-white spotting to agouti. It is to

be expected that these genes should show no linkage.

Little (1912 and 1917) demonstrated the independence

of yellow in relation to black-eyed-white. Since yellow,

agouti, light-bellied agouti, and non-agouti are multiple

allelomorphs, black-eyed-white and agouti should show

the same relation to each other as black-eyed-white and

yellow. Yet no direct crosses had been made to test the

416 THE AMERICAN NATURALIST [Vou. LV

point prior to the beginning of this work, and the data

obtained will, therefore, have value.

Reciprocal matings were made. In the coupling series

agouti black-eyed-whites (AAWwss) were crossed with

non-agouti piebalds (aawwss). Since both parents were

homozygous for piebald (ss), that symbol may be omitted

hereafter. ‘The F, agouti black-eyed-white young (AaWw)

were back-crossed to the double recessive (aaww). The

distribution obtained is shown in the coupling series of

Table II to the right of ‘‘new data.’’ There is an ap-

parent excess of black-eyed-whites and agoutis, perhaps

due to random sampling and inexperienced grading of

some individuals. But this excess is distributed between

both the cross-over and the non-cross-over classes, which

consequently are not affected, the numbers being 176,

186. This gives a cross-over percentage of 48.61 + 1.90.

TABLE II

Back CROSSES BETWEEN F, Aout BLACK-EYED-Wuire MICE ( AaWw) AND

HE DOUBLE RECESSIVE, NON-AGOUTI PIEBALD (aaww)

Test for Data AW Aw aw aw Total bey soa

Coupling ..| Dunn’s data | Obs. 6 22 22 4 114 (38.59 + 3.96

Exp. | 28.5 | 28.5 8.5 |- 28.5

New data.. .| Obs. | 114 90 86 2 362 (48.61 + 1.76

xp 90.5 | 90 90 90.5

Combined ..} Obs. | 150 112 108 106 476 (46.21 + 1.53

Exp. | 119 119 119 119

Repulsion .| Dunn’s data | Obs.!| 77 76 76 77 306 (50.32 + 1.92

| Esp. | 76.5: 76.5'| 765) 76.5

Obs. | 28 27 5 2 112 53.57 + 3.17

Exp 28. 28 28 28

New data...| Obs. 0 9 2 9 230 47.39 + 2.21

Exp 67.6) 87.6 | S5157 57.5

Combined ..| Obs. | 155 172 53 68 648 |49.84 + 1.32

Exp. | 162 162 162 162

Cross-overs Non cross-

overs

Coupling | Dunn’s data | Obs. 258 274 532 48.49 + 1.45

and re- Exp. 266 266

pulsion | New data...) Obs. 285 307 592 (48.14 + 1.38

p. 296 296

combined, Combined ..| Obs. 543 581 1124 48.30 + 1.00

Exp. 562 562 Esa

1 The cross-over and

respectively.

non-cross-over classes were given

as 154 ar 152,

No. 640] DATA CONCERNING LINKAGE IN MICE 417

The deviation from the normal distribution is less than

the probable error and therefore not significant.

In the repulsion series the genes for agouti and black-

eyed-white entered separately into the F, zygote. When

back crossed these agouti black-eyed-white young gave

the distribution recorded in the repulsion series of Table

II opposite ‘‘new data.’’ Here the cross-overs number

109, and the non-cross-overs 121, whereas equal numbers

are expected. This is a fair distribution for the small

numbers raised. Combining the new data on the recipro-

cal crosses gives 285 cross-overs, or 48.14 per cent. of the

592 young raised, with a probable error of + 1.38.

Combining Dunn’s data with the new data, there are

543 and 581 animals in the cross-over and non-cross-over

classes, respectively. The cross-over percentage is 48.30

+1.00. The deviation from the normal distribution is

less than twice the probable error. The data, separately

and combined, clearly show the independence of the two

genes,

Tue RELATION oF BLACK-EYED-WHITE SPOTTING TO

Pink-Eye ;

Previous work by Haldane et al. has shown pink-eye

to be linked with albinism. Detlefsen (1916) described

black-eyed-white mice carrying pink-eye, which resembled

albinos, but the genes segregated in the second genera-

tion. No back-crosses were made in his experiments.

No direct intensive crosses hitherto published have dealt

with the possible linkage relation between these genes.

In a cross made for such a study, the genes for black-

eyed-white spotting and pink-eye entered separately.

Accordingly the F, gametes should give equal numbers

of the parental non-cross-over types (WP and wp), and

of the new cross-over combinations (Wp and wP), if the

genes W and p are independent. If linked, they should,

in the back-cross, show repulsion and the former combi-

nations be in excess of the latter. In Table III oppo-

site ‘‘new data’’ of the repulsion series is shown the dis-

418 THE AMERICAN NATURALIST [Von LV

tribution of the young obtained when F, black-eyed-

whites from the above cross were mated with pink-eyed

piebalds, the double recessives. The cross-over classes

number 134, or 46.68 + 1.99 per cent. of the total 281 ani-

mals raised. The deviation from the distribution ex-

pected is within the probable error and indicates inde-

pendent segregation.

The reciprocal cross was also made in which the two

genes entered together in the same parent. The F,

gametes should again consist of equal numbers of two

classes, WP and wp (the cross-overs), and wP and Wp

(the non-cross-overs), if independent segregation occurs.

If linkage is shown, the latter class should be larger.

The actual distribution is shown in Table III opposite

‘‘new data’’ of the coupling series. The agreement with

expectation is poor due to small numbers (77). But since

the cross-overs exceed the non-cross-overs, it is clear that

no linkage exists. The cross-over per cent. is large

(59.74 + 3.77), but the deviation is still less than three

times the probable error. By combining both coupling

and repulsion results recorded in the new data, a total

of 358 animals is obtained, of which 180 are cross-overs.

This gives a cross-over percentage of 50.27 + 1.78. The

deviation from the normal distribution is 0.27 per cent.,

a figure well within the probable error.

These crosses, as well as Dunn’s, show consistently

the independence of the black-eyed-white and pink-eye

genes. Further evidence may be had by adding together

the cross-over and non-cross-over classes of the two sets

of data as shown in Table III. The cross-overs number

330 and constitute 51.48 + 1.32 per cent. of the total 641

animals raised. The deviation from the expected dis-

tribution is 1.48 and not significant. The number of

young raised does not equal that of other crosses, for

the consistency with which these data, in agreement with

Detlefsen’s observations, show the independence of the

two genes, warranted the discontinuance of further

breeding.

No. 640] DATA CONCERNING LINKAGE IN MICE 419

TABLE III

BACK CROSS BETWEEN F, BLACK-EYED-WHITE, HETEROZYGOUS FOR p AND W,

AND THE DOUBLE RECESSIVE, PINK-EYED PIEBALD (wwpp)

Series Data PW Pw pw pw Total [Cross-over Pet.

Coupling . .| Dunn’s data| Obs. | 20 11 9 19 59 66.10 + 4.14

Exp 14.75 | 14.75 | 14.75) 14.75 |

Obs. | 2f i7 21 21 80 (52.50 + 3.76

* T Bp, 120 20 20 20 |

New data...| Obs. | 19 20 5 it 27 77 59.74 + 3.77

Exp. | 19.25) 19.25) 19.25 | 19.25 |

| Combined ..| Obs. | 60 |48 |41 |67 |216 |63.42 + 2.23

| Exp. | 54 54 54 54 |

Repulsion .| Dunn’s data | Obs. | 16 21 22 23 82 52.43 + 3.70

| Erp. | 20.5 1-205 | 20:5 1- 20.5

Obs. | 18 14 12 18 62 41.90 + 4.22

Exp. 115.6 | 15.0: 16.5} 165

New data...) Obs. | 75 62 72 72 218 |47.68 + 1.99

Exp. | 70.25; 70.25 | 70.25 | 70.25

Combined ..| Obs. |109 97 106 113 425 47.76 + 1.63

Exp. 106.25 106 9R INR OF 106 oR

Coupling Cross-overs Non-cross-

an overs

repulsion| Dunn’s data Obs. 150 133 283 |53.04 + 1.99

Exp. 141.5 141.5

New data...) Obs. 180 178 358 [50.27 + 1.78

xp. 179 179

Combined ..| Obs. 330 311 641 /51.48 + 1.32

Exp. 320.5 320.5

SUMMARY

In agreement with the views of previous investiga-

tors, it is shown by new and conclusive experimental

data that the following pairs of genes in mice assort in-

dependently and are not linked. On the chromosome

hypothesis, the members of each pair are located in dif-

ferent chromosomes: (1) the genes for agouti (A) and

for piebald spotting (s); (2) the genes for agouti: (A)

and for black-eyed-white spotting (W); (3) the genes

for pink-eye (p) and for black-eyed-white spotting (W).

BIBLIOGRAPHY

Castle, W. E ;

1919. Gametic Perey in Yellow Rats. Carnegie Inst. Wash. Publ.,

241: 175-

1920, Linked i in Rabbits. Science, 52: 156-157.

Castle, W. E., and Wright, S.

1915 Two Color Mutations of Rats which Show Partial Coupling.

Science, 42: 193-195.

420 THE AMERICAN NATURALIST [Vou. LV

reen A.

n i Results of Crossing Japanese Waltzing Mice with Albino

Mice. Biometrika, 3: 1-51.’

Detlefsen, J. A.

1916. Pink-eyed-white Mice Carrying the Color Factor. Am. NAT.,

50: 46-49.

Dunn, L. C.

1920. Independent Genes in Mice. Genetics, 5: 344-361.

1920a. Types of White Spotting in Mice. Am. Nat., 54: 465-495.

S., Sprunt, e, N.

Re se ulleditoi in Mice. Jour. cf ba 5: 133-135.

Harv nit E.

1920. à Review of the Chromosome Number in the Metazoa. Jour.

- of Morph., 34: 1-67

Little, C. C.

1912. Yellow £49 m Factors in Mice not Associated. AM. NAT.,

46: 491-—

1915. The ee of Black- er -white Spotting in Mice. Am.

Nat., 49: 727-740.

1917. The Relation 3 Yellow ae Color and Black-Eyed-White

Spotting in Mice. Genetic —445.

1920. Note on the Occurrence of a p a Sex-linked Lethal Factor

7-458

in Mammals Am. Nart., 54: 45

SHORT m AN AUTOSOMAL MUTATION IN

HE HOUSE MOUSE*

CLARA J. LYNCH

INTRODUCTION

ALTHOUGH the house mouse has been one of the favorite

mammals used for the collection of Mendelian data, the

number of known loci falls far short of the number of

chromosomes observed in the germ cells of this species.

Therefore, any addition to the list of Mendelian char-

acters in this form should be a matter of interest to the

geneticist.

For several years, mice with unusually small ears have

been known to exist, but, I believe, no description of the

variation has been reported in the literature. The ster-

ility exhibited by many of the individuals used in the

following experiments has hindered the collection of data

so that the amount is not large, but it has seemed advisable

to put on record the results thus far obtained.

The data are based upon observations made at the

earliest stage at which it was possible to distinguish with

accuracy between the long: and short-eared types. De-

pletion of litters occurring previous to that time was dis-

regarded.

DESCRIPTION OF THE CHARACTER

The mutation was found in stock which originally came

from the Lathrop mouse farm and consists in a noticeable

difference in the size of the ear. The pinna is about one

half as long as that of the normal ear and usually one or

two millimeters less broad but the position in which it is

held lying close to the head makes it appear smaller than

it actually is. In outline it is less regularly curved than

the normal, a flattening near the tip of the ear and one in

1 Received for publication from the Rockefeller Institute for Medical

Research. `

421

422 THE AMERICAN NATURALIST [Vou. LV

short ea

B, F, hybrid bley a cross between a as. eared and a long-eared mouse.

Ears pe TS to nor eg

A. Mouse showing es mutation

the outer margin being fairly constant features. It is

usually thick and rather fleshy in appearance. The dis-

tribution of hair on the surface of the ear is similar to

the normal.

EXPERIMENTAL INVESTIGATION

The Mutant Out-crossed.—The first two experiments

showed immediately that the character ‘‘ short ears’ is

recessive and that it is not sex-linked. Seven crosses

were made between short-eared males and a number of

long-eared females which were taken from sources other

than the Lathrop stock or from Lathrop strains which

No. 640] SHORT EARS IN THE HOUSE MOUSE 423

had never been known to produce any short-eared mice.

In the F, there appeared seventeen young of which ten

were males and seven females (Table 1). The ears of all

TABLE aA

SHORT-EARED MALES CROSSED TO LONG-EARED FEMALES

Mating Fı Long

h oy

vi, AEE A E E os VC EAA an E bo oa 1 1

IE Lito Ss DMN ae art N te Meret abstr See aM 1 1

1309-9 Sor ek oe err ee ee O ENN 2 0

TIGR SB ic iiss SW OR ee ee 2 0

Dee E O a a a ea ee DE C 2 2

TOMI ie ee eee Se Re ee 1 2

pr EAS: AE ret yen ape rc eu at A E rele nn ea a 1 1

Total 10 a

these mice were long—in fact, indistinguishable from the

ears of normal mice. In the reciprocal cross, where the

male parent was long-eared and the female short, the

offspring were again all long-eared and also comprised

members of both sexes. The numbers obtained in the

second case (mating 1790, 1 male and 2 females) are very

small, partly owing to an unusual amount of destruction

of the young which happened to occur in this type of

cross, but they indicate that sex-linkage is not involved.

Were the gene for short ears located in the sex chromo-

some, ‘‘ criss-cross °’ inheritance would result from this

mating and the sons would resemble the female parent

and the daughters would be like the male parent since

the sons would receive their single X chromosome carry-

ing the short-ear gene from their short-eared mother,

while the daughters would receive one X containing the

short-ear gene from the mother, and the other X carrying

the dominant long-ear gene from the father. This event

was not realized—both sexes were of the same type,

showing that both kinds of sperm formed by the long-

eared father, whether they were male-producing or female-

producing, carried the normal allelomorph of the gene

for short ears and determined the appearance of the

424 THE AMERICAN NATURALIST _ [Vou LV

long-eared F, individuals. Since it occurred in both

types of sperm the short-ear gene must be located in one

of the autosomes.

The Back-cross—The back-cross between long-eared

mice, heterozygous for short ears and short-eared mice,

gave 51 long-eared animals to 43 with short ears (the

sums of the figures in Tables II and III). On the assump-

TABLE II

Back-cross. HETEROZYGOUS P MALES CROSSED TO SHORT-EARED FEMALES

Long Short

Mating :

| | Sex Not Sex Not

| oo! | 92 | Recorded Joo! $8 Recorded

ire. eee 1 3|

t612- HP oann e ie a PU AE pee 3 1

Tere emote he 2 ee Ee 2 1

TODA P ee eee eats | 3

1656-B 3 |

T Es E A E A E ek s E eet N tis 3

ROBO aea aa ae Poteet a Re eile: yates E RE S 1

1606-8 te 1 1

1656-RE aaa She eae 3 |

1006 cc eee 3 1 |

1000. iaa a a a A I a eC E e eer e r a a 3

10 ri 6 8 2 4

Total ei in | 23 | 14

tion that the character is due to one gene, the expected

back-cross ratio is 1:1. In a total of 94, the expectation

for the two classes would be 47 :47. The actual numbers

obtained fit the calculated sufficiently well to justify the

conclusion that short ears depend upon a single pair of

genes.

In conformity with the results of the first test, the back:

cross also shows that there is no sex-linkage concerned

in the transmission of the new character. The long-eared

F, males (obtained by crossing normal females with

short-eared males) were bred to short-eared females. If

the new gene were sex-linked, the F, male would have but

e ‘‘dose’’ of the allelomorph long ears. It would be -

No. 640] SHORT EARS IN THE HOUSE MOUSE 425

TABLE III

Back-cross. HETEROZYGOUS F, FEMALES CROSSED TO SHORT-EARED MALES

Long Short

Mating AES

Sex Not | Sex Not

F | FF | Recorded | Po | 92 | Recorded

nL n RE E E EES A aa le N A aD 2 S

LOOD- os aN ans 1 Bi Naa oc Cae Aaa a 2

BGO ais E A CaN Se he has Be ees eet tee 1

Sas es ea Se 1 ee ewan 1 3

105a -i e G 2 n RE TEE P es i

“af oes Pete mode EE a he 1 Me cae ET A Py ine. 2 1

SGODRI 2 i ee 1 Beret be a ate 1

C9 Spa eet aap ERR ASN SO ES ae na yi SE ea) aon OR ern Ea 3

pA Ay See Pera ae L Bay Ne EEN 1 1 1

UE L e AE E E N E: EES ae See E E E T 1

PO ee ote 1 1 1 1 1 2

11 14 3 | 10 13 6

To a o a 28 | 29

carried by the single X chromosome and distributed only

to his daughters, which would have long ears. His sons

would not receive it; therefore they all would be short-

eared like the maternal parent.

The offspring from this cross are listed in Table IT.

In a few cases the mice escaped or were destroyed by the

parents before the sex was recorded. There were, in ad-

TABLE IV

HETEROZYGOUS LONG-EARED MALES By HETEROZYGOUS LONG-EARED FEMALES

Mating Long Short

OD aea u aa

REDO o N Use eee bt owe cb cv unser 2 4

Bi he a eae Cae awa cas Serer 3

ee os ve i be een yea ee enon ees 2 1

Total 11 B

TABLE V

SHORT-EARED MALES BY SHORT-EARED FEMALES

Mating . Short Gt Short 9 Q

rar

slaak

o| w to

426 THE AMERICAN NATURALIST [Vor. LV

dition, ten males and seven females which had long ears

and eight males and two females with short ears. The

appearance of individuals of each sex in each class shows

that the gene must not be carried by the X chromosome.

The First Filial Generation Inbred.—A very small

number of mice were obtained from inbreeding the long-

eared heterozygous F, offspring. These gave, in the F,,

11 long to 6 short (Table IV). In a total of 17 individ-

uals on a one-factor basis, the numbers calculated for a

3:1 ratio would be 12.75 to 4.25. The results are con-

sistent with our previous conclusion as to the number of

genes involved.

The Inbred Recessiwe.—The test of the inbred reces-

sive demonstrates that the character breeds true. Three

matings between short-eared males and females have

yielded 13 young, all with short ears (Table V).

CONCLUSION

The data given above show that the mutation ‘‘ short

ears’’ which appears as a perfectly definite and easily

distinguishable character in mice behaves as a recessive

and is dependent upon a single gene which is not sex-

linked. Other possible linkage relationships have not yet

been worked out.

VARIATION AND HEREDITY IN LUPINUS

DR. LEONAS L. BURLINGAME

STANFORD University, Cat.

Tue great majority of genetic investigations have been

carried on with domesticated plants that have been long

under cultivation. There are, of course, many entirely

good reasons for this fact. It has, however, always

seemed to the author that the principles of genetics

might, perhaps, be more successfully approached by be-

ginning with wild plants. For this reason he has been

bringing into the garden during the past several years a

considerable number of wild species which 50 se mee to

offer desirable material for experimentation.

His attention was first directed to Lupinus through the

difficulties that appeared to beset the systematist in arriv-

ing at a satisfactory classification of the species. It

seemed possible that the difficulty might be due to the fact

that the species are of recent origin and, so, very close to-

gether. Under the circumstances that a genus is either

now or has recently been in a mutable state it would

seem likely that natural selection would not yet have had

time to weed out all the forms doomed to perish through

lack of adaptation or insufficient vigor. |

‘While discussing this question one of his colleagues

mentioned the fact that the form recently (1911) de-

scribed under the name of Lupinus pipersmithii Heller

(5) grows near and that it sometimes has pink flowers.

An investigation during the spring of 1914 not only con-

firmed this statement, but also showed that pink flowers

are not uncommon in two closely related species—L. val-

licola apricus (Greene) (4) and L. nanus Dougl. (7).

These three species are all close to one another but ap-

parently satisfactorily separable by the systematist.

427

428

THE AMERICAN NATURALIST

[Von. LV

This will be clear from the parallel descriptions given

below. That each of the species may produce variant

types with some characters of the others will appear in

our further discussion.

L. pipersmithi Heller

Annual

Somewhat Ratan

below with white hairs,

densely so above, espe-

cially so in the inflo-

rescences, `

Branches several from

the base, m. high,

st colored, rather

idged.

rowed and a

ingly appressed pubes-

cent above, e be-

neath, light green in

or.

Stipul enlists

green, 6-7 . lon

Peduncles z em. lo ong,

or less, wah ese ior

about the len =

Flowers Ti 4

whorls, e PA A ex-

cept permos t

longer finn the flowers

Bracts arbre 4 mm.

long, caduc

Pedicels mm. long,

slender, villous with short

white hair

Calyx ae villous but

parted for nearly 3 m

, . long,

2 mm. wide, 3-toothed at

the apex.

m Le

phita Sgir

L. vallicola

(Greene

Annual

Rather shortly and

sparsely pilose.

apricus

)

Branches few to

several, almost upright,

firm rom near the

base, 30 to 45 em. high.

aflets about 7,

abou

thin, sparsely aie

- pubes

Racemes sub-sessile,

pa ig em. long in fu i

flow

Pow 5

eitia cet

‘Seas

indistinct.

Calyx—broad upper

lip deeply cleft, the

ng, lower entire, eae ” ole

exterior appr ed

villous but not patie

L. nanus Dougl.

Annual

or finely

Villous

pubescent.

Slender, not succu-

lent, often branching

eT near the base, 15

o 37 cm. high.

Petio ied 14 to 3 times

longer than leaflets

Leaflets 7 or 8, linear

- oblanceolate, "1.2 to

2.5 em. long, usually

acute.

Raceme loose, short

oo 7.5: to 17.5

. long.

powe rs fragrant and

in several distinct or

somewhat indistinct

whorls,

Bracts omane the

calyx, deciduo

Pedicels EAT 6 mm.

long.

Upper lip deeply

times obscure or want-

ing.

No. 640] VARIATION

Corolla ea ran

blue, 7 mm.

deep, Tesine arrori

apice of wings and

Bann f

shorter than the wings,

above, obovate- cuneate

when spread out, 6 mm.

wide, with a shallow me-

turni id

pex

expo the keel, the

connate cone i o $ ed,

raise har p

eep

middle, the

Pods Sraa A villous,

2.5 em. long, 4 mm. wide,

sour. f- seeded,

Seeds pale flesh-colored,

slightly yellow-brown

mottled, 2 mm. long, 1.5

mm, wide.

AND HEREDITY IN LUPINUS

olla

(b1 ne)

Seely aly than aig he mm

breadth about 10 mm

with white

; d

sulcus, turnin ng reddish

purple in

ers wy onda

bea the

arte at (Smith).

Pod 2.5 em. long,

appressed villous, 3 to

7-seeded.

Seeds not cake a

compressed, obliquely

round-oval, grayish with

a few markings an

many sma ots

minent ‘dark

grayish brown. (Smith)

429

Corolla blue, 10-12

. long.

with

cular

the sides reflexed.

ican lightly a

ing an = ame ly

petra inflated s

Keel faleate, ciliate

from above the middie.

The above descriptions have been taken from the origi-

nal sources for L. pipersmithii (5) and L. vallicola apri-

cus (4) and from Jepson’s Flora of Middle Western Cali-

The three species are illustrated in

the accompanying text-figures.

fornia for L. nanus.

Field S

tudy of Variation.—Field work can be carried

on profitably only in years when lupines are abundant.

1914 was a very favorable year for L. vallicola apricus

and 1919 and 1920 for L. nanus.

430 THE AMERICAN NATURALIST [Vor. LV

In 1914 apricus and pipersmithiw were very numerous

in a field near the university and much study was devoted

to them. Besides the pink forms already mentioned a

considerable number of other color variants were noted

and seed collected of some of them. Seeds were collected

of both the pink and blue forms of pipersmithii and have

since been cultivated. Smith (10) had earlier collected

seeds of these varieties, but had not been successful in

cultivating them. Jepson (7) makes note of the variabil-

Fie. 1. Lupinus vallicola apricus, showing the flower and its parts.

(After C. V. P. Smith.)

ity of nanus and Smith mentions variations of size of

plant and flower in apricus.

The normal color of apricus is dark blue and white.

After some study it became very obvious that, mixed with

the usual types, there were occasional plants with a de-

cidedly lighter blue. One variation consists in a mere

lightening of the blue. This type was called light blue

No. 640] VARIATION AND HEREDITY IN LUPINUS 481

and plants were bagged for seed collection. Another type

resembled this in color except that both banner and wings

were distinctly striped. The amount of blue and white

varies somewhat but the type is always recognizable. In

some cases one might correctly speak of a light blue flower

is Ky

La T i Le. i

\ RS ER

> A

Bo en.

eo,

ee,

Lay Ae

Sea

;

Fic, 2. Lupinus piper-smithii, showing floral parts. (After C. V. P. Smith.)

ste 3. Lupinus nanus. 1. Normal plant. 2.- Single flower, front view. 3.

wn on flower showing particularly the banner shape, 4. Sid

of mek flower, showing sca of hat spot. 5. Front view with wings re-

St Erd position of tal groove and spots. 6a. Pod of normal plant.

6b. “ curly pod,” pen 11 in tabis T. Seed from Pedigree V, showing

432 THE AMERICAN NATURALIST [Vor. LV

with white stripes or veining. In other, and rarer, cases

it is apparently a white flower veined with blue. Seeds

of this type were collected and have since been grown

under the name of striped white. This first season it

was not realized that the two lighter forms were distinct.

Fortunately the seeds of each individual plant were saved

and grown separately, so that the distinction became evi-

dent in 1917, the first year in which successful cultures

were grown.

Pink apricus were also fairly common that year and

have been occasionally noted since. Seeds were saved

and have since been grown.

Two white forms were observed that year. One ap-

peared to be plain white. This type has been frequently

observed since in this and other species. A single plant

was also observed in which the color was a cream or

yellowish white. Seeds of both sorts of plants were saved

but have so far failed to grow.

A few scattering observations were made on nanus

from year to year, but it was not until 1917 that a field

was found in which variations were numerous and con-

spicuous. Pinks, light blues, striped whites, and whites

were all abundant in a field near the town of Woodside.

Specimens of all these types were staked out for seed but

unfortunately attracted the attention of the small boy

and no seeds were secured. In 1919 3 fields were found

some 3 miles west of the university in which numerous

variations were in full bloom. A careful study was made

and an attempt made to discover and describe every

variation. Twenty-seven plants were labeled and de-

scribed for seed collection. They represented 12 separate

and distinct variations or combinations. All the varia-

tions but two are concerned with flower color. In the

description of nanus it will be noted that the normal

plant has a blue and white corolla having a reflexed

banner with a white front down the middle of which there

is a groove marked with a number of small dark spots.

In the variants the normal dark blue of the banner and

No. 640] VARIATION AND HEREDITY IN LUPINUS 4833

wings may be replaced by the light blue or striped white

described for apricus or the whole corolla may be white

or pink and white.

All the dark blue, light blue and striped white plants

had a white front on the banner with blue, blue-green,

or no spots in the groove. White corollas had either

white or lemon-yellow fronts with or without the dark

spots. Normal dark blue plants have the color uniform

on banner and wings or else with a deeper spot at the

angle where the reflexed banner joins the wings. In light

blues this spot may be dark blue, light blue, greenish blue

or absent altogether. Whites have this spot faintly blue,

bright orange, or absent. The orange spots may occur

with either white or lemon-yellow fronts.

A reddish-purple corolla was found on a number of

plants which were distinguished by unusually large and

fine flower clusters. In some of them the flowers had

fallen and pods formed on some earlier branches. The

pods in every case were of a very peculiar type. Instead

of the elongated, nearly straight and somewhat flattened

and constricted type commonly found these plants had

pods long, tapering, nearly round, and frequently much

coiled. This was at first supposed to be due to attacks

of insects or fungi. Diligent search failed, however, to.

show any such infection or infestation. And when it

was later discovered that the plants which would pro-

duce these curly pods could be identified in the early

flowering stage it was realized that this whole group of

characters constituted an interesting variation. A num-

ber of plants were labelled and some bagged for seed.

Although many seed capsules were examined, no good

seed were secured. It appears highly probable that this

form is completely sterile, though there is a chance that

the failure to set seed may have been due to un-

favorable weather. In 1920 neither this nor any of the

other variations were to be found in that same field.

Diligent search, in fact, has failed to discover it any-

where. It is much to be regretted that good seed of this

434 THE AMERICAN NATURALIST [Vou. LV

form can not be secured, inasmuch as the somatic varia-

tions involved appear to be more profound and com-

plicated than in any of those so far studied.

The only other form variations, aside from mere size,

so far observed are two modifications of the banner. In

one plant of nanus in 1919 the banner was hooded in the

fashion of some varieties of sweet peas. This same varia-

tion had been previously noted in apricus in 1914. Apri-

cus also occasionally yields plants with the abbreviated

banner characteristic of pipersmithui. So far seeds of

neither of these variations have been collected, though

attempts have been made.

The table below summarizes the variations observed in

1919, This season’s study yielded no new types and

failed of some of those found the year before.

Color and Markings Front Lateral

Number | Other

Corolla Veining Front Spots Spots Characters

s S Light blue| Veined White I Dark blue Lisco e Tanala

e PE eas ANO PAE eter Bes bi A Dark blue | Dwarf

So ee Ore Se cee Os Mf ise Famt blue hea So

Soy tex: A EE begets ear ama r? Greenish | Greenish |;.......4<

5. care: Witte oto or Demons fe os saa Paint blue}. 6.53 o.5

Boar eh eee ae ey PROW LEs hi os e eno so ies eee

TEO T Sees SP E ainat ei White Darkblue) Orange ...... a ee

are ee al, Me ee See ROMO eae erate ried cee, yee ea

yellow

em cig Pile Spoor "WMG OT) onI Faint blue | Dwarf size

ji SAFRA i E S OE a e S lang = Dark blue Bisse ke Vue se

Ps oe, Reddish: foc ek abe E IRD onlOr pete mee Curly pods

rple ‘

Bou Reddish oi i a Blue Hooded

purple | banner

The variations in size of all three species are very con-

siderable, no doubt due in large part in the field to differ-

ences of soil, moisture, shade and exposure. When

grown in the garden side by side nanus is the largest and

most vigorous and pipersmithii the smallest. In the

field nanus is often smaller than apricus, though the

flower cluster is usually much larger. This difference is

possibly due to the fact that it grows in more open, ex-

No. 640] VARIATION AND HEREDITY IN LUPINUS 435

posed places and usually in poorer soil. No cultures of

different-sized plants of the same species have yet been

grown and critically compared. Some casual observa-

tions, however, indicate that there are probably heritable

size differences in apricus.

Reference to the specific descriptions will show that

some taxonomists have been inclined to base classifica-

tion in part on the color, shape and markings of the seeds.

Our observations would indicate that this is not a very

safe criterion, particularly in reference to color and

markings. Seeds of nanus are the largest. They could

usually be picked out of a mixture with apricus by size

and shape, but not by the color or markings. No char-

acter is more subject to variation than this. This does

not mean that the character is a fluctuating one, but that

there are a great many differently colored and marked

seed varieties. All the seeds of any one plant are alike

(with certain exceptions to be noted in a later para-

graph), as one would anticipate since the color and mark-

ings are seed-coat characters and hence genetically all

alike on any particular plant.

Dark blue nanus and apricus plants have dark seeds,

whites and pinks have light seeds, light blues and striped

whites have seeds of intermediate color. Pinks and light

blues have a yellowish-brown tone and are flattened and

without other markings. Striped whites are longer and

thicker in proportion to the width, of either a bluish or

yellowish tone with a conspicuous crooked line on either

side forming a continuous ring about the shorter. perim-

eter of the seed. This ring does not occur on the seeds

of pinks or whites. It is present on some races of dark

blues and absent on others. The statements just made

apply particularly to those races which have been culti-

- vated for some time in the experiment garden. Field

observation confirms them in part, but reveals a consid-

erable number of other variations in color and marking

which have not yet been cultivated and about the genetic

behavior of which nothing is yet known. The only case

436 THE AMERICAN NATURALIST (Von LV

so far noted in which the seeds of a single plant of apricus

are not entirely uniform is that of two tones in a striped

white variety. An attempt has been made to cultivate

the two sorts separately and to determine the ratio in

which they occur. Cultures have not yet proved success-

ful. In some collections of seed the ratio approximates

a 3:1, but in others it is nearer 8:1, but whether any

significance is to be attached to these results has not yet

been definitely determined.

It is a curious fact that, although lupines yield an

enormous number of seeds and the plants often literally

cover acres of space they are nevertheless very capricious

in their occurrence. Some illustrations of this may be

taken at random from our notes. The field near the

university where the apricus mutants and the pink piper-

smithii were collected in 1914 has been under observation

every year since and has not at any time had many plants

or produced a second display of these mutants. A num-

ber of pink forms of both species were staked for seed

that first year and the stakes left in the ground to mark

the site, in the expectation that the same forms would

reappear the following year. In no single case were pink

flowers found at any of these stations, although they were

found the next year at other locations in the same field.

The field near Woodside studied in 1917 has been visited

each year since. Not only have no mutant forms been

found there, but there have been exceedingly few normal

ones. The field from which the notes were made for the

table on a preceding page was very thoroughly searched

again this spring. There were a few dark blue plants, but

not a single one of the types which were more or less

abundant there last year. These vagaries of distribu-

tion doubtless depend in some manner not yet clear on

the difficulty of germinating the seeds.

Pollination and Seed Collection.—In 1914 when these

observations were begun it was assumed that the lupines

were probably frequently cross-pollinated, inasmuch as

they appeared to be freely visited by bees. It was a

No. 640] VARIATION AND HEREDITY IN LUPINUS 437

matter of some surprise to find that the forms of apricus

and pipersmithii which were brought into cultivation

did not indicate this to be true. The pinks bred true in

both species. Experiments to determine self-pollination

by bagging or screening the plants with fine-meshed wire

cages showed no diminution in the harvest. Further-

more, in the following years different strains grown in

adjacent rows showed no sign of crossing. It was then

assumed that the same would be true of nanus and flower

clusters were bagged in the field for seeds—not to insure

selfing, but merely to prevent the seeds from being scat-

tered by the explosive dehiscence of the pods. The re-

sults were wholly negative, resulting in a failure to

Secure any seed that year. The appearance of the

contents of the paper bags first used led to the supposi-

tion that possibly the failure to set seed was due to the

bags. Careful experiments were therefore made the

following season by inclosing whole plants in cages of

fine screen wire or cheesecloth. In no ease did this result

in setting seed. It appears, therefore, that nanus is de--

pendent on bees for pollination. On the other hand, cul-

tures derived from white-flowered nanus have shown that

the bees act in part merely as a mechanical agent, for

part of the progeny was white and part blue. These

results agree well with and serve to explain the greater

number of variations found in nanus.

Pollen Sterility.—It has been maintained by a number

of authors at one time or another that variability in

nature is very largely a matter of hybridity and that

sterile pollen is a more or less certain indication of the

hybrid nature of a species (6). Having found two closely

related species both variable and in a closely similar

manner, it became a matter of interest to study the com-

parative sterility of close-pollinated and cross-pollinated

species. In order to determine this matter a large num-

ber of plants belonging to all the varieties in cultivation

at the time in the garden were examined. Mounts of

pollen from three different flowers from each plant were

438 THE AMERICAN NATURALIST (Vou. LV

made. Each slide was so prepared that about 100 pollen

grains would be visible in a single field of the microscope.

All the grains in the field were then counted and the

percentage of sterile grains caleulated and averaged.

Only two plants showed more than 3} per cent. of sterile

pollen. One normal dark blue plant showed 40 per cent.,

39 per cent. and 0 per cent. in three flowers. One striped

white showed 9 per cent., 6 per cent. and 0 per cent. in

three counts. One isbiid hight blue, one dark blue and

one pink showed no infertile pollen. The one plant with

a high degree of sterility was a selfed dark blue apricus.

Only a few plants of nanus were available, but they

showed no poorer pollen than apricus and pipersmithi.

Seed Germination—In a previous paragraph it has

been pointed out that, although lupines produce immense

quantities of seeds, field germination is apparently poor

—at least an abundant seed harvest is likely to be fol-

lowed by a poor stand of plants the following season:

Our earlier attempts to grow them in the garden were

practically a total failure.

The first attempt to grow controlled cultures was made

in the winter and spring of 1914-715. Four hundred

seeds were planted in pots in ordinary unsterilized

garden soil and 400 more in pots of sterilized soil (ster-

ilized in an Arnold steam sterilizer). The seeds them-

selves were planted dry without treatment of any sort.

The results were almost a total failure. About 5 per

cent. of the seeds produced seedlings. Of these all but

5 or 6 were killed as seedlings by the attacks of soil fung:

or just simply died. Two plants lived to flower, but failed

to set seed. It appeared evident that greenhouse cultures

under the conditions then available were likely to be un-

profitable and the remainder of the original seed collec-

tions were held over until an experiment garden could be

secured. `

In 1917 the remaining seeds were planted without treat-

ment in the open garden. Nine plants in all lived to

mature seed—one light blue, one striped white, one pink

No. 640] VARIATION AND HEREDITY IN LUPINUS 439

pipersmithii, three pink apricus, and three dark blues.

The seeds since used have all been derived from these

nine plants, each of which has been assigned a pedigree

number.

In 1918 seeds of each of the 1917 plants were again |

planted. Although about 400 seeds of each were planted

in each culture, two pedigrees failed to yield a single plant

that year, though a few plants came up the following

spring. Pedigree VIII, Light Blue, yielded a culture of

85 plants and pedigree V, Striped White, 28 plants. Thus

the highest per cent. of germination did not exceed about

20 per cent., ranging from that down to zero.

In 1919 a series of germination tests were carried out.

This had not been possible before on account of the small

number of seeds available. A considerable variety of

methods were employed. Soaking the seeds in tap and

distilled water for weeks is of little value, owing to

the failure of the seed to imbibe water. Breaking or

cutting the seed coat brings about prompt imbibition of

water and consequent swelling, but does not produce a

high percentage of germination. Seeds were soaked in

water under air pressures up to 140 pounds to the square

inch with no noticeable effect.

Since the seeds are small cutting. or filing the saoi

coats is a very arduous affair where cultures of any size

are to be grown. Attempts were, therefore, made to find

some other means of bringing about the same result. The

seed coat appears to be difficult to wet and this was

thought to be due to the presence of some oily or waxy

constituent. Attempts to dissolve this by means of KOH,

various percentages of alcohol, ether, etc., proved en-

tirely unsuccessful.

Soaking in concentrated H.SO, proved the most effica-

cious method tried. This was applied in parallel series

of seeds from the same plant for periods from 5 minutes

up to 2 hours. The shorter treatments seemed to produce

“no effect at all. The longer periods of two hours or

over killed the embryos. After some experimentation it

440 THE AMERICAN NATURALIST [Vou. LV

was found that a treatment of one and one half hours

would sufficiently char the seed coat to secure practically

100 per cent. of swelling (8). At first it seemed im-

possible to say whether this might not also injure the

embryos. Later it was discovered that whenever any

acid penetrated the cotyledons of dark blue plants they

turned pink. This color is probably due to the presence

of a chromogen similar to or identical with the one which

eventually produces the pink or blue flower pigments.

Since this color reaction is brought about by even faintly

acid solutions, it was thought that it would serve as an

effective check against overtreatment with the acid.

After the preliminary experiments had shown the sul-

phurie acid method to be the best at our disposal a com-

plete series of tests were run on each of the 70 pedigrees

available for planting at that time. The dry seeds

were placed in the concentrated commercial acid and left

for 90 minutes. They were then washed rapidly through

several changes of sterile water until the water failed

to affect litmus paper after the seeds had stood in it for

20 to 30 minutes. It was found necessary to carry out

the washing rapidly since the weak acid readily pene-

trated into the cotyledons.

After washing the seeds were subject to one of three

treatments. Some were left in sterile water until

sprouted. Others were removed as soon as they had

swelled. The great majority of the pedigrees swelled

within 18 hours. A few were completely swelled within 4

hours. In most cases they were left about 18 hours. A

third method was to transfer the seeds as soon as washed

to a nutrient solution. This did not show any advantage

over plain sterilized water. The seeds which were re-

moved from the water after washing and swelling were

placed on or between moist blotting papers. It was soon

found that those left in the water or placed between

papers kept wet and soggy excelled those which were

placed on papers merely kept moistened. Twenty-one

lots of seed failed to sprout at all, although the percent-

No. 640] VARIATION AND HEREDITY IN LUPINUS 441

age of swelling was 95 per cent. or better except in two

cases, each of which had only 40 per cent. of the seeds

swelled. The remaining lots averaged 51 per cent. of

sprouted seeds. Two gave 100 per cent., one gave 95

per cent. and 14 were below 25 per cent.

From these results it appears that many seeds which

show no observable defects are nevertheless either dead

or in a state of dormancy not readily overcome. That

the latter is probably the true explanation is indicated

by the fact that more seeds treated in this manner actu-

ally sprout than when the seed coats are mechanically

ruptured. The treatment with the acid seems to act in

some manner, possibly by dehydration, as a slight stimu-

lant to sprouting. It is not certain, however, that a

larger percentage of viable seedlings is actually produced.

Many seeds put forth the radicle in an apparently normal

manner, but do not continue growth. Others die at later

seedling stages apparently from internal causes, for they

have not had opportunity of infection and are growing

under the same conditions as others in the same culture.

This spring and winter a number of cultures were tried

in which the seed were treated with acid, washed, and

then planted directly in the soil out of doors. In every

case a good stand was secured averaging about 50 per

cent. of the seeds planted. The seedlings of the preceding

season were planted in pots and kept in the greenhouse

until a vigorous young plant was secured, and then trans-

planted to the garden. They did very poorly after being

transplanted and produced practically no seed. It is

uncertain whether the failure was due to faulty technique

in transplanting or to the failure of the plant to adapt

itself to the change of environment. It is not unlikely

that both causes had something to do with the matter.

They had been grown in 5-in. pots and transplanted with

the whole mass of dirt, but even in that way some dis-

turbance of the root system was unavoidable. In addi-

tion to this there was a very hot, dry wind lasting three

days about flowering time. This seemed to do a lot of

442 THE AMERICAN. NATURALIST [Vor. LV

damage to plants in the field and was no doubt highly in-

jurious to those in the garden as well. It is planned to

repeat the experiment again, using paper pots which may

be set in the garden without disturbing the roots. at all

and without interfering with subsequent growth. (These

pots have no bottoms:

Garden Cultures have now been carried through four

seasons in some'lines and through two or three in the

others. Pink apricus and pipersmithti have proved en-

tirely constant, with one exception. Dark blue apricus

also breeds true.

The striped whites of Pedigree V for three seasons

produced both striped whites and dark blues.. In 1918

the culture produced 28 plants which flowered. There

were 20 striped whites and 8 dark blues. In 1919

the cultures were so badly injured by transplanting and

unfavorable weather that little reliance could be placed

. on numbers. Some cultures, however, did produce white

. plants. In the larger, but still unsatisfactory, ones this

spring whites have again been produced. Although the

numbers are very small, the fact that striped whites give

rise to dark blues, striped whites, and some whites is

significant. Two cultures of 100 seeds each this spring

produced two white plants each and no other sorts.

From these data, unsatisfactory as they are, it is probable

that striped whitgs are hybrids between white and dark

blue and that white differs from dark blue by a single

factor. : |

The light blues of Pedigree VIII have also been grown

through four seasons. In 1918 the original plant pro-

duced a progeny of 85 plants, of which 20 were dark blue,

60 light blue, and 5 failed to flower. They have never

produced any whites. This season 8 cultures out

of 27 produced only light-blue plants. In most cases

_the numbers were small, but one culture of 200 seeds pro-

duced 32 light blues and one plant which did not flower.

Three others respectively produced 22 light blues out of

200 seeds, 18 light blues out of 170 seeds, and 9 light

No. 640] VARIATION AND HEREDITY IN LUPINUS 448

blues out of 150 seeds. Taking into account that four of

the eight cultures had only one or two plants and so

might have produced dark blues also, this 8 : 27 is prob-

ably as close an approximation to a 3:1 ratio as could

be expected. The data now available would appear to

justify the conclusion that the light-blue color is due to

a single factor difference and that heterozygotes and ho-

mozygotes are phenotypically indistinguishable.

Owing to the fact that a single cross yields only five

or six seeds, it has not seemed profitable to attempt hy-

bridizing the light blues and striped whites either with

one another or with other forms, until a techinque has

been perfected that insures a higher percentage of ger-

mination. In nanus, which is naturally crossed, certain

observations have been made which indicate something

of the relations of certain factors to one another. Seeds

of a white-flowered plant collected last year from un-

protected flowers were grown this year and produced

both whites and blues apparently like normal wild ones.

- This would indicate that the mother plant had been vis-

ited by a bee which had effected pollination in part with

its own pollen and in part with that of a neighboring

dark blue. If this be the true explanation this white was

a recessive one. It might have been a heterozygous domi-

nant white, of course, which would be in agreement with

the nature of the whites in apricus, Pedigree V. Seeds

of pink nanus collected last year yielded only a half

dozen plants, all dark blue. This would point to the con-

clusion that pink is also recessive to blue. However,

these results are too meager to have more than a sug-

gestive value.

Seed characters have also proved constant in inherit-

ance. Dark-blue apricus plants invariably have dark

seeds, but the particular type of marking differs accord-

ing to the origin. It is suspected that the two types of

seeds found in Pedigree V, striped whites, indicate a

genetic difference, possibly distinguish striped whites

444 THE AMERICAN NATURALIST [Vor. LV

from pure whites, but the facts now known do not suffice

to prove this.

Mutations have apparently occurred in culture in re-

spect to both flower color and seed-coat markings.

This spring (1920) a single dark blue appeared in a cul-

ture of pink apricus which had bred true through the three

preceding generations. In 1918 a pink arose in a culture

of. seeds from a wild dark-blue nanus. If the relations

between pink and dark blue are the same in the two

species one of these cases must be a mutation. Hybridi-

zation is exceedingly improbable in the case of the pink

apricus. It is not likely that the dark-blue nanus was a

hybrid either since it was not collected from a location

where this cross would have been likely to occur and

only one plant out of a large culture was pink.

In 1918 plant VITI-27 with dark-blue flowers produced

seed of the light color characteristic of light blues. These

seeds this year produced both light blues and dark blues.

In collecting these seeds it was necessary to read the

label on the plant and that on the seed box. It is very

unlikely that they failed to tally with each other or with

the color of the flowers still in bloom on the plant. Sev-

eral collections were made over a period of two or three

weeks and the plant label put in the box at the final

collection. Owing to the fact that the seed pods were

not opened until after the collections had been finished,

no suspicion of anything unusual was entertained until

after any sort of check was no longer possible. A mis-

take might have been made several times in succession,

but this is certainly very improbable. The seeds this year

are of the usual type. Three light-blue plants produced

light seeds and two dark blues produced no seeds.

If this is not a case of mistake in records it is very

difficult to offer any explanation of it. Since plant VIII

was a hybrid and the factor for dark-blue flower color

is linked with that for dark seeds, plant VIII-27 could

have arisen either through a mutation in the recessive

factor for coat color or by a cross-over of its dominant

No. 640] VARIATION AND HEREDITY IN LUPINUS 445

allelomorph, so that dark flowers would then be linked

with light seeds. In either event plant VIII-27 would

be homozygous for dark flowers. It might have been

either homozygous or heterozygous for seed-coat color.

In the one case it would yield a progeny with dark-blue

flowers and light seed coats. In the other all plants

would have dark-blue flowers, but there would be three

light seed coats to one dark one. The results, however,

are both light-blue and dark-blue flowers.

Pedigree VII is dark blue and has bred true for three

seasons. This spring culture VII-3-1 produced a single

white plant. It was the only plant from 77 seeds. The

parent plant was expepronat in that its seeds were lighter

in color than usual.

Discussion.—It is realized that the facts presented in

the preceding pages are regrettably incomplete. It is

hoped, however, that they are sufficient to interest others

in lupines as suitable materials for genetic investigation.

The striking parallelism in the mutations occurring in

the two species, apricus and nanus, is certainly a signifi-

cant phenomenon. All recent work with both plants and

animals proves that in varietal crosses homologous chro-

mosomes are freely interchangeable and that allelomor-

phie factors oceupy identical loci in their respective

chromosomes. The work with multiple allelomorphic

systems clearly indicates that a particulat factor may

undergo a number of different changes. In L. apricus the

evidence at hand likewise indicates that the factors pro-

ducing striped white and light blue respectively are each

allelomorphie to that for dark-blue flower color. Whether

they are allelomorphic to one another remains to be

shown, though that would be a probable supposition.

From the data presented in this paper it can not be

said whether flower color and seed-coat color are both due

to the same factor or to linked factors. The latter is,

however, indicated by two facts. In the first place we

already know more heritable patterns of coat color than

there are flower colors associated with them. At least

446 THE AMERICAN NATURALIST [Von LV

three seed patterns are found in association with dark-

blue flowers in different pedigrees. In the second place

there is the case of the dark-blue plant, VIJI-27, which

produced seeds with the light color characteristic of light

blues.

In the great majority of mutations described up to

the present time the new character is recessive to the nor-

mal one (1). They are apparently due to the loss or

inhibition of a previously existing factor. It is interest-

ing to note that the two factors for light blue and striped

white here reported are both dominant, the former com-

pletely and the latter incompletely so. Furthermore,

many of the other variations observed are in the nature

of additions. Yellow color on the front of the banner

and orange lateral spots, although their heredity is not

yet known, are certainly to be considered as in the nature

of additions.

The mutations in Lupinus represent three categories

(11). In the whites a positive character has been lost. In

the light blues and striped whites a character has been

replaced by another. This would naturally be supposed

to be due to some alteration of the factor governing the

somatic character. In the case of the orange spots and

lemon-yellow fronts one is led to suppose the addition

of a new factor, especially so in view of the fact that they

are not constantly associated with one another or with

white flowers. White flowers may occur without either

lemon fronts or orange spots or with both or with one

and not the other.

These three categories of characters naturally suggest

that there are also three sorts of factorial bases for them.

Loss of a character appears readily explicable either on

the assumption of an actual loss of the factor in the chro-

mosome or of its becoming latent. Multiple allelomor-

phic factors seem to be located according to the work of

Morgan and his associates (8) at identically the same

loci in their respective chromosomes and must therefore

be thought of as different changes of the same factor.

No. 640] VARIATION AND HEREDITY IN LUPINUS 447

Additional characters when inseparably linked with an

old one might be due to a change in the original factor,

but when not linked, or so loosely so as not to maintain a

constant association, they would have to be considered as

due to a new factor. A new factor might originate by

the subdivision of an old one and the subsequent differ-

entiation of one part. In this case as well as in the case

of the actual loss of a gene, homologous chromosomes of

the hybrid between the new and old form would present

the situation originally conceived in the Presence and

Absence hypothesis (2) of an actual gene paired with its

absence. In the other cases of modified factors this

would not be true.

Summary.—l. The genus Lupinus presents an as-

semblage of closely related and difficultly separable

species.

2. The present paper reports some results of a 6-year

field and garden study of L. apricus vallicola, L. piper-.

smithu, and L. nanus.

3. The variations described concern the form and color

of the flower, the shape and size of the pod, and the

color and markings of the seeds.

4. Dark-blue and pink-flowered races breed true.

5. Striped-white flowered races are heterozygous for

a single factor, which in the homozygous condition pro-

duces white flowers.

6. Light-blue flowers are due to a single dominant fac-

tor, indistinguishable in the homozygous and heterozy-

gous condition.

7. Dark seed coats are linked with dark-blue flower

color, but probably due to separate factors.

8. The factors for light-blue and striped-white flowers

are both allelomorphic to that for dark-blue and not im-

probably constitute a system of multiple allelomorphs:

9. Mutations are frequent, some are already known to

be dominant, and others appear to be in the nature of

additions of new characters and factors and so progres-

sive in the sense of de Vries.

448 THE AMERICAN NATURALIST [Vor. LV

10. On account of the frequency of dominant and pro-

gressive mutations and notwithstanding the difficulties

of seed germination this genus merits the attention of

geneticists.

LITERATURE CITED

1. Bateson, W. Mendel’s Principles of Heredity. RATA 1909.

2. Bateson, W. Problems of Genetics. New Haven

3. Duerden, J. E. Parallel Mutations in the iti go N. S., 52:

165-168. 1920.

4. Greene, E. L. Two New NE Leaflets of Botanical Observation

and Criticism, “i 67-68. 1910. ;

5. Heller, A. A. The North cis Lupines—V. Muhlenbergia 7:

85-95. 1911.

6. spain E. C. Spore eae wane in Hybrids and the Mutation Hypoth-

of de Vries , 58: 322-336. Pls. 22-25, 1914.

E ue WL, seg a ee Wontan California. San Francisco,

911

8. Martin, Í. N., and H. S. Coe. Sweet Clover Seed. U. S. D. A. Bull.

4, 1920.

9. Morgan, T. H. The Mame Basis of Heredity. Philadelphia, 1919.

10. Smith, C. P. Notes upon some STR Lupines of the Micranthus

Grou kagi 6: 135-1 1911.

11. r À. s A ape pation in Drosophila funebris.

Sci , N. 8., 48: 72-73. 1918.

12, de Vries, Bae “2 ae Tn Chicago, 1909.

STUDIES ON PARASITIC COPEPODS OF THE

GENUS SALMINCOLA*

NATHAN FASTEN

OREGON AGRICULTURAL COLLEGE, CORVALLIS, OREGON

THE parasitic copepods afford a very interesting group

of animals for the biologist. Not only do these organisms

offer many fascinating problems for the pure scientist,

but from the commercial standpoint they are extremely

important as parasites of our food and game fishes.

There is every degree of parasitism amongst these crus-

tacea, from those which spend a very small portion of

their existence as parasites to those which are parasitic

throughout almost their entire life. My own studies,

covering a period of nearly ten years, have been confined

to some of the most highly specialized members of the

latter group, belonging to the genus Salmincola of the

family Lerneopodide.

The Salmincola are parasitic on the Salmonide,

which include such important fishes as salmon, trout, lake

herring and whitefish. These copepods are all built on

a similar plan and can be easily recognized. The adult

females are the ones which are usually encountered, and

these may be attached to the delicate membranes of the

gills, gill chambers, fins and mouth of the host. Here

they hang on and are supplied with a constant stream of

fresh blood, which serves as their sole food.

These adult copepods can be readily seen with the

naked eye. They are quite large, measuring a few milli-

meters in length, and are yellowish-white in color. An-

teriorly they are fastened to the host by means of two

second maxille and a chitinized bulla. This last named

structure is imbedded in the tissues of the host. Pos-

teriorly each female possesses a pair of slender freely

1 Delivered before the Biological Club, Oregon Agricultural College.

; 449

450 THE AMERICAN NATURALIST [Vor. LV

dangling egg-sacs within which the embryos undergo

complete development.

During the last few years our knowledge of the Sal-

mincola has been increased largely through the efforts of

Wilson and the present writer. Wilson, in 1915,? pub-

lished a key to the various species of Salmincola found in

North America, while the present writer (1912-1919)*

has published a series of papers on the behavior, mor-

phology, life-history and economic importance of two of

these forms, namely Salmincola edwardsii (Olsson)

Wilson, which parasitizes the brook trout of our middle-

western and eastern states, and Salmincola falculata

Wilson, which is parasitic on salmon and trout of our

Pacific states. In speaking of the former species, Small-

wood, in a recent paper,‘ says:

These parasites are widespread in the United States in the native

trout streams, and in Canada and Europe. The first scientific record

of this particular parasite is by Linneus in 1761. It seems strange

that an animal could be known for so long and its habits not be under-

stood until within the past five years.

The writer is at present engaged on other species of

Salmincola which dwell on various salmon and trout of-

the northwest section of the United States. From all

appearances the different stages in the life histories of

the various species of Salmincola seem to be more or less

similar and, therefore, will be briefly outlined. :

As already mentioned above, the young larval copepods

undergo development within the egg-sacs of the attached

females. When these larve are mature, they rupture the

egg-sacs and escape into the water as minute, freely-

swimming organisms that closely resemble free-living

pelagic copepods. Although they measure about one

thirty-fifth of an inch in length, they are very active and

2 Proc, U. S. Nat. Mus., Vol. 47, pp. 565-729.

3 Report Wis. Fish. Com., 1911-12, pp. 12-22. Jour. An. Beh., Vol. 3, pp.

36-60. Biol. Bull, Vol, 27, pp. 115-127. Biol, Bull., Vol. 31, pp. 407-419.

Pub. Puget Sound Biol. Sta, Vol. 2, pp. 73-77. Pub. Puget Sound Biol. Sta.,

Vol. 2, pp. 153-181.

4 Amer. Nart., Vol. 52, pp. 322-352.

No. 640] PARASITIC COPEPODS 451

swim about with a snappy spiral dart. They may thus

swim about for nearly two days, constantly searching for

a host to which to attach themselves. They dart here,

there and everywhere: if not successful in meeting a host,

they soon die, but if one is found they attach themselves

and carry on their life-cycles to completion.

In Salmincola édwardsii it has been found that the

larval copepods swim about near the surface of the water

throughout the day, but at night they sink down to lower

depths near the bottom of the stream. These migrations,

although contrary to the general migrations of free-living

copepods, are, nevertheless, of great benefit to these para-

sitic forms in that they are parallel with the migrations

of the hosts. Brook trout generally feed near the upper

surfaces of the streams during the day and at night they

sink down to lower levels. This similar behavior on the

part of the parasite and the host makes it much easier

for the parasite to meet its host and thereby carry out

its life-cycle.

The manner in which the larval copepod attaches itself

to the host is extremely interesting. Each larva posses-

ses powerful mouth parts and a peculiar attachment fila-

ment which aid in the attachment of the organism. On

coming in contact with a desirable portion of the host,

the parasite first rasps a hole in the tissues by means

of its mouth parts. Then the attachment filament is

brought in contact with this cavity and by means of the

contraction of numerous thin head muscles which are

attached to the proximal end of the attachment filament,

the bulb-like distal end of the filament is driven into the

cavity. The glue-like secretion of the attachment fila-

ment as well as the regenerating tissue of the host soon

attach the copepod quite securely.

The copepod now undergoes degeneration. It loses its

segmentation as well as its plumose swiniming feet. The

abdomen rounds out, becomes larger and more bag-like

in outline. The mouth parts also change their appear-

ance. The mouth itself grows into a prominent tube-like

- 452 THE AMERICAN NATURALIST [Vot LV

piercing organ, which is capable of puncturing the tissues

of the host for purposes of sucking blood. This is the

exact method by means of which the attached parasite

feeds itself.

About a week after attachment to the host, the modi-

fication of the parasite has been so complete that one can

hardly recognize any resemblance between it and the free-

living larva from which it was derived. In another week

and a half, that is, about two and a half weeks after

attachment, the copepods have reached sexual maturity

and are ready to undergo fertilization. The males can

now be easily distinguished from the females. This was

not possible previously. The females are veritable giants

as compared with the males, being about three or four

times the size of the latter.

The only male ever discovered in the genus Salmincola

is that of Salmincola edwardsii, which has been described

and figured in the Biological Bulletin for 1914. The

writer has just completed the study of another male of a

different species of Salmincola, namely, Salmincola beani

Wilson which he recently discovered on the gills of the

chinook salmon.’ This new male shows the same size

difference when compared with the female as does the

male first mentioned.

Prior to fertilization, the males and females hang side

by side on the tissues of the host. In order to accomplish

fertilization, the male undergoes a rather peculiar be-

havior. He begins circling movements and somewhere

in his vicinity he comes in contact with a female. As

soon as this occurs, the male clasps the female with his

maxillipeds and at the same time he releases his hold on

the tissues of the host. The male then creeps towards

the posterior region of the female’s body, in the neighbor-

hood of the genital pores, and here he attaches himself in

position for fertilization. The male next bends his abdo-

men upward toward the genital openings and soon

extrudes two pear-shaped pouches known as spermato-

5 In press, Biol. Bull.

No. 640] PARASITIC COPEPODS 453

phores. These are manipulated by the free maxille of

the male and are ultimately attached near the genital

pores of the female. The spermatophores contain a

cement-like material which aids in their attachment.

They are also filled with large numbers of mature sper-

matozoa which wander through the genital pores and

become stored within the spermatheca of the female.

Here these male gametes remain dormant until the ova

of the female are ripe for fertilization. When the wander-

ing of all the spermatozoa has been completed the

spermatophores collapse and soon come to resemble

transparent, shell-like, yellowish spheres. The female

may be fertilized more than once. Oftentimes as many

as six spermatophores may be found clinging to the

genital pores of some of the females, showing that these

have been fertilized three times.

After fertilization, the male drops off the body of the

female and soon dies. The female, however, lives on and

completes the life-cycle. She now undergoes extreme

degeneration, increases enormously in size, and develops

a large number of eggs which become clearly visible with-

in her abdomen. At the same time two slender mem-

branous egg-sacs make their appearance at the posterior

margin of the female’s body. When the ova are ripe

they are passed down through the oviducts and as they

migrate past the spermatheca they are fertilized by the

stored spermatozoa. The embryos are then transferred

to the egg-sacs where they carry on their complete de-

velopment. In about a month the young are liberated as

free-swimming larve ready to begin the cycle again. In

Salmincola edwardsii two batches of young are produced,

each numbering about one hundred and twenty individ-

uals. After all the young have been liberated, the adult

females die and soon deteriorate on the tissues of the

ost.

Although these copepods are not, ordinarily, very

dangerous to fish in their natural haunts, yet from the

standpoint of fish-eulture they are of considerable eco-

454 THE AMERICAN NATURALIST [Von. LV

nomic importance. When once they make their appear-

ance in our hatcheries they cause a great deal of damage

and loss amongst the fish. Here conditions are ideal for

parasitism. The ponds are small and large numbers

of fish are crowded into them. Because of this situation

the parasitic larve have very little trouble in finding their

hosts. At the same time the current of water which cir-

culates through the hatchery ponds is not swift enough

to interfere with the movements of the parasitic orgau-

isms. It is therefore a matter of a short time before

most of the fish become heavily infested with copepods.

While the young fish as well as the adults are attacked

in the hatchery ponds, nevertheless it is mainly the adult

fish which are most heavily parasitized. These are

attacked by so many of the copepods that they are ulti-

mately killed. It is by no means uncommon to find as

many as two hundred and fifty copepods on one trout.

Recently I found around five hundred copepods on the

gills of a single chinook salmon. In such eases of para-

sitism the injury to the host is considerable. In the first

place, the parasites suck enormous quantities of blood,

thereby depriving the host of a large amount of nourish-

ment. Secondly, when the copepods attach themselves,

they injure the tissues of the host, thereby making it

possible for injurious spores and bacteria to enter and

set up secondary infections of a serious nature. And

lastly, the injured tissues swell and develop into so-called

‘“ sear tissues,’’ which interfere with the normal functions

of the host. Taking all these facts into consideration,

- there is little wonder that fish succumb under the attacks

of these parasites, particularly in hatchery ponds where

conditions are just right for parasitism. In one Wiscon-

sin hatchery the author found that in a single year about

twelve thousand adult trout out of fourteen thousand

kept in outdoor ponds died from the attacks of these

copepods. .

Many states have had this trouble for years, with very

serious losses. The writer has devoted considerable

No. 640] PARASITIC COPEPODS 455

attention to the control of these parasites, and has recom-

mended the following remedies in the state of Wisconsin.

These have also been found useful in other states where

the same type of parasitism has made its appearance.

1. When the water supply is polluted, sand filters

should be installed at the mouth of the water stream as

it makes its way into the hatchery ponds. The sand

catches most of the free copepods before they enter the

hatchery, thereby preventing them from attacking the

fish.

2. The young fry should be given salt baths quite often.

The salt solution kills the copepods during the early .

stages of attachment. At the same time this solution

makes the fish more resistant to the attacks of the

parasites.

3. Since the adult fish are the ones most heavily para-

sitized, it is better to do away with these as soon as

possible and to keep only the younger fish for spawning

purposes.

4. Inasmuch as the free-swimming stages of the cope-

pods are strongly attracted by intense light, powerful

are lights should be erected at various points over the fish

ponds. By means of fine gauze bags towed over the illu-

minated regions, a large number of the copepods can be

gathered and removed.

5. The introduction of certain types of minnows into

the hatchery ponds tends to keep the parasites down.

These minnows feed on the free-living larve of the cope-

pods, thereby destroying many of them before they have

the opportunity of coming in contact with the proper

ost.

Another means of overcoming this sort of parasitism

which has often suggested itself to the writer is, through

breeding, to develop a strain amongst the hosts which

would be practically immune to the attacks of the para-

sites. This appears to be possible when one considers the

fact that under similar conditions the hosts show varying

degrees of resistance to the parasitic organisms. Some

456 THE AMERICAN NATURALIST [Vou. LV

are attacked very lightly, while others become heavily

parasitized. Doesn’t it seem logical to speculate that

through intelligent selection and breeding, one could de-

velop resistant strains of fish, which would be attacked

by so few of the parasitic copepods that the parasites

would be almost a negligible quantity?

These remedies, of course, are not absolute, but they

may help a great deal in reducing the loss of the fish. In

cases of such parasitism there is no absolute cure known.

A most desirable remedy would be one which would de-

stroy the adult copepods while they are attached to the

structures of the fish, without in any way harming the

latter; but all attempts in this, direction have thus far

been without success. The hosts are so delicately consti-

tuted that they can withstand only a very slight change in

their environmenta! medium. The adult copepods, on the

other hand, can resist powerful chemical solutions by

virtue of their resistant body walls. It is obvious that

the weak link in the chain of the life-history of these

parasites is the free-living period, and in view of this,

the real solution seems to be quite clear. One must catch

the organisms as they break out of the egg-sacs of the

mother and kill them before they come in contact with

their hosts. As with a good many of our modern dis-

eases, ‘‘ prevention before parasitism occurs ’’ should be

our motto, rather than ‘‘ cure after parasitism.”’

SHORTER ARTICLES AND DISCUSSION

COLLINS’S REMARKS ON THE VIGOR OF FIRST

GENERATION HYBRIDS

IN a review of the theories regarding hybrid vigor Collins’

has attempted to show that the two objections which were long

upheld as precluding the possibility of dominance accounting

for heterosis were without foundation. The suppression of dele-

terious factors, he considers, is adequate to account for the ob-

served facts without considering the phenomenon of linkage.

The two objections which were raised against the hypothesis

of dominance as a factor in hybrid vigor before the importance

of linkage became generally known are as follows: (1) Domi-

nance of independent factors would give an asymmetrical distri-

bution to the progeny populations of those individuals which

show an increase in growth when crossed. (2) Free assortment

would make possible a recombination of all the dominant favor-

able growth factors into a homozygous fixed race which would

not be reduced by inbreeding. Neither of these objections holds

when linkage is taken into consideration. Collins believes that

they also do not apply when linkage is left out of consideration.

Collins shows numerically and graphically that with a large

number of factors involved the skew curve of the theoretical dis-

tribution of independent dominant factors approaches the type

-of the normal curve. He points out that, with characters de-

pendent upon a large number of factors, only populations with

larger numbers than have been dealt with statistically would

exhibit any noticeable tendency toward skewness. This is a

good point well brought out which previously had been neglected.

But this would apply only to progenies which have a restricted

range in comparison with their parental populations. In those

cases where the range of the segregating generations with small

numbers nearly equals the combined range of the original races

as exhibited by the characters which show heterosis the number

of main factors which govern the expression of this particular

character can not be large. Therefore, if it were merely a mat-

1 Dominance and the vigor of first generation hybrids. Amer. Nar., 55:

116-133, 1921.

457

458 THE AMERICAN NATURALIST [Vor. LV

ter of dominance without linkage, such distributions would be

expected to show right-hand skewness. But they do not con-

sistently do so.

In regard to the second objection, that of recombination of all

favorable factors, Collins has given a large number of figures to

show what was already well known, that with a large number of

factors the chances for recombination are remote with the small

progenies grown in experimental plots. It was not intended to

maintain that pedigree cultures were adequate to show that such

a recombination could not be made. The point in mind, if not

clearly expressed, was that natural selection in isolated popula-

tions of cultivated plants had not brought about any noticeable

approach to stability. In the hills of New England maize has

been grown for long periods of time in isolated fields. Some

varieties have probably been grown for at least fifty years with-

out admixture. Yet these varieties when self-fertilized show as

rapid a reduction in growth as other varieties which are lately

the produet of extensive hybridization.

There is an enormous difference in the possibilities for imme-

diate recombination with and without linkage. To illustrate:

with twenty independent factors the chance for the bringing

together of all dominants in a homozygous state in one genera-

tion is theoretically one in 4%. With the same twenty factors

distributed by twos in ten different chromosomes, each being

separated by ten units of crossing-over, the chance for recombi-

nation is theoretically one in 20%. This is a difference in total

numbers so vast as to be almost inconceivable. Working over

long periods of time, linkage may not be a hindrance to recom-

bination, as factors once brought together tend to stay together

as firmly as they once resisted separation. Many cross-fertilized

species in the wild whose age is measured in geological periods

rather than years are stable. But cultivated forms even when

isolated for a considerable time show no noticeable approach to-

ward this condition. It seems reasonable to suppose that the

arrangement of factors in the chromosomes has something to do

with this state of affairs. Therefore, until the chromosome

theory of heredity was developed, there was considerable plausi-

bility to the older view that something besides mere dominance

was responsible for heterosis.

Even so, I am perfectly willing to admit that there is no clear

way of deciding the argument as to whether or not the old objec-

No. 640] SHORTER ARTICLES AND DISCUSSION 459

tions were valid. But how important is it now to make this

decision? Linkage is a fact and must be taken into considera-

tion. True, the evidence in support of the chromosome hypoth-

esis from maize is not extensive. But hybrid vigor is a wide-

spread phenomenon shown by many organisms. The dominance

hypothesis apples to Drosophila as well as to maize.

This failure to look outside of the corn field has led Collins

to make certain statements to which I must take strong objec-

tion. He is inclined to believe that the suppression of deleteri-

ous factors is all that is involved in the vigor derived from cross-

ing. This may be true for Drosophila, but there are many cases

of wild species of both animals and plants as well as of naturally

self-fertilized varieties of cultivated plants which show an un-

mistakable increase in growth after crossing. Take Naudin’s

Datura crosses which doubled in height, Kélreuter’s Nicotiana

hybrids which astonished their producer, the hybrid walnuts,

both natural and artificial, and Gerschler’s fish hybrids, to name

a few notable illustrations. Collins himself has given us several

good illustrations of remarkable vigor shown by hybrids of many

varieties of maize from different parts of the world. Here it is

clearly not a matter of suppressing deleterious characters. The

parental types are normal, vigorous and perfectly capable of

maintaining themselves in their own way. But crossing brings

about a new combination of hereditary qualities. By utilizing

the best from both parents the hybrid is able to obtain a sur-

passing development. As long as variation exists different indi-

viduals will have unlike germinal potentialities. Crossing tends

to bring these different possibilities together. Dominance en-

ables the offspring to take advantage of the more favorable _

factors. This is as true of domesticated races as it is for wild

species.

Furthermore, there is abundant evidence that many factors

are without effect unless working in consort. In plants, colors

of various parts, and in animals, coat patterns, are conspicuous

examples of this complementary action. These characters are

possibly of no importance in growth, yet they illustrate a state

of affairs which is probably of real significance. Crossing makes

it possible to assemble the component parts.

As Collins says, to consider hybrid vigor as the suppression of

deleterious heredity as compared to the bringing together of a

greater number of favorable growth factors is, to a certain ex-

460 THE AMERICAN NATURALIST (Von. LV

tent, merely a different way of looking at the same thing. But

it puts the emphasis on the wrong side and is wholly inadequate

to account for all the manifestations of hybrid vigor. It would

be unnecessary to discuss this were it not for the fact that his

way of looking at the matter leads him to think that there is no

essential distinction between the Darwinian view of inbreeding

as a process leading toward extinction and the more recent con-

ception that the results of this system of mating depend upon

the inheritance received. Collins says:

Many of the older writers on heredity have held that inbreeding is a

cause of degeneration. In avoiding ambiguous words “ cause” is one of

the first that must go. If forced to define their position this school

would probably be content with the statement that degeneration is a

necessary consequence of inbreeding, the intermediate step or nature of

the process being unknown. Is this. conception really at variance with

the idea that degeneration results from the increased number of un-

favorable recessive characters brought into expression by increased

homozygosity? Does not this conception rather amplify the older,

general and indefinite position by explaining how degeneration may

be brought about? (P. 124.)

Leaving aside all question of definition of terms, let us con-

sider the results of the two views when applied in practise. To

say that abnormal and undesirable individuals appear after close

mating is very different from supposing that such forms have

their origin in the system of mating. Whether or not this is

stating the matter fairly, breeding practises have been in accord

with the latter view. As a result of inbreeding we now know that

aberrant individuals bordering on the teratological often come

. to light. Along with these types which are truly degenerate in

any sense of the term (but inbreeding has nothing to do with

their origin) there are perfectly normal individuals which suffer

‘in comparison with their more heterozygous parents in that they

are only slower in growth, are not so resistant to unfavorable

conditions and are not so productive. Inbreeding is solely a

process of sorting out. Some bad material is brought to view

which can be discarded. But along with this there is all the

good material that was in the stock, and this can be used to re-

build a better breed than existed at the start. Before the era

of Mendelism there was little conception that it was the stock

that was at fault and not the system of mating. Even though

it was the appearance of abnormal and bizarre forms which gave

No. 640] SHORTER ARTICLES AND DISCUSSION 461

the bad name to inbreeding in the past, the less vigorous off-

spring frequently resulting from inbreeding, although healthy,

were also considered to be valueless for further propagation and

were quickly disposed of.

This is still the belief and practise of live-stock breeders.

Those who do not know the principles involved think that in-

breeding has permanently injured the families with such weak-

ened individuals. Equipped with the results of two decades of

genetic investigation, we can say, ‘‘ No! this is not so. Nothing

has been lost. These less vigorous inbred individuals of no ap-

parent worth have potentially great value.” A widespread re-

ception of this idea has possibilities of great practical outcome.

Not to see clearly the important distinction which there is here

between the present and former views is not to appreciate the

real progress which the combined genetic research of twenty

years has made along this line.

D. F. Jones

CONNECTICUT AGRICULTURAL EXPERIMENT STATION,

NEw HAVEN, CONN.

AN APPARENT CASE OF SOMATIC SEGREGATION

INVOLVING TWO LINKED FACTORS!

Somatic segregation as an ordinary occurrence, and especially

as a source of definite progeny ratios in subsequent sexual repro-

duction, seems highly improbable. The evidence connecting nor-

mal segregation and recombination with meiosis and fertilization

is too strong. As a matter of occasional mitotic abnormality in

heterozygous poi however, the question of somatie segre-

gation is still o

Any ‘‘bud speek? involving apparently simultaneous change

of two or more non-allelomorphie factors is therefore of special

interest, since the probability of its occurrence through two

nearly simultaneous factor or point mutations seems very

remote. Either deficiency mutation, which seems to mean

(Bridges, 1917) the loss of a normally present portion of a

chromosome, or the development or resolution of a condition of

‘*duplication’’ (such as vermilion-sable duplication in Drosophila

melanogaster; Bridges, 1919, p. 646) might produce the effect

ih question. So, also, might a process properly descri

‘‘somatic segregation,” in which at some mitosis one daughter .

t Paper No. 60, University of California, Graduate School of Tropical

AS and Citrus Experiment Station, Riverside, California.

462 THE AMERICAN NATURALIST [Vor. LV

cell received both halves of one mother-cell chromosome, while

the other daughter cell received both halves of the homologous

chromosome. Such a variation, then (Muller, 1920, p.

he an almost eee decision ‘between factor mutation and

‘‘mitotie irregularity.

A ‘‘bud variation’’ Bo involving two linked factors

was observed in a culture of Matthiola annua, at Riverside, Cali-

fornia, in 1916-17. Unfortunately the factorial relations are

not entirely clear, but the case seems decidedly significant never-

theless.

The plant in question occurred among progeny of a ‘‘slender’”’

parent (25b—6—8-6; Frost, 1919). The slender type (S’) is one

of several aberrant forms evidently dependent on factors linked

with the factor ‘‘for’’ single (normal) flowers (D). Slender

parents have given (Frost, 1919) on the average about 32.5 + 2.0

per cent, of slender progeny, most of the rest being ‘‘Snowflake’’

(normal). The constitution of the slender 'single parent men-

tioned appears to have been S’D/s'd. Both S and D (or a

factor completely linked with D) appear to be imperfectly re-

cessive for a lethal effect; no functional pollen carries D, and

S'S’ zygotes appear to be non-viable, while S’s’ zygotes are some-

what weak and probably are selectively eliminated before germi-

nation.

Plant 25b-6-8-6 gave the following progeny: slender, 18 or

19 (2 double, 1 undetermined, rest single); Snowflake, 25 (1

single, 24 double) ; total, 44. One plant was noted, at the age

of about seven months, as having the upper main stem leaves

like Snowflake, but the rest slender. When mature this plant

had produced from one side of the main stem at least three

primary branches, all slender and single, two at least yielding

seed. The main cluster was stout, and, although its flowers seem

not to have been noted as peculiar while in bloom, it produced

persistent sterile pistils; at least two of these pistils were ab-

normally broad, each enclosing a cluster of petal-like parts. One

2 This paper was written, aside from some revision of this second para-

graph, before I saw Muller’s paper here cited.

3 All the singles of such a ‘‘double-throwing’’ race are therefore hetero-

zygous for doubleness, while the doubles (dd) are sterile (Frost, 1915). A

back cross of two Snowflake plants by pollen of 25b—6-8-6 (crosses 23ca and

23ea; Frost, 1919, table 36) gave about 22.4+ 2.6 per cent. of slender

progeny, including only 2 (or 1) doubles out of 26 slenders, while the —

flakes were about half doubles. Evidently both eggs and pollen carried som

factor or factors os to this result,

No. 640] SHORTER ARTICLES AND DISCUSSION 463

stout flowering branch, well above the others, evidently was

similar to the primary inflorescence. Near the level of the

uppermost of the slender flowering branches, on the opposite

half of the stem, arose two stout branches, which bore Snow-

flake-like leaves and sterile double flowers.

It would seem that some change eliminating the factors 8’ and

D occurred, probably-in a single cell, at the growing point of the

young main stem. The Snowflake double (s’d/s’d) cells result-

ing, perhaps because of their normally greater vigor of growth,

gradually obtained the ascendency in a large portion of the

stem. The primary infiorescence and the high branch beside it

perhaps remained in a chimerical condition, while the two lower

stout branches received the new type nearly or quite unmixed.

That the double flowers were somewhat abnormal‘ hardly

lessens the force of the evidence in relation to the improbability

of factor mutation...Even the two lower stout branches may

have been periclinal chimeras, or the new factorial constitution

may have been (as for example through a ‘‘duplication’’ shift-

ing of chromosome material) somewhat different from that of a

normal double. Plainly some change occurred that involved,

nearly or quite simultaneously, two factors in linked loci some

distance apart. This change was probably not. factor (point)

mutation. It may have been deficiency mutation, itself prob-

ably a mitotic abnormality, or it may have been some other ab-

normal shifting of a chromosome or a portion of a chromosome.

A further consideration is pertinent here. The slender form

and at least one or two others, in arising (Frost, 1916, 1919) in

very small proportions from the normal (Snowflake) type,

show evident linkage phenomena which indicate segregation

rather than immediate mutation. As has been suggested (Frost,

1919), the apparently mutant factor (as 8’ above) may be pres-

ent in ordinary Snowflake singles, but concealed because of the

4 The usual double flowers are ‘‘ petalomanous ’’ (de Vries, 1912, p. 330);

that is, inside the sepals they consist of nothing but an indefinitely pro-

liferated floral axis bearing numerous petals. The flower lives long after

anthesis, and often develops into a short branch bearing secondary flowers

in the axils of its leaves (petals). No trace of stamens and carpels can be

found. These abnormal double flowers, on the other hand, had the four

petals of the typical cruciferous flower, followed by an indefinite number of

smaller curved petals probably representing petaloid stamens; finally, in

he central mass of petals seemed to atise from within a modi-

fied pistil, somewhat as in the case of the less abnormal flowers of the

terminal cluster mentioned above.

464 THE AMERICAN NATURALIST [Vor. LV

action of an epistatie or inhibiting factor 7. Thus the constitu-

tion of the Snowflake singles giving rare slender progeny may

be IS’D/is’d or DS'I/ds'i. A serious theoretical difficulty seemed

to arise in the apparent necessity for several specific ‘‘

hibitors’’ all giving the same ‘‘normal’’ type, and also for rela-

tively frequent dominant mutations. Perhaps, however, the

apparent mutation may usually consist in the development or

disappearance of some condition of duplication in one chromo-

some of the pair concerned. Origin of apparent mutants

through duplication of whole chromosomes, as seems to have

been demonstrated for a remarkably similar series of mutant

` forms in Datura (Blakeslee, Belling and Farnham, 1920), seems

to be precluded in these cases ehy the evident linkage phenomena.

Howard B. Frost

in-

UNIVERSITY OF CALIFORNIA

LITERATURE CITED

Blakeslee, Albert F., Belling, John, and Farnham, M. E.

192 P TEER EAN, Duplication m Mendatlas Phenomena in Datura

Mu = nts. Science, N. S., 388-390.

ie Calvin

1917. Dass: Genetics, 2: 445-465. 14 tables.

1919. Vermilion-deficieney. Jour. General Physiol, 1: 645-656. 9

tables

Frost, Howard B.

1915. The Inheritance of Doubleness in Matthiola and Petunia. I.

The Hypotheses. Am. NAT., 49: 623-636. 1 fig., 2 ae aa

TN Mutation in Matthiola annua, a ‘‘Mendelizing’’ Species. Am.

Jour. Bot., 7: 377-383 figs.

1919. Mutation in Matthiola. Vale. Caif. Pub. Agr. Sci., 2: 81-190.

, 4 charts, 40 table

Muller, H.

1920. Or Changes in the White-eye Series of N ped

mo Bearing on the Manner of Occurrence of Mutation

r. Exp. Zool., 31: 443-473. 3 figs., 2 tables.

de Vries, Bae

1912. Species and Varieties: their Origin by Mutation. Ed. 3, xviii

+ 847 p. Chicago, Open Court Pub. Co.

Tue discovery of a neuromotor apparatus in Diplodinium

ecaudatum (Sharp 1) and Euplotes patella (Yocom 2) con-

firmed by Taylor (3). leads me to expect similar conductile fiber

systems in other ciliates. This expectation has been met in the

No. 640] SHORTER ARTICLES AND DISCUSSION 465

well-known ciliate Paramecium. The neuromotor apparatus of

this organism consists of fine branching fibers in the periphery

and the cytopharynx, converging to the neuromotor center.

In the periphery these fibers are connected to the basal gran-

ules of the cilia and also to the trichocysts. From these organ-

elles they may be traced to the neuromotor center (Fig. 1, n. c.)

located in the endoplasm just anterior to the cytostome. They

)

p” =

—

VAAAALLAARALAASAA

-- ---p. fib, -+- -

J. Wig. 2.

Fic. 1. Paramecium caudatum, diagrammatic sketch showing oral whorl of.

peripheral neuromotor fibers, the neuromotor center, ciliary grooves, trichocyst

ridges, ciliary suture, cytostome and cytopharynx with anterior and posterior

membranelle zones.

Fic. 2. Paramecium caudatum, diagrammatic aboral view, showing aboral

whorl of neuromotor fibers,

Abbreviations: ¢. v., contractile vacuole; c. 8., ciliary suture; o. gr., oral

ve; ac., macronucleus; mic., micronucleus; n. C., neuromot :

o . ora ; *

aboral whorl; cyp., cytopharynx; p. memb., posterior membranelle zone; t. r.,

trichocyst ridges; p. fib., peripheral neuromotor fibers.

466 THE AMERICAN NATURALIST (Vou. LV

are arranged in whorls, one on the oral side, the other on the

aboral side (Fig. 1, or. wh.; Fig. 2, ab. wh.).

The oral whorl is the more extensive. In the oral groove

(Fig. 1, o. gr.) the fibers run obliquely caudad to the cytostome

where they turn and converge obliquely cephalad to the neuro-

motor center. From other parts of the oral surface they run in

gracefully curved lines directly to the neuromotor center.

The fibers of the aboral whorl converge in a large apex on the

right, opposite and slightly posterior to the cytostome. Here

they dip into the endoplasm and run direct to the neuromotor

center.

The entire periphery of the animal is supplied with the diverg-

ing fiber ends of these two whorls. On the left side those of the

oral whorl meet those of the aboral whorl about midway between

the two sides. On the right the inner ends of the fibers of the

oral whorl mingle with those of the aboral whorl near the con-

verging apex of the latter.

~ Two sets of fibers connect the organelles of the cytopharynx

with the neuromotor center. One, a fan-shaped set, runs in the

right wall of the cytopharynx to the anterior membranelle zone.

The other set consists of two fibers which run from the neuro-

motor center’ to the peristomal cilia around the. cytostome and

meet in the posterior margin. From here they run in the oral

side of the eytepharynx and branch profusely into the posterior

membranelle zone and the endoplasm. This posterior zone, the

cilia of which beat in an opposite direction to those of the an-

terior membranelle zone, has not been previously described.

The cytopharyngeal fibers are heavier than the peripheral fibers

and may be seen in living unstained animals under oil immersion.

_ The trichocysts are arranged with reference to the peripheral

neuromotor fibers in whorls. They reach the surface of papille

which constitute interrupted ridges (Fig. 1, t. r.). The cilia,

however, spring from longitudinal grooves. The grooves from

each side of the oral surface meet in a series of V’s, the apices

of which lie in a line, the ciliary suture (Fig. 1, c. s.) which ex-

tends obliquely through this surface from the anterior to the

posterior end,

Fibers were found in 2.5,» sections connected to the basal

granules of the cilia and running into the endoplasm. Khainsky

(4) also found these fibers and called them ciliary rootlets.

‘They are here interpreted as the ends of the peripheral neuro-

_ No. 640] SHORTER ARTICLES AND DISCUSSION 467

motor fibers. Similar fibers have been found connected to the

inner ends of the trichocysts.

From the foregoing it is seen that the neuromotor system of

Paramecium consists of fibers running from the neuromotor

center to the membranelle of the cytopharynx and also from the

same center to the basal granules of the peripheral cilia and to

the trichocysts. Its morphology suggests that it is conductile in

function adapted to coordinate the movements of the peripheral

cilia and the cytopharyngeal membranelles.

The peripheral fibers were discovered in whole mounts fixed,

stained and dehydrated in centrifuge tubes. The best stain was

Heidenhain’s iron alum hematoxylin. They were not seen

with this stain when the animals were attached to the slide by

ege albumen. However, contrary to Neresheimer (5), the distal

fiber ends were sometimes seen in such preparations when Mal-

lory’s triple connective tissue stain was used. Complete cyto-

logical details were worked out only from the hematoxylin

preparations.

This idea of staining non- distorted animals in centrifuge

tubes resulted from micro-injection studies, It was found that

animals survived such operations only when isolated in rounded

drops. Blisters invariably formed when they were held flattened

to the cover by water-glass surface tension (Taylor 3). An

apparatus embodying the principle of Taylor’s micro-injection

pipette (Taylor 6) was constructed by means of which isolation

was accomplished in such small drops that only a very limited

movement of the animal was possible. To secure rounded drops

the cover was coated with a thin film of oil as described by Barber

(7)

Three kinds of experimental methods in attempting to dem-

onstrate that the fibers are conductile were carried out as

follows:

Griibler’s methylene blue which stains nerve fibers in meta-

zoan tissue (Wilson 8) gave negative results when injected into

Paramecium

An antéro-postérior gradient was demonstrated as follows:

The organisms were isolated in 4 per cent. to 6 per cent. alcohol,

Yo per cent. nicotine, 1 per cent. antipyrin, or 1 per cent. mor-

phine hydrochlorate. In all cases the anterior cilia ceased

beating at least ten seconds earlier than the posterior cilia

and those of the cytopharynx. The animals did not disintegrate

as do planarians and annelid worms in these solutions so that

468 THE AMERICAN NATURALIST (Vor. LV

the physiologically anterior end (Child 9) could not be deter-

mined. But the antero-posterior gradient is what one would

expect in animals possessing fibers which conduct efferent im-

pulses (Tashiro 10).

Contrary to Neresheimer (5) the animals are narcotized in

these solutions.

Micro-dissection experiments showed that the coordination of

movement of the cytopharyngeal membranelles is interrupted

when the neuromotor fibers are cut. Those posterior to the cut

beat slower and with smaller amplitude than those anterior to it.

Extensive destruction of structures in the region of the neuro-

motor center or motorium destroyed coordinated movement of

the peripheral cilia. In one case in animals isolated in gelatine

four zones of cilia were seen. Those of one side beat in oppo-

site directions to those of the other.

CONCLUSIONS

A complex fibrillar apparatus has been differentiated in Para-

mecium, It connects the membranelles of the cytopharynx and

the peripheral cilia and also the trichocysts with the neuromotor

center. Therefore, the morphology of this system suggests that it

is conductile. Experimental data strengthens this morphological

evidence; first, because the antero-posterior gradient that exists

here is that which would be expected in an animal possessing a

complex system of fibers which conduct efferent impulses from

the anterior end to the neuromotor center; second, the micro-

dissections indicate that coordinated movement of the cytopha-

ryngeal membranelles is interrupted when neuromotor fibers are

severed and coordinated movement of the peripheral cilia is

interrupted when the neuromotor center is destroyed.

CHARLES W. REES

ZOOLOGICAL LABORATORY, UNIVERSITY OF CALIFORNIA

LITERATURE CITED

1. Sharp, R. G.

1914. Diplodinium ecaudatum with an Account of its Neuromotor

gy ratus. Univ. Calif. Publ. Zool., 13, 43-122.

2. Yocom, H. B

1918, The Neuromotor Apparatus in Buplotes patella. Univ. Calif.

Publ. Zool., 18, 337-396, pls. 14-16.

3. Taylor, C. V

192 Vecasiatenkive of the Function of the Neuromotor Apparatus

in Euplotes by the Method of Micro-dissection. Univ. Calif.

Publ. Zool., 19, 403-470, pls, 29-33

No. 640] SHORTER ARTICLES AND DISCUSSION 469

4, Khainsky, A.

1910. Für pE ets a und see aad einiger Infusorien (Para-

m) a und einer neuen histologischen

Memede” peg Prot., P Se 3.

5. Neresheimer, E. R.

1 Die hohe histologischen eT bei heterotrichen Cilia-

ten. Arch. Prot., 2, 305-324,

6. Taylor, C. V.

1920. An Accurately Controlable Micropipette. Science, N. S., 51,

617-618

7. Barber, M. A.

1914, The pets Method in the Isolation of Single Microorganisms

the Inoculation of EI into Living Cells. Phil.

tan “Bok. Sec. B, Trop. Med.,

8. Wilson, J. G.

1910, Intravitam Staining with Methylene Blue. Anat. Rec., 4,

267-277.

9. Child, C. M.

1915. Senescence and Rejuvenescence. Chicago. Univ. of Chicago

Press.

. Tashiro, S.

1917. A Chemical Sign of Life. Chicago. Univ. of Chicago Press.

A NEW MUTATION IN THE HOUSE MOUSE?

A NEw and distinct mutation in mammals is not a frequent oc-

currence, and therefore the record of a recent dilute form of the

house mouse, allelomorphie to color and albinism, is perhaps

justifiable.

The infrequency of mutations in mammals may be due to

greater stability of the germ plasm than in such forms as in-

sects, for example, Drosophila; or may be due to our lack of

opportunity for examining as large a population of mammals

as of insects; or possibly may be due to a more frequent lethal

effect associated with mutations in mammals. Whatever the

cause is, a tendency toward similar mutations in closely related

groups of mammals is apparent and suggestive. The fact that a

given type of mutation has occurred in one group is some promise

that a corresponding mutation is possible and may occur in a

closely related group. The pink-eyed mutation (giving ping-

eyed colored varieties) in mice has been known for some time.

A similar mutation in rats was described recently (Castle, *14).?

1 Paper No. 17, Geneties Laboratory, Illinois Agricultural Experiment Sta-

tion.

2 Castle, W. E., 1914, Am. NAT., Vol. 48, p. 65.

470 THE AMERICAN NATURALIST [Vor. LV

In both rats and mice, this gene greatly reduces the production

of black and brown hair pigment, but leaves yellow undisturbed.

The linkage of pink-eye to albinism in both forms gives added

evidence that these genes are similar if not actually identical.

A gene for pink-eye is also known in the guinea pig (Castle

14)*, but it has not yet been established that pink-eye and

albinism are linked in this form. It may be unsafe to homologize

strictly the gene for pink-eye in mice and rats with the gene in

guinea pigs which produces similar somatic effects. A deep

red-eyed (almost black-eyed) yellow rat has also been described

(Castle, *14).4 The gene for red-eye is linked to the gene for

pink-eye. This mutation has not yet been observed in mice.

Brown varieties of mice and guinea pigs have been known for a

long time. Brown rabbits (the Havana variety) and the brown

. roof rat (Patterson, ’20)° are recent productions. While no

brown variety of the common rat (Mus norvegicus) is known,

the form should be possible and its discovery is only a matter of

time and opportunity. A set of quadruple allelomorphs in the

guinea pig includes intense color, dilution, ruby-eyed dilution

and Himalayan albinism (Wright, 715).° Soon after these forms

were found, Whiting and King (’18)* reported ruby-eyed dilu-

tion in rats, an allelomorph of both color and albinism. While

Whiting and King used the same symbol, ¢,, for this gene that

Wright used for ruby-eyed dilution in guinea pigs, they point

out that the somatic effect is somewhat different. The two cases

of ruby-eyed dilution may not be identical, for it is conceivable

that numerous dilution effects in the color-albino series are pos-

sible. Ordinary color dilution still remains to be found in rats,

in order to make up a series of color allelomorphs as elaborate

as in the guinea-pig. Similar examples of apparently parallel or

identical mutations in closely related groups can be shown in

other animals, as in the color varieties of the horse and the ass,

for example, and in the case of the Drosophila species.

In an effort to homologize the genes which affect the quality

and distribution of hair pigment in the mouse, rat, guinea pig

and rabbit, I was impressed by the dearth of color allelomorphs

in mice, such as cause the various grades of dilution in the rat

8 Loe. cit.

4 Loc. cit.

5 Patterson, J. T., 1920, Science, N. S., Vol. 52, p. 249.

6 Wright, S., 1915, Am. Nart., Vol. 49, p. 140.

* Whiting, P. W., and King, H. D. 1918, Jour. Exp. Zool., "Vol. 26, p. 55.

No. 640] SHORTER ARTICLES AND DISCUSSION 471

and guinea pig. It appeared that similar changes in the mouse

were possible and could be found, if persistently sought. With

this general thought in mind, I attempted by corresponding with

fanciers et cetera to locate dilute mutations in mice, thinking

that the dark red-eyed yellow mutation (parallel to the type

found in rats) or the dilute forms allelomorphie to color and

albinism (parallel to the guinea-pig and rat series) might be

possibilities. My correspondence with fanciers brought no re-

sults, but a dilute mutation appeared from a rather unexpected

source. On August 31, 1920, Mr. J. E. Knight of Weldon,

Illinois, who exterminates rodents from corn cribs, poultry houses

and the like, and who has much opportunity to examine a large

number of these mammals, brought to my laboratory a young

male mutant mouse which he had captured in a corn crib, located

on a farm seven miles from the nearest town. This animal on a

first and cursory examination, gave the appearance of being an

ordinary black-eyed white in which the hair was apparently

very ‘slightly stained or dirty. Realizing that such a form would

mean a double (and therefore much more improbable) mutation

from the wild, in which both dominant and recessive spotting oc-

curred simultaneously, I made a more careful examination about

one month later and found that the hair had become darker.

I have since learned that it is characteristic of this form to be

practically white on the first pelage, but the dorsal hair eventu-

ally acquires a brownish shade,—a little lighter than an ordinary

pink-eyed brown with a slight dull yellowish cast. There is no

clear evidence of an agouti pattern, the base of the hair being

light and the apical portion pigmented. The ventral surface is

almost white, at least in the presence of agouti. The eyes at

birth are somewhat less heavily pigmented than the wild. This

difference persists for some time, but when the mutant is full

grown I am not sure I can distinguish the eye from the wild

type. Dark pigment is quite pronounced in the skin of the ears

and scrotum, in which respect this mouse differs from the ruby.

eyed rat. The dark eyes and yellowish tinge in the hair at first

suggested that the mutation was similar to the dark red-eyed

yellow rat. Recent matings have, however, demonstrated quite

clearly that.it is a third allelomorph in the color-albino series, and

may therefore be homologous to the ruby-eyed dilute rat which is

allelomorphie to color and albinism.

While the hair of the mutant mouse is lighter and the eyes are

472 THE AMERICAN NATURALIST [Vou. LV

apparently darker than the ruby-eyed dilute rat, the genetic

behavior agrees quite closely in both forms. Nevertheless it is

hardly safe to insist that these two mutations are identical, for

there may be numerous possible grades or conditions of the

color gene. We are also unable to prove that they are different,

for the genes may be identical, but simply give different somatic

effects since the residual inheritance in the two forms can not

be the same. If a new dilute type of mouse can be found which

is more like the rat in both genetic behavior and somatic ap-

pearance, then we shall be able to state with more assurance that

‘the present mutation is not identical with the dilute rat. Until

that event occurs, we can only regard these dilute color mutations

in the rat and mouse as samples of a series of possible mutations

in the color gene (ef. the red, white, eosin, cherry, et cetera series

of multiple allelomorphs in Drosophila melanogaster). A sim-

ilar interpretation applies in any attempt to homologize the mem-

bers of the guinea-pig series with those of the rat or the mouse

series.

Three types of crosses between the mutant male mouse and

other color varieties have been made as follows:

1.-Mated to homozygous blacks, all of the F, offspring were

wild gray. In the F, the mutant and other expected forms

segregated out.

2. Mated to pink-eyed spotted brown, all of the F, offspring

were wild gray. In the F, the expected forms occurred including

the mutant type. We have not yet had opportunity to identify

the mutant form when homozygous for pink-eye.

. 3. Mated to albinos, all the F, offspring were of the mutant

type, that is, they are white at birth with eyes rather less heavily

pigmented than the wild type. As they grow .older, the hair

soon approaches the mutant color type, but I can not yet state

whether the mutant hair color is incompletely dominant, as in

the case of the ruby-eyed dilute rat and guinea pig.

From present indications, this new dilute mutation is certainly

recessive to color, and I am inclined to believe it will prove to

be incompletely dominant to albinism. The three genes (color,

color dilution, and albinism) probably form a series of triple

allelomorphs. I shall designate these genes by the symbols C,

c4, and c, respectively. The mutant mouse is homozygous in

agouti, black, dark-eye and self pattern, and therefore represents

a single factor difference from the ordinary wild type, from

No. 640] SHORTER ARTICLES AND DISCUSSION 473

which it arose. The new gene should prove to be linked with

dark-eye, like its allelomorph, albinism. Since it occupies, in a

scale of dominance, an intermediate position between color and

albinism, the mutant should give a coupling series when mated

to albinos carrying pink-eye ( Fi = e

but a repulsion series when the same mutant is mated to pink-

C p

GENITAL ORGANS OF HERMAPHRODITIC FUR SEALS

THE resumption of commercial killing of the surplus and use-

less males of the fur-seal herd resorting to the Pribilof Islands

of Alaska furnished an opportunity to study the life history and

anatomy of this group of mammals. After a lapse of six years,

killing was begun again in 1918 by the Bureau of Fisheries.

Upwards of 33,000 males, mostly young, were secured and their

skins preserved for sale by the goyernment. Two hermaphro-

ditic animals were killed among this large number and the

writer had an opportunity of examining the sexual organs of

both. Such abnormalities of the species have not previously

been recorded and since they are rarely found among mammals

of any species it seems desirable to note the occurrence with a

brief description. The organs of both animals have been deposited

in the United States National Museum.

Normally the female Alaska fur-seal has two kidney-shaped

ovaries located just forward of the pelvis and loosely invested in

the folds of connecting ligaments. Blood vessels, ureters, fal-

lopian tubes and uteri are attached to the same folds. The

ureters pass above the genitalia but bend down below to reach

the tip of the bladder. The uterus is bicarinate, an ovary being

attached to the distal end of each horn. It is pretty well deter-

mined that each side functions alternately every other year. The

horns unite in the median line and the vagina continues to the

exterior, a distance of about 20 em. Attached to the lower side

of the vagina is the pear-shaped and very muscular bladder.

The urethra leading therefrom is deeply embedded in the mus-

cular walls of the lower vagina a as it passes to its oo of dis-

charge near the exterior.

474 THE AMERICAN NATURALIST (Von. LY

The first of the hermaphroditic animals to be described was

found at the Northeast Point hauling grounds of St. Paul Island

on August 6, 1918. It was four years old and thought to be a

male by the external characters of the head. It was not dis-

covered to be bisexual until skinning had started and the mam-

mary glands were found fairly well developed and containing

a small amount of milk. This is one of the first characters to be

noted when a female has been killed. As the native skinner

thought this had happened, my attention was immediately called

to the matter. Two or three dozen females are unavoidably

secured when large killings are made in the manner followed in

1918 and it was thought that this was one of these unfortunate

accidents. But when the penis opening on the abdomen was

seen in the usual place is was known that a freak had been

` found and its organs and skull were preserved.

The right ovary was found to be smaller than normal and pear-

shaped instead of flatly oval and it was entirely divested of the

usual covering membrane. Its fallopian tube was thicker and

fleshier than usual. But the left ovary was much larger than

normal and the membrane was firmly. attached all over its sur-

face by adhesions. Both horns of the uterus were normal in

shape but smaller than usual in a four year old female. The

walls of the vagina leading backward from the uterus were

extremely heavy and firm. The opening grew smaller and

smaller posteriorly until it reached the point of junction of the

vagina walls and the penis. Then it followed the latter organ

forward on the ventral side as a small duct. The opening to the

exterior was near the distal end of the penis.

There were no testes; rudiments even could not be found.

But the penis was well developed and in the normal position.

The os penis was only about one fourth as large as would be

found in a male of equal age. The cartilaginous continuation

of this bone and the continuation of the walls of the vagina

were one and the same.

The urinary system was normally developed, the bladder being

attached to the vagina.. The urethra followed the penis forward

on the side opposite from the duct of the vagina. The muscles

for the retraction of the penis were well developed.

It would seem that the presence of the female reproducing

organs would preponderate in affecting other characters of the

animal such as the skull. (This is widely different in the two

No. 640] SHORTER ARTICLES AND DISCUSSION 475

sexes.) But such was not the case. Although possessing de-

cided features of both sexes the skull resembles, far more, one of

a male of the same age.

The other specimen found in 1918 was very much less in-

teresting. It was secured on Lukanin field, St. Paul Island. The

organs were brought to me after the killing was over in a some-

what mutilated condition but the relations seemed to be about as

follows:

There was a pair of testes, apparently in the usual position

of ovaries. The spermatic cords united above the neck of the

bladder and seemed to discharge through a large blind glandular-

walled pouch. This latter was taken to be a pathological vagina

and was all that remained of the female system. The bladder

was normal in size and the urethra passed straight backward

from it as in the female. A rudimentary penis two centimeters

long with a minute bone projected posteriorly beneath the anus

and the opening of what was taken to be the vagina. The

urethra discharged through this penis. This animal was prob-

ably a two year old, but its skull was not preserved and the ex-

ternal characters were not reported to me.

An old native sealer once told me that he had seen a half male

and half female seal about five or six years old. It was as large

as a male of that age but had the beautiful soft pelage of the

female or the young male. In this connection it is worth while to

recall an attempt which was made with poor success in 1896 to

castrate pups. It might be that he saw one of these but it is

more likely that he saw a hermaphrodite. It has been rec-

ommended that castration be attempted on animals older than

pups, say two year olds. If successful it is probable that they

would develop to the size of the full-grown male but would re-

tain the very valuable fur of undeveloped males or of females.

As the males are when full grown four to six times as large as

the females, the pelts should be proportionately increased in

value.

G. DALLAS HANNA

CALIFORNIA ACADEMY OF SCIENCES,

SAN FRANCISCO, CALIF.

476 THE AMERICAN NATURALIST [Vor. LV

INHERITANCE OF BELTING SPOTTING IN CATTLE

D SWINE

Ir has long been noted that when two belted individuals are

bred together only part of their offspring show true belts.

Taking the Dutch Belted breed in cattle, although the new

belted Galloway (Ashton) may later serve just as well, we find

that in practically all pure-bred herds there appear individuals

with imperfect belts and more often those that are pure black.

Kuiper has furnished us with the most promising results so

far. He shows that the characteristic markings of the Laken-

velder or Dutch Belted breed may be obtained by crossing within

the breed or by crossing with spotted cattle. In his experiments

a belted bull was bred to 55 Holstein-Friesian cows and produced

as offspring 27 belts, 24-25 self-black, and 3-4 spotted. The `

identity of one calf was doubtful.

To explain these results he takes two pairs of allelomorphic

factors L-l for belt, epistatie over E-e for self, and a repulsion

between L and E in the reduplication series 1-7-7-1. A fairly

high correlation exists between white feet and wide belts, This

correlation agrees closely with Walther’s work on horses.

Kuiper’s work may be criticized on the fact that he has no

definite grounds for assuming the presence of allelomorphic

factor pairs. He does not assume a factor for white spotting

that will take in all parts of the animal.

In crosses between the single colored reddish-brown Nether-

land cattle (Richardson) a very few self-color individuals

were produced. Crosses between the reddish-brown Netherland

cattle (Kiesel) and Holstein-Friesians produced in the F,, 90

well spotted, 84 medium spotted, and 6 self-colored individuals,

showing that the dominance of either character was not complete.

Crossing the F, individuals together gave 22 self-colored and 29

spotted. These results show that the Holstein-Friesian markings

are of a heterozygous type and Holstein-Friesians when bred to

Dutch Belted gave practically a 1-1 ratio. Assuming that all

possible factor combinations were made this would easily prove

that the belting in the Dutch Belted breed is a simple hetero-

zygous condition and explains the appearance of offspring other

than belts when belts are bred together.

In Hampshire swine there seems to have been considerable

selection within the breed. Originally they were white or black

and white (Youatt). Later, selection brought them to their

No. 640] SHORTER ARTICLES AND DISCUSSION 477

present color either black or black with a white belt (Day).

White spotting other than belts appear, also an excessive white

belting condition covers all but the extremities of the ears and

the tail. Spillman states that ten per cent. of the progeny

of registered individuals are without belted areas. He supposes

two types of belting (a) homozygous, occurring very rarely, and

(b) heterozygous, occurring as the common type. In crosses of

other breeds on Hampshires (Simpson) (Severson) the spotting

condition proved to be heterozygous, for when bred to recessive

colored breeds the ratio of belts to non-belts was t 11-10 (ex-

pected ratio 10.5-10.5).

Summing up all evidence so . far on the inheritance of belting

spotting one would be safe in saying that this character is due

to a single heterozygous factor pair Ss for white spotting.

G. B. DURHAM

LITERATURE CITED

Ashton, John.

1920. The Belted Galloway. Breed Gaz., 78: pp. 1173, 1207, 1217.

1913. Produetive Swine Husbandry.

omens

1913. Uber RETEA shai ee beim Rind. Zeit fiir Ind. Abst.

und Verbung

Kuiper, K., Jr.

1920. Researches on Color and Markings in Cattle. Based on Experi-

ents by R. Houwink Hzn. Genetica, 2: 137-161.

Richardsen, —.

191 Heredity of Color in Cattle. Deut. Landw. Tierzucht., 18, No. 6,

pp. 61-65

Ben By O,

1919. Color Inheritance in Swine. Jour. Her., 8: 170-181.

Spillman, W. J.

1907. Inheritance of the Belt in Hampshire Swine. Science, N. S.,

25: 841-549,

Walther, Ad. R.

1913. i Vererbung unpigmentierten Haar und Hautstellen bei Rind

d Pferd als Beispiele transgressive fluktuierenden Factoren.

Zeit für Ind. Abst. und Verbungsl., pp. 1—43.

STANDARDIZED MICROPHOTOGRAPHY

THIRD CONTRIBUTION: THE EXPOSURE FACTOR

THE great convenience to be derived from the use of only two

tables of factors is so apparent that anybody who is at all

familiar with microphotographie work should not for a moment

478 THE AMERICAN NATURALIST [Von. LV

hesitate to spend some time m the preparation of such tables.

It is, of course, natural that differences in the source of light,

optical equipment and ray-filters necessitate the preparation of

new tables. Our own outfit being one of the standard Bausch

and Lomb apparatus with a Zeiss microscope and Cramer ray- —

filters, it occurred to me that those possessing a similar equip-

ment might be spared the tediousness of preparing an exposure

table if our table of exposure factors were published with such

instructions as would permit of identical arrangement of the

apparatus. Moreover, it will be remembered that these factors

are dependent upon the source of light, substage position, mag-

nification, and numerical aperture of the objective, and that

other conditions being identical, exposure varies as the square of

the numerical aperture.

A word must be said about the use of the substage condenser.

The position of the condenser indicated in the table is such as to

give the greatest detail without apparent bad effect on definition.

As is well established, the depth of focus in a microscopic ob-

jective depends upon its numerical aperture. The greater the

latter, the smaller the depth of focus. Since definition increases

with numerical aperture, one has to sacrifice the one or the other.

From a practical point of view depth of focus is often more

desirable than perfect definition. The numerical aperture of an

objective may be conveniently cut down by increasing the dis-

tance between the substage condenser and the objective. This

can be done with safety. only to a certain point, which I called

the optimum, and beyond which definition is visibly impaired.

I hope to be able later to return to this subject in greater detail.

Meanwhile, we may state as a general rule that the higher the

magnifying power of an objective, the sooner the optimum will

be reached. With other words, the substage may be lowered a

great deal more in low-power than in high-power objectives.

In the following table the position of the substage is indicated

in millimeters, assuming that it is at zero when moved up as far

as it will go. In the Zeiss microphotographie stand the substage

condenser has a numerical aperture of 1.40, and when placed at

zero the surface of its upper lens is still 0.8 mm. below the

surface of the microscopie table.

The source of light for which the table of exposure factors

holds good is the Bausch and Lomb Microprojector No. 4301

with hand feed are lamp and rheostat for 4.5 amperes, 110 volts,

No. 640] SHORTER ARTICLES AND DISCUSSION 479

D.C., placed so that the distance between the uprights carrying

the projector and the microscope table is 20 inches, while half-

way between the projector and the microscope a water cell is

placed for the absorption of heat rays. The projector has a

movable condenser of its own. The position of this condenser

influences the intensity of light. Our table is made for such a

position of this condenser that the light on the focusing screen

is brightest. It is not the same for every objective and must be

found by experimenting, with the aid of an assistant, who moves

the condenser while the observer watches the field. For this

experiment the microscope substage is best lowered about 2 mm.

to insure even illumination. When the position of the condenser

of the projector has been found and marked on the mounting,

TABLE OF EXPOSURE Factors FoR ORTHONON PLATES WITHOUT Ray-FILTER,

UsED WITH A B. & L. Open Arc D.C. 110 VOLT MIĊROPROJECTOR

AND AN ABBE SUBSTAGE CONDENSER WITH N.A. 1.40.

| i |

» |g22 (asg [aay lee (ase lage | ase | :

$ fos |dos (Ese [Ess [ESE |ESe | eos B EL| Zeis

®@ j/a¢8 |546 [des (edo (448 |45 | 445 | mat E. \chrom’t

¢ | Srey lee psa ge ekg eR 2 F. 1.9 E. F. 2

= 21%2Z_, | oxo Ze | 822 Zo: Oe CNOA LN A

z TER Si KEF Hag. Bee fae mf pt REE a

25 | 0.0005

0 | 0.0007

40 | 0.0013

-0020

75 | 0.0045 | 0.013 | 0.011

100 | 0.0080 | 0.0: 0.020

125 0.037 -031

150 0.0 0.045

175 0.073 | 0.061

0.096 .080

250 0.150 | 0.125

300 0.216 | 0.180 .| 0.108 | 0. 0.048 | 0.040

350 0. .245 147 | 0.122 | 0. 0.0

400 0.384 | 0.320 | 0.192 | 0.160 | 0.085 | 0.071

500 0.300 | 0.250 | 0.133 | 0.111 0.040

0.360 | 0.192 | 0.1 0

700 0.588 | 0.490 | 0.262 | 0.218 0.078

800 0.351 | 0.284 0.102

0.432 | 0.360 0.130

1000 0.532 | 0.444 0.160

1250 0.250

1500 0.

1750 0.490

2000 0.640

2500 1

3000 1.440

480 THE AMERICAN NATURALIST [Vou. LV

the experiment need never be repeated again. The table itself

applies only to an Orthonon plate used without ray-filter. For

use with Cramer’s rayfilters the factors given in this table must

be multiplied by the factors shown in my first paper.

The so-called ‘‘ Pointolite ’’ is an excellent source of light for

low powers. It requires about 5-10 times longer exposures. In

our laboratory we use the Complete Illuminating Apparatus with

D.C. open arc. After some experimenting it was found that the

most satisfactory arrangement of this illuminating system is

when the are is moved forward to within 214 inches from the

first condensing lens, a water cell placed between second and

third condensing lenses, in contact with the mounting of both,

and the distance between the third and fourth condensing lens

fixed at 15 inches, while the distance from the latter to the

microscope table is 714 inches. Under these conditions the ex-

posure is double that obtained with the microprojector for

which the table in this paper is given. I omit giving a separate

table for this system, because any one may obtain it in a moment

by simply doubling the values given here. But it must be re-

‘membered that change in the arrangement of the illuminating

system, no matter how small the deviation, will result in wrong

values.

ALEXANDER PETRUNKEVITCH.

YALE UNIVERSITY

THE

AMERICAN NATURALIST

Vou. LV. November—December, 1921 No. 641

EXPERIMENTAL STUDIES ON THE DURATION

OF LIFE

I. [INTRODUCTORY Discussion oF THE DURATION oF LIFE IN

DROSOPHILA ?

RAYMOND PEARL AND SYLVIA LOUISE PARKER

Sucua quantitative knowledge as exists of fundamental

principles in the general biology of the duration of life

has, in the main, been derived from an examination by

purely statistical methods of human mortality records.

Of course a good deal of information about the biology

of death and duration of life of a general and non-quanti-

tative character has been gained from experimental work

on lower organisms. This literature has recently been

reviewed by one of us (Pearl (1) to (7) inclusive). But

the outstanding fact is that most of the existing quantita-

tive data about duration of life are purely statistical, and

derived from man as material.

The statistical method of acquiring knowledge of nat-

ural phenomena has a number of distinct and important

limitations (cf. Pearl (8)). It is the settled policy of this

department to check every conclusion drawn from purely

statistical methods by an independent experimental in-

vestigation of the same problems, wherever in the nature

of the case this is possible. Most problems of human

vital statistics can not, in the nature of the case, be in-

vestigated experimentally, in any direct way with man

himself as material. Probably this is chiefly the reason

why all of the immense mass of data collected, and work

done upon vital statistics has contributed so little in the

1 Papers from the Department of Biometry and Vital Statisties, School of

Hygiene and Public Health, Johns Hopkins University, No. 45.

482 THE AMERICAN NATURALIST [Vou. LV

way of general principles to the science of biology. In

outlining the plans of the department at the time of its

inauguration provision was made, as a major element in

the whole organization scheme, for experimental work on

the duration of life, to parallel as closely as possible, in

respect of its problems and foci of interest, the statistical

work of the department. The present paper is the first

of a series which will appear dealing with the exper-

imental side of our work.

Originally it was planned to use mice as the material

for experimentation on the duration of life, and a large

and flourishing colony was bred up in accordance with the

most critical genetic standards for experimental material.

Just as the colony was ready to start definitive exper-

imentation with, an accident completely destroyed it.

It was then decided, after advising with a number of

persons, notably Professor T. H. Morgan and Dr. Jacques

Loeb, to take up Drosophila as material for the extensive

program of experimental work which we had planned.

This organism has the great advantage over any other

which could be used, that its genetic behavior and poten-

tialities. are more thoroughly understood than those of

any other animal, thanks to the epoch-marking researches

of Morgan (9) and his students. It has the further great

advantage that under certain conditions, which we now

rather clearly understand, its duration of life, both in

respect of means and of the ls or ds distributions of a life

table, is extraordinarily like that of man, with one day in

the life of the fly corresponding roughly to one year in

the life of man.

The first paper in the series aims to present, as a back-

_ ground of reference for further contributions, the follow-

ing. essential items:

1, A brief review of what has been noted by previous

workers regarding duration of life in Drosophila, and

other insects in so far as the observations are quantita-

tive in character.

2. The details of our material and methods of exper-

No. 641] THE DURATION OF LIFE 483

imentation, which are critically standardized and have

been used in the work whicli will be described in sub-

sequent papers. ;

3. The general form and characteristics of the mortal-

ity curves of Drosophila, presenting mortality tables for

certain strains.

4. The influence of certain phases of the experimental

technique employed upon the results.

Specific problems regarding the duration of life will be

presented and discussed in the subsequent papers in the

series.

LITERATURE

The earliest mention we have found of observations on

the duration of life of Drosophila is a casual reference

in a paper by Moenkhaus (10) published in 1911 in which

he makes the following statement in connection with egg

counts: ‘‘ We have kept females alive 153 days.’’ There

are no details of any kind, as to conditions or numbers

involved.

The first paper to make more than casual reference to

the duration of life of Drosophila is a paper publishea in

1913 by Hyde (11). In studying fertility and sterility in

different strains of flies he found two strains which dif-

fered to a marked degree in respect of length of life, and

made crosses to study the behavior of the shortered

length of life of the mutant ‘‘ truncate ’’ in heredity. ilis

numbers are small, but they show the characteristic in-

creased vigor of F, hybrids. The shorter average age

of the lumped F;,’s indicates that there have segregated

out in the F, generation some short-lived flies. His data,

however, do not give, or allow us to get, separate averages

for the truncate F,’s and the normal winged F’,’s.

His data are summarized in Table I.

In 1914 Baumberger (12) published a paper in which

he gives data on the length of life of different orders of

insects without food as affected by different constant tem-

peratures, and by exposure to two different temperatures.

Since the insects were caught in a net as imagoes the total

484 THE AMERICAN NATURALIST [Vou. LV

longevity is not known, so that the results have little sig-

nificance from the standpoint of exact studies. The 359

‘insects had at 72° F. an average longevity of 4.8 days with

a maximum of 15 days, at 62° F. an average of 6 days

with a maximum of 23, and at 42° F. an average of 10.9

days with a maximum of 39. For the second part of the

experiment 184 larve of the oak-tree moth were used.

The results are too conflicting to allow one to draw any

definite conclusions.

TABLE I

HYDE’S DATA ON INHERITANCE OF DURATION OF LIFE IN DROSOPHILA

Truncate

In- xX | Recip-

bred | Trun- Inbred ro-

Type of Flies Wild | cate F, Q eal | Total

Fe F2

No. of Flies. 191 272 42 128 89 722

Mean duration of life | po and 9 | 37.4 | 21.4 | 47.0 29.5 POG he R

in days. ofl 40.5 | 26.9 | 47.8 32.8 DET a.

Q 34.5 | 18.5 | 46.4 25.9 vg Ge Ni E

In 1915 appeared Lutz’s (13) paper on natural selec-

tion in which he finds in each sex a slight negative correla-

tion between the length of adult life and the duration of

the embryonic periods. The distributions which he has

for normal length of adult life with varying temperature

give the 250 d's an average duration of life of 36.3 days,

and the 263 9s an average of 28.9 days. He also gives

distributions in hours of duration of life of flies which

were given water but no food, and the correlations of

duration of life of these starved flies with wing measure-

ments.

During 1916 and 1917 Loeb and Northrop (14-16) pub-

lished a series of papers on the effects of food and tem-

perature on duration of life in Drosophila. The first pre-

liminary paper in 1916 gives the duration of life of cultures

of Drosophila in water and in cane sugar at temperatures

from 28° to 9° C., showing a temperature coefficient for

the duration of life of about the order of magnitude of

that of chemical reactions, namely of about 2 for a dif-

No. 641] THE DURATION OF LIFE 485

those found by Hyde and Lutz because, of course, of the

inadequate food. At 19° C. the culture in water had an

average duration of life of 4.1 days and those in 1 per cent.

cane sugar of 12.5 days. In 1917 the experiments were

repeated, using sterile flies on 2 per cent. glucose agar

which was found to be a more adequate food. Similar

results were obtained, getting a similar coefficient for the

duration of the larval and pupa stages, and finding that

the ratios of the duration of the three different stages re-

mained approximately constant for the different tem-

peratures. The averages here of the life of the imago

are more of the order of those found by previous workers,

. 228 flies at 30° lived an average of 13.6 days, 70 flies at

25°, 28.5, and 49 flies at 20°, 40.2. Later in 1917 thev

published another paper in which they give 92.4 days as

the average duration of life of 143 flies at 15°, and 120.5

days of 105 flies at 10° C., together with the frequency

distributions from which the averages were obtained.

They also present results with different food mixtures,

and for the two sexes separately, finding that isolated

males live a little longer than isolated females, or than

the males when mixed with females.

In another paper in the same issue Northrop (17)

gives the results of some experiments undertaken to de-

termine the effect on the duration of life of the imago

of prolonging the life of the larva by inadequate feeding

(omitting yeast for different lengths of time). In this

way the embryonic periods were prolonged from 8 to 17

days, but the duration of life of the adult remained the

same in every case, ranging between 10.5 and 11.9 days

at 27.5° C., at which temperature the four experiments,

involving 644 flies, were performed.

In a recent paper Arendsen Hein (18) gives a few

observations on duration of life in the meal-worm Tene-

brio molitor. Thirty-two male beetles lived an average of

60 days, with a range from 39 to 113, and 32 females

averaged 111 days, with a range from 89 to 132 days.

486 THE AMERICAN NATURALIST (Vor: LV

MATERIAL AND METHODS

The flies furnishing the data set forth in this paper be-

longed to five different basic laboratory stocks or strains

of Drosophila melanogaster. Four of these stocks were

obtained from Professor T. H. Morgan in December,

1919, and have been bred continuously in this laboratory

since that time. The original individuals of the fifth

stock were collected by one of us (R. P.) as wild flies at

Eagle Point, Lake Memphremagog, Vermont, in the sum-

mer of 1920.

The stocks may be listed as follows:

1. Old Falmouth. Wild kiz fly, long bred in Morgan’s laboratory.

ore inbred than

2. New Falmouth. vu type fly bred for about 6 months in Mor-

gan’s laboratory before we got our sample of it.

3. Sepia. A mutant stock carrying one third chromosome mutation,

sepia eyes, in homozygous form. Other characters wild types.

(Morgan.)

4. Quintuple. A synthetic stock, carrying five EE chromosome

mutations, each in homozygous form, as follows: Purple,

arc, speck, vestigial, and black. Other characters wild type.

( Morgan.)

5. Eagle Point. Wild type collected in summer of 1920, and since

bred in this laboratory.

An account of the second chromosome mutations men-

tioned will be found in Bridges and Morgan (19) and

Sturtevant (20). The discovery of the mutation sepia is

noted by Muller (21).

The stocks are carried along in the laboratory in pure

mass cultures in half-pint milk bottles. Those in an ex-

periment on duration of life are tested relative to this

character in one ounce shell vials.

The flies are all kept on a standard food mixture made

up fresh each day, according to the following method:

For each 100 c.c. of water add 2 grams of agar-agar. Boil the

agar and water until the agar is thoroughly dissolved. For each

100 c.c. of solution add 100 grams of ripe peeled mashed bananas.

Boil five minutes. Pour into bottles which have been well heated in

oven (or sterilized in autoclav). In the breeding bottles pour a layer

3/4 inch deep; in duration of life bottles a layer 1/2 inch deep. When |

No. 641] THE DURATION OF LIFE 487

the food has partly cooled sprinkle on top of food the smallest possible

amount of pulverized dry magic yeast (shaken from a can with one pin

hole in cover). Put in breeding bottle a folded square of filter

paper, and stopper with cotton batting.

The purpose of the filter paper is to furnish a dry

place for the larve to crawl up and pupate on, and also

to absorb some of the excess moisture which often forms

' on top of the food from the growing yeast. Filter paper

has not been used in the small duration of life bottles,

since no young are pupating there, and since it furnishes

too many hiding places for the flies in the frequent trans-

fers which have to be made in the duration of life tests.

Excess of moisture on top of the food may become a

source of error in duration of life experiments, because

flies may drown in a small drop of water. . Throughout

our work we have been constantly on guard against this

source of error and have tried a number of plans, with

varying degrees of success to eliminate it entirely. Some

of these experiments will be reported on in detail later

on in this paper. In general it may be said here that

this source of error from flies drowning need never be a

significant one if due precautions are taken. We know

that it has not been in our work.

Occasionally the yeast becomes too active at the edge

of the food and causes the whole food mass to rise in the

bottle. In an attempt to eliminate this accident yeast in

dilute solutions was added to the boiled bananas and agar

and was sprayed on top of the food. The results were

not particularly favorable. The pulverized dry yeast

added in the most minute quantity possible is the most

satisfactory standard method yet found. We expect to

continue attempting to get the food conditions more and

more nearly ideal and identical in every experiment, but

we feel reasonably certain that in all of the experiments

we shall report, even including the very first in point of

time, the precautions taken to standardize food and to

guard against accidental death were sufficient to insure

statistical accuracy in the results. Whatever environ-

488 THE AMERICAN NATURALIST [Vou. LV

mental differences in respect of food have existed in our

experiments have been randomly distributed among the

different groups in any given experiment. We have had

very little trouble at any time with moulds in the cultures,

the frequent transfers in the duration of life bottles in

an experiment preventing them from getting any start.

The stock bottles holding reserve stocks of flies have

been kept at the varying temperature of the room, but all

experimental flies have been kept in electric incubators

at 25° C., in which recording thermometers have been

placed to insure that no fluctuations of temperature have

occurred without our knowledge. All the experiments on

duration of life and their results recorded in this paper

have been carried through at the constant. temperature

of 25° C. We have settled on this as a normal for this

- particular element of the environmental complex.

During the first year of the experimental work no at-

tempt was made to keep the different generations sep-

arate in the stock bottles, the process being merely to

keep enough bottles (generally 4) of each stock to insure

always having pup and newly emerging flies on hand

for any matings and experiments to be started. Each

week all the flies old and young together from the oldest

bottle of each stock were transferred to a fresh bottle. In

this way each bottle was kept 4 weeks and there were al-

ways on hand bottles with flies in all stages of develop-

ment.

In January, 1921, it was decided that it would be desir-

able to keep the generations separate in the stock bottles.

All stock bottles were emptied on January 11, and flies in

the stock bottles on January 14 were arbitrarily called

generation O. From that time on the procedure has been

to empty out all the parent flies from each stock bottle 7

days after the bottle was started (when there are usually

a large number of larve and some pupe formed). The

bottles are then left for 7 days longer, during which time

enough flies emerge to start a fresh bottle for the next

generation. Several bottles are kept of each stock, as

No. 641] THE DURATION OF LIFE 489

before, to insure always having on hand newly emerging

flies with which to start experiments.

When any experiment is to be started flies are taken

from the stock bottles and etherized within 4 hours of

emerging (usually sooner), before the wings have un-

folded, so that they are surely virgin. Matings are made

up as desired, putting the mated flies in half-pint milk

bottles in the incubator. The parents are taken from

the mating bottles in 8 or 9 days (before any young begin

to emerge), and removed to a second mating bottle if a

larger sample of progeny is desired than can be obtained

from one bottle. As the offspring begin emerging they

are shaken out every morning from the mating bottle

to a small shell vial. Thus all the flies in a small bottle

are the same age, and are properly labelled with mating

number and date of emergence. Then every morning all

these small bottles are looked over and those with dead

flies separated out. After all have been looked over, the

live flies in the bottles which have dead ones are shaken

across to fresh bottles, the dead flies taken out and sexed,

and all the pertinent facts as to duration of life, ete., re-

corded on printed blanks, from which the records are later

(when all the flies of an experiment have died off) coded,

punched on Hollerith cards, and sorted and tabulated by

Tabulating Machine Company electric sorting and tabulat-

ing machines. Flies from any small bottles which are not

changed (because of dead flies) within five days of the last

previous transfer are transferred on the fifth day to

fresh food. The physical manipulation is too great with

the numbers we desire to use to admit of changing all the

bottles every day, which would be the ideal way to keep

food conditions absolutely identical for all the flies.

Changing every five days keeps them approximately so.

We desire to record our indebtedness to Mr. James

Krucky, technical assistant in this work, for his pams-

taking care and fidelity to the highest ideals of exact

experimental work.

In the work discussed in this paper no attempt has

490 THE AMERICAN NATURALIST [Vou. LV

been made to keep the flies in aseptic culture as had been

done by Loeb and Northrop (eve. cit.) and other workers.

Our choice in the matter has not been dictated by tech-

nical difficulties, which are not great, but has been deliber-

ate. Aseptic life is by no means normal life for Droso-

phila. Normally it is as loaded with a bacterial flora as

we are. It was felt that in the beginning it would be well

to establish norms of duration of life for normal life

conditions. Later we expect to make a special study of

duration of life under aseptic conditions.

Duration of life in this work with Drosophila is always

measured in days, and all of our records relate to dura-

tion of adult or imaginal life. No account is taken in any

figures of the larval or pupal stages. The reason for this

convention is first accuracy and second convenience. It

is far more difficult to measure accurately either larval

or pupal duration of life than it is imaginal. And from

the point of view of these studies nothing significant in

principle is lost by dropping these early stages, so far

as we have been able to discover, either from the lit-

erature or experience with the flies.

Mortauity CURVES

The most exact and comprehensive manner in which

the facts about the duration of life in any organism can

be presented is by means of life tables, of the type used for

many years by actuaries in their work. The biologically

essential features of a life table may be mentioned here

briefly for the benefit of biologists not immediately famil-

iar with the development of actuarial science. A com-

plete life table includes, inter alia, the following items:

1. The number of individuals surviving up to each of

the ages £o, %ı, ete., out of a given number (1,000 or 10,000

or whatever number one chooses) assumed to have started

life together at exactly the same instant of time. These

survival frequencies taken together constitute what is

technically known as the l» line of a life table.

2. The number of individuals dying within any short

No. 641] THE DURATION OF LIFE 491

interval of time, say between v and w+ L o dis

ls —le+ı. These frequencies of death taken together

constitute the so-called dz line of a life table.

3. The death rates at each time (or age) «; i.e., the

ratio of the deaths between time x and x +1 to the sur-

Vivors at the time x. These observations together con-

stitute what is known as the t line of a life table q: =

(le — le + 1)/le.

4. The curtate expectation of life of individuals at a

given age x. This is the mean or average after life time of

all those individuals alive at age x neglecting fractions of

the x interval. These observations together constitute

the ex line of a life table.

aes lan + less i læ+s + E

These simple definitions state with entirely sufficient

accuracy for present purposes the significance of the con-

stants which we shall present. Any one wishing to go

more particularly into details of actuarial methods will

find a useful elementary introduction in Henderson (22)

or Dawson (23).

It is our purpose to present here life tables for four

groups of flies, to serve, first, to show the general laws

of mortality in Drosophila as compared with man, the

only other organism for which we have extensive and ex-

act life tables; and second, as a normal base for compar-

_ ison in experimental work on Drosophila to be reported

in subsequent papers.

The ‘groups for which complete tables are presented.

are these:

1. Long-winged males. This table includes all our data

up to June, 1921, on normal (7.e., not experimentally mod-

ified) duration of life of male Drosphila individuals at 25°

C. belonging to the following stocks (cf. p.486 supra): Old

Falmouth, New Falmouth, Sepia and Eagle Point. In

these stocks all the individuals have wild type wings,

hence the designation ‘‘ long-winged.’’

492 THE AMERICAN NATURALIST [Vou. LV |

2. Long-winged females. The corresponding table to

1, but for females.

3. Short-winged males. This table includes all our data

up to June, 1921, on the normal duration of life at 25° C.

of males belonging to the Quintuple stock. These flies

eariy the wing mutation vestigial; hence the designation

“ short-winged.’’

4. Short-winged females. The EROM a ie table to

3, but for females.

We have tried a number of different plans for the

graduation of these tables, and wish to acknowledge

gratefully the helpful suggestions of our colleague in the

department, Dr. Lowell J. Reed, in connection with this

phase of the work. It was first found that a rather

satisfactory result could be obtained by fitting a logarith-

mic parabola of the type

y =a + ba + cx? + d logg

to the qz data. Working from this as a basis we finally

decided that, as a practical matter, results on the whole

most satisfactory could be got by the following type of

graduation.

log ls = cea 2(a + bs + ca? + dz’). (i)

This amounts to asserting that the instantaneous death

rate increases with age as a modified logarithmic func-

tion of v. :

The actual equations for the four calculated l» lines of

Tables II to V inclusive, together with the absolute num-

ber of individuals on which the curves are based, are as

follows:

Long oe fod (4,581 586 flies) :

log Ip == esor (3,0041905 — .02937911” + .0001402452* i

— 00000158974"). (ii)

Long pony Ea (5,426 flies) :

ewisisssw (3,0042303 — .01869993x + .000059620z

— ,00000204382°). (iii)

Short T gs (854 flies) :

log ls = ese (3,0085931 — .17931770a + .0040106302°

— 00003325012"). (iv)

Short peit Qs (906 flies) :

ls = emma (30116555 — 149486152 + 0028512192"

,00002106422"). (v)

No. 641] THE DURATION OF LIFE 493

The plan of arrangement of Tables II to V is as fol-

lows: The first column gives the age of the flies in days,

starting theoretically from the time of the emergence of

the imago from the pupa as zero. Since the flies spend

on the average a day in the breeding bottle before they

are taken out into the small duration-of-life bottles, and

the deaths are not observed for this interval, our distri-

butions as recorded actually start with age 1 instead of

age 0. The next two columns give the observed deaths

and survivors on the basis of a thousand individuals at

‘‘ birth’? (here emergence as imago). The next three

columns give the calculated (graduated) values deduced

from equations (ii) to (v) above; first the l» line, next

the qz, and finally the ez, the latter values being of course

in days. Owing to the fact that no premium rates are

likely to be caleulated from these life tables, we have not

thought it necessary to keep but one place of decimals in

the case of the qe and ez lines, and none whatever in the

le line. Of course, in the computations more decimal

places were kept, and these life tables may be regarded as

accurate to a considerably higher degree than the figures

as here published indicate. But, on the other hand, so

far as we can see, the figues here tabled are sufficiently

detailed for any use to which they are ever likely to be

put.

The le lines of Tables II to V are shown graphically in

Figs. 1 and 2. The diagrams are plotted to an arithlog

grid, the scale of the abscisse being divided arithmet-

ically, and that of the ordinates logaritl lly. Field

(24) has shown the advantages of this method of plotting

life table l+ lines. He says:

In the natural-seale diagram the descent of the curve expresses

the number of deaths in a year among the survivors to a given age.

This is not the usual way of stating death-rates; nor is it a conveni-

ent method, since the absolute number of deaths is a joint resultant

of two factors which might ‘better be considered separately—the

probability of death at the specified age, and the number of persons

at thet age and subject to that hazard. We are ordinarily more

concerned with the probability alone, or, which is much the same

494 THE AMERICAN NATURALIST [Von. LV

TABLE II

LIFE TABLE FOR DROSOPHILA—LONG-WINGED MALES

ee Observed Calculated ren | Observed Calculated

in EENE TET in

Days dz ly qz er

ooo 5 63| 368| 45.8 14.2

ae ot 12 348| 351| 47.7] 13.8

Ca Sa E 6 328| 334| 49.6 13.5

I S ta 13 3 18| 51.6) 13.1

Bee us 10 287| 301| 53.7) 12.8

eoa, 10 271| 285| 55.7|: 12.4

Tic hs 15 258| 269| 57.9) 12.1

Boar 9 239| 254) 60.2) 11.8

Co 9 227| 238| 62.5) 11.5

Wie os 9 208} 224) 64.8] 11.2

Enn 12 195| 209| 67.3) 10.9

ee la a 8 95| 69.8 10.6

Boo oo 8 165| 181| 72.4! 10.3

A 11 157| 168] 75.2) 10.1

Tee 144| 156| 77.9, 9.8

E de 14 132} 143] 80.8 9.5

Wie 8 119} 132| 83.6) 9.3

E 13 105; 121 s67) 9.0

19.3) 10 98} 110) 89.8! 8.8

PoE EE 11 90 92.9 8.6

Pa RASS A 16 85} 91 96.1) 8.4

7 ae oe 6 77| 82| 99.6 1

oe, 13 72| 741029 7.9

pe ee 1l 65! 67|106.4) 7.7

io ee: ll 57| 59/110.0| 7.5

Ps EE 10 113.8) 7.3

7 I o 10 45) AT|117.3). 7.1

a 14 43} 41|121.5| 6.9

W ee 11 35| 36)125.4| 6.8

80. 2500515 15 81} 32/120% 6.6

Sijs. 3 27 133.6 4

e are 3! 24) 941137.5: 6.3

RE eS. 15 21; 2111420 1

wooo 7 17| 18|146.4| 5.9

BO, onn. 18 15) 15)151.0) 5.8

TS oa 5 14, 13/156.2} 5.7

Ls SR a 19 12| 111160.2) 5.5

SB 13 10 9/164.5) 5.4

80 22 9 8170.6) 5.2

Ch EE 15 7 6 175.6) 5.1

Al ee 13 5 5 180.4 0

to o 23 5 4 185.0 9

3G... 19 4 s beat 4.8

ao 22 3 3/196.3} 4.6

ab an 18 2 ig as 4.5

M600 22 2 207.5| 4.4

> SEE 19 1 1\212.8| 4.3

Se o 15 1 1/218.3| 4.2

W.. 20-

No. 641]

THE DURATION OF LIFE

LIFE TABLE FOR DROSOPHILA—LONG-WINGED FEMALES

495

| |

Äge | Observed Calculated Age Observed | Calculated

in | r in |

Days | dz | lz le | qr ez Days dz ls | lz qz ex

oo, | 5 |1,000/1,000! 9.7) 38.8| 46........ 19 | 386| 384! 46.4] 14.3

3o. | 14 | 995| 990! 9.71 38.1147........ 19 | 367| 367| 48.3] 13.9

Be a. | 10 | 981] 981] 9.8] 37.5]48........ ‘13 | 348| 349] 50.2] 13.6

ae Satie 13 | 971] 971] 99| 36.9] 49.. 23 | 335| 331) 52.1] 13.2

Bon | 12 | 958 961 10.0, 36.2] 50. 16 | 312| 314) 54.1] 12.9

e oo | 10 | 946] 952| 10.1) 35.6] 51.. 20 | 296! 297| 56.1| 12.6

eee | 13 | 936] 942] 10.3] 34.9] 52........ 15 | 276) 281| 58.2] 12.3

ee | 8 | 923] 932] 10.5] 34.3]53........ 15 | 261| 264| 60.4| 12.0

eee | 11.| 915] 923] 10.8] 33.6] 54........ 17 | 246| 248| 62.6] 11.7

[of | 13 | 904] 913] 11.0] 33.0/55........ 11 | 229) 233 64.9! 11.4

11........| 12 | 891] 903] 11.4] 39.41 56.. 17 | 218| 218| 67.2) 11.1

2o.. | 6 | 879| 892| 11.8| 31.7| 57.. 13 1) 203| 69.6) 10.8

E | 9 | 873 382 Bizs.. 12 | 188| 189| .72.0| 10.5

t4 o l 40 | Seu s71 124 3041 0. o 15 | 176| 175| 74.5| 10.3

B o o. l8 | 853 s61 12.9] 29.8160........ 12 | 161! 162| 77.1] 10.0

16. 12 | 844! 949] 13.3] 29.2161........ 11 | 149| 150 79.8) 9.8

Wo | 9 | 832{ 838] 13.9] 28.6] 62........ 11 |- 138] 138| 82.4) 9.6

eee | 11 | 823 827| 14.4| 28.0163........ 9 | 127| 126) 85.1] 9.3

es 17 | 812| 815| 15.0) 27.4] 64. 9 | 118| 116] 87.9] 9.1

eee | 10 | 795 an 15.6} 26.8] 65........ 12 | 109! 105 908 8.9

| Bans | 22 | 788| 790| 16.2) 26.2|66........ 10 96| 93.8] 8.7

aooo | 10 | 763! 7771 16.9| 25.61 67....... 7 87| 96.7| 8.5

pre 16 | 7831 764| 17.6] 25.0168........ 11 78| 99.8) 8.3

w 14 | 737| 751| 18.4| 24.4169........ 5 71102.9| 8.1

aE 10 | 723, 737 19.2| 23.9170........ 10 63/106.2) 7.9

a 12 | 713 Ot Oe Fi. 3 57|109.2| 7.7

Fe 14 | 701) 708) 20.9| 22.8|72........ 7 50/112.5| 7.5

SB o.. 18 | 687|} 693] 21.9| 22.2173........ 3 115.9| 7.3

aga tafe 16 | 669] 678} 22.9] 21.71 74........ 4 401119.4) 7.2

w 16 | 653|. 663], 23.0] 21.2175........ 5 123.0) 7.0

rh ea pine 12 7| 647| 24.9] 20.71/76........ 1 311126.2 6.8

eee 13 631] 26.0! 20.2) 77........ oy 27|130.2; 6.7

Ao 15 | 612| 614| 272 19.7|78...-..... 4 ı 231133.41 6.5

ano 14 | 597| 597| 28.4| 19.2| 79.. 1 201137.5 6.4

3o o 18 581| 29.6! 18.8|80........ 3 17 141.6. 6.2

3 ge 3 583| 30.9] 18.3]81........ 2 15 144.9 6.1

ea 16 | 848] 546| 32.2] 17.9182........ 4 13 149.0 6.0

a. 17 | 532| 528| 33.6 17.4183........ 1 111153.0) 5.8

E 19 | 515) 511| 36.0! 17.0 84........ 1 9157.8! 5.7

aoo 17 | 496 493 36.6 16.6185. 1 8 161.5 5.6

Eoo 15 | 479| 475| 38.0] 16.2)86........ 2 6|164.9| 5.4

w 13 | 464| 4671 39.7| 15.8187........ 0 5/169.7| 5.3

PC a 15 | 441 439) 41.2] 15.4188........ 1 4|173.3| 5.2

mS 21 | 426| 420. 42.9] 15.0189........ 1 4|180.1| 5.1

Chee: 19 | 405| 402: 44.6 14.6] 90.. eo 183.6! 5.0

Coo: 1 2/188.7| 4.9

ne 0 2/1921. 4.8

a... 1 21195.1| 4.7

Ho. 0 1204.5) 4.6

oo 0 1209.5 4.5

496 THE AMERICAN NATURALIST [Vou. LV

TABLE IV

LIFE TABLE FoR DRoSOPHILA—SHORT-WINGED MALES

Ane Observed Calculated Ags Observed Calculated

in | in cr

Days ae ck k Gat ee Days Ur lz G G

onoo. 6 |1,000|1,000] 25.6| 14.2| 24........ 16 | 151| 155|107.5) 8.0

Bio fees 2 poa: O74 S15) AsO) a 13 135) 139/107:3) 7.8

r E ER 30 9 44 IRA S 2) E i R 11 122; 124)107.1| 7.7

i U sea N 34 D31! 908i 43.3) IZA Si oo ee: 6 111; 111107.2: T:5

E T F 38 903| . 869| 48.9; 12.0] 28........ 20 1065]; 99/1073) 7.2

f oR agua 36 865| 826) 54.6) 11.6129... 2. 4 85 881107.9. 7.0

Eee OES 85 $29| T51! 60.0) 11-4530. È 77 79/109.2| 6.7

Byes: 66 T44: 134: 65.5 10.8;31........ 13 70 TTIR 64

Gas whee 678} 686) 704) 10.5133.. ... 7i 57 62/114.5| 6.1

LO ee 52 G23) - 6388) 75.2) 10:2) 333. 22. 9 50 55/119.3) 5.8

Fie STi 590 798 O03. ee: 5 41 49;126.2| 5.4

i e ee 48 527| 543 0 O a 4 36 42/135.4| 5.0

E ES ee Ie 21 479i 497| 87.9; 9.5136....:... 6 32 37|147.4| 4.7

i Nae ee 49 4 54 5) 9.3) 37 4 26 $1/162.7| 4.3

ioa 53 409) 412| 94.8! 9.1] 38 1 22 26/182.2| 3.9

oat

16327 43 SOG) 373i 94:7! Dl OOo ia 6 21 21 (206.2) 3.6

a FRA O 24 313| 3371100.2). 8:91 40:....... 2 15 17 234 Tl. oe

Ia- onra 28 289: 303102.) 874i... 4 13 13,268.60; 8:0

CS R A 22 2 SUG Le SGN ae ee 1 9 1030; ot

pai Ea ee Oe 19 239). 2441105.3).. 8.5143 .°. 2... 2 8 71352.9| 2.4

rA E OSes 24 220) 218106.3) Salat oo ce. 5 6 4403.0) 2.2

a) US 17 196) 105106.8: 8.3146 2. 3.50. 0 J 3 457.9 2.0

bts ME BE ane 28 I7% 1740072 SHB. 0 1 1 16.0 1.8

thing, with the proportion of those persons of given age who die in

the course of a year. Precisely this relative mortality rate de-

termines the slope of the curve in the logarithmic figure, for here,

as always, a given distance on the logarithmic scale denotes a cer-

tain proportion of change. Hence the more steeply the logarithmic

curve descends, the higher is E relative mortality which it ‘indi-

cates. Hence, too, it is possible to provide a key to the diagram in

the form of standard sample slopes and corresponding numerical

death-rates, which hold true for all parts of the curve

From these tables and the diagrams, the following

points are to be noted:

1. It is obvious that the laws of mortality are funda-

mentally similar in Drosophila to what they are in man,

with the one striking and outstanding difference that

since in the case of the Drosophila life tables we are deal-

No. 641] THE DURATION OF LIFE 497

TABLE V

| |

| Observed Calculated | Observed | Calculated

Age | Age |

in | | in | | | | |

Days dz & iE | Go & Days dz | Ly | G | Ge | tz

ee | 10 /1,000/1,000} 30.2} 15.8] 27........ | 13 | 177| 165| 93.1} 8.5

- Es? | 26 | 990 970) 33.9] 15.2 e o wid 49| 94.6) 8.3

a. | 37 | 964) 937| 37.6| 14.7 | 11 | 146] 135| 96.11 8.0

ac: 31 27, 902| 41.2) 14.3] 30 | 20 | 135] 122] 98.1) 7.8

Ee Rae ae | 47 | 896 865| 44.7; 13.9 31........ 15 | 115) 110)100.2) 7.5

B | 27 | 849) 826| 48.1] 13.5] 32.. 15 100° 99 102.8 7.2

Pie. oe 50 822 786| 51.5| 13.1] 33.. AER 85| 89|106.3| 6.9

Eeo 78 | 772| 746| 54.7| 12.7| 34.. i Š 77| 80/110.0| 6.6

OF 45 | 694} 705| 57.9} 12.41 35........ Heed 72| 71)114.5) 6.3

1622. os 63 | 649} 664] 60.9] 12.11 36........ 10 65| 63/120.2) 6

Tes 34 586 624) 63.8) 11.8] 37 6 55| 55/127.1) 5.7

a. 35 2} 584| 66.6] 11.438........ 7 49| 48/134.8 5.4

sa ae 25 | 517, 545) 69.2) 11.3] 39 8 42| 42)144.4), 5.1

Fea 39 [BOT TAT 11.0 a0... 8 34| 36/155.8 4.8

mo oo si | A 471] 74.1) 10.9. 41........ | 8 26, 30 pe 4.5

woo 30 | s12 436] 76.2] 10. 42. 3 18; 25/183.8) 4.2

i ia 98 | 882) 408]. 78.3|.10.848........ 1 15| 20/201.3) 3.9

Boy ays Os | O64) S71) 00.9) 10-01-46: ... 5.2: 3 14| 16/221.2| 3.6

woo 25 | 326] 341] 82.0) 10.1) 45......] 1 11| 13|243.8| 3.3

W ae a 14.1 BOll 318) S36) OG 46... o.. 0 10) 10/269.9| 3.1

2o o Pe 20. 987: 387i 85.1 O47 o o: 2 10| 7297.0) 2.8

oe 14 | 267) 209 86.6 9948... 0 5|328.9)° 2.6

eee 21 | 253| 240| 87.9] 9.3} 49.. 2 8 31363.3) 2.4

ee 21 | 232| 219| 89.2} 9.1] 50. 3 6 2/401.0} 2.2

Mis Fae | 16 | 211) 199| 90.4) 8.9) 51........ 1 3) -1/457.2} 2.0

| f

iy. | 18 | 195) 181| 91.7) 8. | |

ing only with the duration of imaginal life, the important

infant and early childhood mortality component of the

human dv line is entirely omitted. With this difference in

mind, it is apparent that the remainder of the l» curve

for the flies is essentially similar to any human I+ curve.

Further on, we shall make a more detailed comparison be-

tween Drosophila and human curves.

2. There is evidently a fundamental and marked dif-

ference between the long-winged and short-winged groups

in respect of the duration of life. This difference is some-

what more marked in the case of the males (Fig. 1) than

in the females, but it is sufficiently definite and clear-cut

498 THE AMERICAN NATURALIST [Vorn LV

IG Diagram showing the observed and graduated J, poirts for long-

winged and se -winged males, respectivel The small circles are the observa-

tions and the smooth lines the fitted ait from equations (ii) and (iv). In

order not to p Porina” the diagram, only every second observation is shown. es

)

45 49 53 57 6l 65 69 73 77 G& -85 89 33 97

AȘ

B3 IL- Al £0 £9 33 IT Al

100

IODA

LYCU

SYONAYNS

DAS co FLY. LIFE

No. 641] THE DURATION OF LIFE

2. Diagram showing the observed and graduated 1, points for long-

Olan and short-winged females, respectively. The small circles are the obser-

vations and the smooth lines the fitted curves teoth equations (iii) and ae In

‘order not to overcrowd the diagram, only every second observation is show

meat

5 49 53 57 61 65 69 73 77 8l 85 89 939

£5 £9 $3 37 Al

13 H H

1000

100

SAONAANS

ve O FLY LIFE

500 THE AMERICAN NATURALIST [Von LV

in both eases. Broadly speaking, the wild type long-

winged flies have from two to three times as great an

expectation of life at any age as do the flies of the Quin-

tuple stock. Since all of these flies lived under substan-

tially identical environmental conditions, as has been set

forth earlier, it follows that the basis of the great differ-

ence in expectation of life between these two groups, as

exemplified in Figs. 1 and 2, is hereditary and not en-

vironmental.

3. It is apparent that on the whole the graduations

given by equations of the type of (i) are very satisfac-

tory, and as good as could reasonably be expected on ex-

perience bases of the magnitude of those here dealt with.

Undoubtedly the curves would be slightly more smooth if

we had larger experience, especially in Tables IV and V,

where we are dealing with less than a thousand flies in

each case, but in the main the curves fit the observations

very well.

4. The death rates ( 5) generally inerease steadily with

advancing age. An exception to this rule is the slight

dip between ages 25—28 in the short-winged J table.

In Figs. 3 and 4 the Drosophila le lines are compared

with the human l- line taken from Glover’s (25) 1910

U.S. Life Tables. In order to make a just comparison,

the human l: line is displaced to the left in the diagrams

until age fifteen of human life coincides with age one

of the fly curve. This drops out the infant and childhood

mortality component of the human curve. It will be

understood that in the present instance, this is a some-

what arbitrary and purely graphic procedure. Whether

the point which exactly corresponds on the human curve

to the beginning of the Drosophila imago curve is exactly

15 years or 13 or 14, or some other near-by value, is a

matter for further research. In a broad way, however,

it is clear that the two lines must be taken to correspond

at something like this point in the human curve.

From these diagrams, it is apparent that, after leaving

out the infant mortality component of the human curve,

No. 641] THE DURATION OF LIFE 501

the essential difference between the human and Droso-

phila lo curves is that, up to what may be designated as

the end of the middle life portion (and even into the early

part of the old age portion), human beings have a rel-

atively better expectation of life than does Drosophila.

On the other hand, in the extreme old age portion of the

curve, the Drosophila expectation of life is relatively

0 E Ea are u

74g

ae ae

4 ag

NI h

r D

Š \

ie \ \

\ \

i ‘

DAYS oF FLY UFE T OF A 5 29 7 4i 549 453 57 61 65 69 T T7 8l 85 89 3 BW

WSN ip 25 27 9 a DEA G 55 59 63 67 TI 75 79 83 87 A 95 99

Fic. 3. A comparison, of human 1, and Drosophila 1, lines for males.

better than the human. The result then is as thouglr

some external power had seized the Drosophila l- line at

about the middle of its course and bent it to a sharper

angle in that region, stretching it at that point upward

and to the right and by this process converted it into the

human curve. Suppose one of the Drosophila l- lines,

as shown in Fig 3, to be a thin, flexible whalebone rod,

possessing mass. Then move a point on that rod stand-

ing say just above the final A in ‘‘ Drosophila ”’ in Fig. 3,

up to a point where it exactly coincides with the human

life table curve. Then the whole of the rest of the

Drosophila curve would fall into about the same posi-

502 THE AMERICAN NATURALIST [Vou. LV

tion as is occupied now by the human life table curve.

Put in another way, what appears to have happened is

that, as compared with Drosophila, more human beings

are able to live through middle life, but at the expense of

those who, if the mortality law was the same as in

Drosophila, would live to extremely advanced old ages.

As a matter purely of speculation in the present stage of

our knowledge, it may be suggested that the Drosophila

= WO ~

1000 ee — HAN Lp.

GP,

i ca

100 2%,

\ Za,

NS A

2 c XG

x 9 X bigi

Š AL o

9 \ P 3

10 A

Lt 589 95 97 TOI

DAYS OF FLY LIFE 5 9 13 M 2] 25 29 33 G7 Al 45 49 33 57 6l 65 69 13 TA B

YEARS OF HUMAN 5 19 23 27 31 35 39.43 47 5) 55 59 63 67 N 75 79 83 87 91 95 99 103 107 M 15

LIFE

Fig, 4. A comparison of human J, and Drosophila 1, lines for females.

lz curves represent more nearly the normal, fundamental,

biological law of mortality, and that the human curve has

been warped from this form as a result of those activities

which may be comprised under the terms public health

and sanitation. It is to be understood that at present we

offer this merely as a suggestion and in no way as a

settled conclusion. It is, however, clear that the effect

which we should expect these activities to have upon the

form of the Iz line is exactly of the sort which makes the

human curve different from the Drosophila in fact.

In this connection Fig. 5 is of interest and significance.

No. 641] THE DURATION OF LIFE — 503

It compares, for males, two human le lines nearly 2,000

years apart in point of time, with the long-winged Droso-

phila le line of Table II. The two human lines are (a)

Glover’s, 1910 U. S. table (as in Fig. 3), and (b) Mac-

donell’s (26) observed l» line from the population of

Roman provinces in Africa at about the beginning of the

Christian era, his data having been taken from grave-

stone inscriptions. We calculated the l» line here plotted

from Macdonell’s tabled dz data, determining an le

point at each quinquennium. This smooths the Roman-

African figures somewhat, and makes the l- line so de-

termined lie very slightly higher all along its course than

would be the case if we used a more elaborate and exact

mathematical procedure. The error, however, is so small

that it would scarcely be discernible in the scale at which

Fig. 5 is reproduced.

000 = HAN ,

V dire

DAIN

SURVIVORS

JEN]

Hia

ows oF iv KET 5 8 W I7 2I 23 2I 33 Z TO 65.69 73 a5 a9

YEARS OF HUMAN) 19 23 27 9 35 39 43 47 5I 55 59 63 67 N 15 79 83 87 9 95 99

Fig. 5. A comparison of Drosophila mortality with human mortality at two

periods (a) early in the Christian era, and (b) in 1910.

It is at once apparent from Fig. 5 that the Drosophila

survival curve runs, in general throughout its course,

between the curve for human beings 18 or 19 centuries

504 THE AMERICAN NATURALIST [Vou. LV

ago and that for the present time. As compared with

Glover’s 1910 U. S. tables, the Roman population of

Africa at the dawn of our era was, in respect of the course

of its mortality, even more Drosophila-like than Droso-

phila itself! Now in Roman Africa there was relatively

little of what we now understand as sanitation, hygiene,

and preventive medicine. Men lived to old age, if they

did, by virtue mainly of the strength of their innate con-

stitutions, and their good luck in avoiding fatal accidents.

At the present time hospitalization, the science and art

of medicine and public health, and general sanitation keep

many persons alive well into middle age who would in

those days have died much earlier because of a lack of

constitutional ruggedness. Altogether the data of Fig. 5

seem highly significant in relation to the hypothesis sug-

gested above as to the reason for the difference between

Drosophila and man in respect of mortality curves.

ACCIDENTAL DEATHS

The tacit assumption in all the foregoing is that each

of the 11,772 flies comprised in the four life tables died

a natural death, and that the time of death (or duration

of life) was in each case’ determined fundamentally by

internal factors, since the environment was substantially

a constant for all.

For certainly more than 98 per cent. of the flies this

assumption is unquestionably true. But in view of the

possibility that some few of the flies might be dying ac-

cidental deaths by drowning in the moisture, which some-

times collects on the surface of the food, it was. thought

worth while to attempt to prevent the collection of mois-

ture in some of the bottles and compare the duration of

life of flies kept in such bottles with that of flies kept

under the ordinary conditions. Accordingly, bottles of

food were prepared by putting dises of several layers

of a very absorbent paper, Zorbik, in the bottom of the

bottle and then pouring in the food and letting it solidify

No. 641] THE DURATION OF LIFE 505

on as steep a slant as possible, so that any moisture formed

would drain down and be absorbed by the paper in the

bottom of the bottle. Flies of generation 5 from four

different lines were used in the experiment, two norma!

wild type lines, one of Old Falmouth stock and one of

New Falmouth stock, and two lines of Quintuple stock.

Flies were taken out from stock bottles of these lines as

they emerged every day, beginning March 18, 1921, and

continuing through April 4, putting hatches of alternate

days in the specially prepared bottles and the other

hatches in ordinary bottles. The flies put into the specially

prepared bottles were of course kept throughout their

lives in such bottles, and the controls in ordinary bottles.

Table VI shows the l- lines of the four groups of flies—

long-winged with paper and slant food, long-winged with-

out paper and food surface horizontal, short-winged with

paper and slant food, short-winged without paper and

food surface horizontal. Distributions have been made

for the four lines separately, and for the sexes separately,

but since they all show the same results the separate

distributions are not given.

‘The data of Table VI are shown graphically in Fig. 6.

It is evident that there is no definite or marked differ-

ence between the slant food group and the other. Such

differences as do appear between the lz lines in the two

eases are only of the order of magnitude which might

readily appear in random sampling. This is indicated in

another way by the data of Table VII.

In the case of the short-winged flies the difference in

the mean is plainly not significant. In the case of the

long-winged flies the difference is 2.96 times its probable

error. One would expect a difference as great as this or

greater to occur from chance alone only 4 to 5 times in

every 100 trials, so that the difference is here getting on

towards the magnitude where it must be regarded as cer-

tainly significant on purely statistical grounds. But the

difference is in favor of the horizontal food without

drainage, and against the food with drainage.

506 THE AMERICAN NATURALIST [Vou. LV

TABLE VI

SURVIVAL DISTRIBUTION OF FLIES UNDER DIFFERENT CONDITIONS AS TO

: SURFACE MOISTURE on Foop

Short-winged Flies Long-winged Flies

Age With Without With Without

: Paper and |Paper. Food Paper and |Paper. Food

Slant urface Slant Surface

Food Horizontal Food Horizontal

Pi T PEA 1,000 1,000 1,000 1,000

Ps PAURE NIE os wee 983 1,000 — 7

ie eo oad eh 913 951 | 996 1,000

Or A 804 S57. | — —

2 y EO A TEETE 712 774 | 974 982

gE Noe, ts SO er an NEN 651 660 | -— —

PE Aveda e ag E 6 498 | 898 912

PiE PERT E EEO S 469 442 — —

DRS ees Seek 346 355 834 832

r a E N S ee cin 57 302 | noth

Meee. ey 182 234 | 796 799

h e PEE SE Pecos ees 115 147

a e EE EEE ENEA 84 83 675 770

Go ee es ae oa 8 38 =

BR Oe A eek 22 15 547 661

(Abe Agnes eRe a 17 4 — —

E, PE E E E T RENA 14 0 408 482

BE eos Wee ARES N 8 - —

Pee Ga eee ee Ree 3 —— 294 314

l- r A E AE 0 -— — —

OOo eae ER — —— 204 230

E MR ase T EOE -= —- —

GA cise ches — — 94 157

i RES EN ER AN —— — a

Soe See Sy tay, — — 34 120

TO aee a aN — — — —

i REER E RS -= — 23 51

3 Bee a Nan = — —— —

a EE EAT ee -— —-—- 15 7

BE BG ie re == —— — ne

OO eee. ys noe — 4 0

Absolute number of flies. (265) (274) | (358) (265)

On the whole it seems perfectly clear that these ex-

periments give no justification for going to the consider-

ably greater trouble of preparing this food so that there

is drainage from its surface. As a matter of fact the

drainage of moisture is never entirely complete even with

No. 641] ; THE DURATION OF LIFE 507

TABLE VII

MEAN DURATIONS OF LIFE CALCULATED FROM THE DATA OF TABLE VI

E Mean Duration of Life

nce

With Paper Without 4 pr P.E. Diff.

yar Slant Paper and |

ood

Slant Food |

Long-winged............ | 43. 80+ .73 | 46.91 +. 15 | 3.1041. o6 | 2.96

Short-winged............ 0.10+..40 | 20.58+.42 48+ | .83

1000 prou en

= £ONVg,

= Ly

Bex

\ N :

\ NS '

\ `

100 z

h Yi

S N

Š x

p E E es aces aon ee, mars vet "ewe et A ===

Siemon ona toon tomy gaat See Sanaa Rae ee

: 5 9 13 17 2) 25 29 33 37 4) 45 49 53 ST 6l 65 69 73 TI 8l 85 89 BI

DAYS OF FLY LIFE

Fig. 6. ear survival (lz) eri seng different food

onditions explained i

the slant food and absorbing paper. Small drops still

cling in some cases to the agar, and a fly might drown

in such a drop just as well as in a similar drop on a

horizontal surface. . The important point is that this ex-

periment confirms our general experience in this work,

namely, that accidental deaths occur so extremely rarely

under our conditions that they do not appreciably affect

the results.

508 THE AMERICAN NATURALIST [Vou. LV

SuMMARY

This paper is the first in a series of experimental

studies on the factors influencing the duration of life in

Drosophila melanogaster. An account of the experimen-

tal technique used in these duration of life studies is

presented. Four complete life tables for Drosophila are

given, and it is shown that this organism follows quan-

titatively the same general law in respect of the dis-

tribution of its mortality as does man. As this work

deals only with the duration of imaginal life in Droso-

phila there is no component in the life tables correspond-

ing to the mortality of infancy and childhood in man. It

is shown that there are wide differences in duration of

life in different stocks of Drosophila, and that the basis

of these differences is hereditary and not environmental.

The Drosophila survival line of the life table (l+) runs in

general throughout its course between human survival

lines of (a) the present time, and (b) about the beginning

of the Christian era (Macdonell’s data from Roman

Africa), the curves being superposéd on the basis of the

omission of the human mortality of infancy and child-

hood.

LITERATURE CITED

1. Pearl, R. The Biology of Death: I. The Problem. Sci. Monthly, Vol.

12, pp. 193-215, 1921.

II. Conditions of Cellular Immortality. Ibid., Vol. 12, pp. 321-335,

iad

. Id, IIT. The Chances of Death. Ibid., Vol. 12, pp. a. ae

4. Id. IV. The Causes of Death. Ibid., Vol. 12, pp. 489-516,

my la. V.: pa o of Duration of Life in Man. Ibid., core 13, pp.

46-66, 1

6. id. YL PR Studies on the Duration of Life. Ibid., Vol. 13,

pp. 144-164, 1921.

T. Id. VIL Nataral Death, Public Health, and the Population Problem.

Ibid., Vol. 13, pp. 193-212, 1921.

. Id, Modes of Research in Geneties. New York (Macmillan), 1915, vii +

2 pp.

- Morgan, T. H. The Physical Basis of Heredity. Philadelphia (Lippin-

cott), 1919, 305 pp.

No. 641] THE DURATION OF LIFE 509

10.

pan

Moenkhaus, W. J. The Effects of Inbreeding and Selection on the Fer-

ans Vigor, and SERS of Drosophila ampelophila. Jour. Morph.,

Vol. 2 ee 123-154

. Hyde, R. R. Inheritance P he Length of Life in Drosophila ampelo-

phila. Indiana Acad. Sci. Rept. for 1913, pp. 113-12

i e: J. P. Studies in e Longevity of Insects. Ann. Ent. Soc.

. Vol. 7, pp. 323-353, 1

. Lu g 5 E. Experiments da Deaieehite ampelophila Concerning

Nataral Pelecsion: Bulletin Amer. Mus. Nat, Hist., Vol. 34, pp. 605-

. Loeb, A and Northrop, J. H. Is There a Temperature Coefficient for the

cad. 6

Duration of Life? Proc. Nat. A

~ td a t oo the Duration of Life in Metazoa? Ibid., Vol. 3,

, 1917.

pp. 382-3

. Id. On the “shen of Food and On upon the Duration of

Life. Jou l. Chem., Vol. 32, pp. 103-121

iY ps

. Northrop, J. fs “The Effect of Prolongation of am Period of Growth on

the Total Duration of Life. Ibid., Vol. 32, pp, 123-126, 1917,

. Arendsen Hein, S. A. Studies on Variation in = Mealworm, Tenebrio

molitor. Jour. Genetics, Vol. 10, pp. 227—264, 1

. Bridges, C. and Morgan, T. H. The Second Chemo Group of

vam EAA In ‘* Cont. to the Genetics = tates melano-

r.’’ Carnegie Inst. Publ. No. 278, pp. 123- 919.

` gas

' Sturtevant A. H. Inherited Linkage Variations in A Second Chromo-

1 919.

; Maller, J. H. The Mechanism of Crossing-over, AMER. Nart., Vol. 50, pp.

34, 1916.

195-221; 284-305; 350-366; 421-4:

- Henderson, R. Mort tality Laws and Statistics. New York, 1915, pp. v +

. Dawson, M. M. Practical Lessons in eae Banat. An Elementary

. 2 vols

Text-Book. 2d edit. New York, 1

. Field, J. A. Some PE ae of ae Logarithmic pje in Statistical

Diagrams. Jour. Pol. Econ., Vol. 25, pp. 80

5 Oot J. W: panied States Life Tables, 1910. ee of the Census,

916, pp pp. 65.

; ca W. R. On the Expectation of Life in Ancient Rome, and in

` the Provinces of Hispania and Lusitania and Africa. Biometrika,

Vol. 9, pp. 366-380, 1913.

INHERITANCE OF CANCER IN MICE

: DR. LEO LOEB

THE DEPARTMENT OF CoMPARATIVE PATHOLOGY, WASHINGTON UNIVER-

SITY ScHooL or MEDICINE, St. Louis

In the following we shall give a summary of our in-

vestigations into the part heredity plays in the origin

of cancer. Our interest in this problem dates back a con-

siderable number of years. In 1899, in conjunction with

Dr. Jobson, we made our first observations on the en-

demic occurrence of cancer at the Stockyards in Chicago.

At that time we found cattle coming from a certain ranch

in Wyoming especially prone to have cancer at the inner

canthus of the eve (1). A few years later we observed

an endemic occurrence of sarcoma of the thyroid in a

family of rats. In this case we pointed out that the cir-

cumstances under which these sarcomata originated

pointed to a hereditary condition rather than to an in-

fection. At that time (1904)* we referred to the de-

sirability of investigating experimentally this endemic

occurrence of cancer (2) and we had particularly in mind

an analysis of the etiological factor in breeding establish-

ments of mice or rats. Such an opportunity presented

itself a few years later in the breeding establishment of

Miss Lathrop in Granby, Mass. A preliminary investi-

gation here revealed the fact that cancer occurred with

much greater frequency in certain inbred strains than in

others, and that no indications whatever could be found

of cage infection or direct infection from animal to an-

imal. We concluded in 1907 on the basis of these ob-

servations that there existed a hereditary predisposition

which was responsible for the endemic occurrence of

eancer (3). At that time we planned further much more

extensive experiments on the mode of hereditary trans-

mission of cancer in following isolated families and

strains of mice through several generations and in study-

1 A paper presented in abstract before the II Intern, Congress of Eugenics,

New York.

510

No. 641] INHERITANCE OF CANCER IN MICE 511

ing the effect of hybridization on the cancer rate. This

was made possible to us in 1910 through the interest

which Miss Lathrop, who bred a very extensive stock of

mice, took in these plans and with her cooperation we

carried out these breeding experiments from 1910 on

through a considerable number of years. In the following

year we published the record of a cancer family of mice

which we had observed through several generations (4).

Altogether we carried out observations on approximately

12,000 female mice reaching the cancer age which we fol-

lowed throughout the whole period of their life and which

were observed through successive generations. Many of

them were used for hybridization.

In the meantime E. E. Tyzzer had published in 1907

and subsequently studies of the inheritance of cancer in

mice (5). This author found indications that heredity

may play a part in the etiology of animal cancer. Some-

what later Murray undertook similar studies in which he

used methods similar to those previously employed in

the study of human cancer. This author compared the

frequency of cancer among individuals whose mother or

grandmother had cancer on the one hand, and among

those in whom cancer had not been observed in the direct

ancestry but may have occurred in the great grand-

parents (6). Murray found on the average about 20 per

cent. of cancerous mice among those whose direct an-

cestors had suffered from cancer and 11 per cent. among

those whose direct ancestry had been free from cancer.

In certain age classes the cancer rate of both kinds of

mice showed only a very slight difference and in one age

class the cancer rate was even higher in those whose

direct ancestry had not been affected by cancer.

These results as well as our previous studies and some

occasional observations of Albrecht and Hecht and of

Spronk and a few others made very probable the signifi-

cance of heredity in the etiology of cancer.

Nevertheless there were opposing views such as those

of Borrel and others who referred the endemic occurrence

512 THE AMERICAN NATURALIST [Vou. LV

of cancer not to heredity, but to infection, and as late as

1910 Bashford expressed the opinion that heredity has

no significance in the causation of cancer.

We believe that our investigations which were carried

out with the cooperation of Miss Lathrop established the

importance of heredity in the etiology of cancer beyond

doubt, and they point more accurately to the mode of in-

heritance and the interaction between heredity and other

factors (7). Subsequently Miss Maud Slye began to

use a large stock of mice which she had collected for

biological purposes for a similar study of heredity of

cancer in mice (8). While in general her conclusions

as to the significance of heredity in cancer agree with

ours and are thus confirmatory of ours, she has extended

her researches in various other ways and has made val-

uable contributions. In regard to certain questions our

conclusions differ. To those we shall have occasion to

refer in the following pages.

The following is a summary of our main conclusions:

1. The cancer rate of each strain of family is a definite

characteristic of this strain and is transmitted by hered-

ity to successive generations. The differences in the

tumor rate in various strains are very pronounced; the

tumor rate may vary between zero in certain strains and

almost 100 per cent. in others. All intergrades may be

found. To cite a few examples of tumor rates of various

strains: English 67.6 per cent., European + I daughter,

of No. 10 72 per cent., 344 + 328 = 79 per cent., London

28 per cent., No. 8 27.5 per cent., 8 + German 34 per cent.,

Cream 5.9 per cent., European 9 per cent., German + 8

0 per cent. (344 + Black Cream) + Cream 0 per cent.

While these strains represent composites, they are on

the whole in so far homogeneous, as in the large majority

of cases substrains showed similar tumor rates. Thus in

the case of the English and Cream, for instance, numer-

ous substrains showed the typical tumor rates. In a

number of cases individual families were separated and

followed, and on the whole their tumor rate agreed very

No. 641] INHERITANCE OF CANCER IN MICE 513

well with that of the main strain. Certain deviations

must of course be expected in the case of relatively very

small numbers of individuals obtained in ease of an in-

dividual family or small substrains, and yet quite fre-

quently even small substrains or families agree in their

tumor rate with the main strain.

We may give some examples of the comparative tumor

rates of the main strains and of the substrains: The

strain 84 + 328 had a tumor rate of 56.4 per cent. There

was among them a family No. 1075 consisting of ten fe-

males reaching an age which permitted inclusion in these

records. They had a tumor rate of 60 per cent.; another

family of this strain (No. 1,113), consisting of sixteen

such females, had a tumor rate of 88 per cent. Family

782a (22 females) had a tumor rate of 68 per cent.

Among the substrain English Sable which had in the

corresponding generation an average tumor rate of 75

per cent., there was a family No. 437 (27 females) with a

tumor rate of 89 per cent. The same correspondence is

found in others and among them low rate strains. In

certain cases, however, certain substrains of families

can be split off which differ in their tumor rate. Thus

at an early period of inbreeding there were split off

from the English strain two substrains with low rates:

English Silver and English Silver Fawn, with tumor

rates of 8 per cent. and 12 per cent., respectively.

From the strain London (tumor rate 28 per cent.) two

families were branched off which showed very different

rates: London Blue and White (31 females) 55 per cent.,

and Family 481 (25 females) 0 per cent. But these are

not the usual occurrences. A correspondence between

main strains, substrains and families is the usual finding.

We see then that all kinds of intergrades in the tumor

rates occur in different strains, substrains and families

and that these are on the whole constant and character-

istic of strains and families.

2. These differences in rate persisted through succes-

sive generations in the majority of our strains with a sur-

514 THE AMERICAN NATURALIST [Vov. LV

prising regularity. Thus for instance, in the strain

London, the earlier generations (120 female mice) showed

a tumor incidence of 27 per cent., in the intermediate

generation (61 females) the figure was 38 per cent. and in

the later generation (197 female mice) 28 per cent. Sim-

ilar conditions were found in a number of our strains. In

certain strains, however, variations’ in the tumor rate

did occur. While some variations may of course be ex-

pected, in case the number of mice considered is very

small, there occurred in addition changes which can not

be attributed to this factor.

In the majority of cases in which these latter changes

did occur in our stock, they consisted in a decrease in

the tumor rate in later generations; in a few cases only

there occurred an increase in the tumor rate. These

changes were in all probability due to two factors: (a)

In certain families and strains as a result of long con-

tinued inbreeding a gradual decrease in fertility and

vigor occurred. Associated with this change was in cer-

tain cases a noticeable decrease in the tumor rate. Es-

pecially in the strain No. 8 there seemed to be a connec-

tion between loss in resistance to disease and fertility and

the decrease in the tumor rate. This strain was inbred

for seventeen generations and the changes in the tumor

rate seemed to occur step by step in correspondence with

the progress in inbreeding. Under those conditions the

connection between inbreeding and change in the tumor

rate appears the most probable explanation, although it

can not be considered as definitely proven as yet.

(b) Various factors caused a selection to take place

within the strain; certain families died out, while others,

which happened to be more resistant to a certain disease,

survived, propagated and thus gained a preponderance.

These surviving families differed sometimes in appear-

ance, or in vigor, in the behavior towards certain inocu-

lable tumors. Such changes were accompanied in certain

cases by a change in the tumor rate. In the majority of

our cases a decrease occurred; in a few cases an increase;

No. 641] INHERITANCE OF CANCER IN MICE 515

but even in such cases the increase was moderate; there

was never observed among our material a sudden transi-

tion from a low to a high rate tumor strain. The increase

as well as the decrease in the tumor rate was caused by

the same factor; whether one or the other should prevail

depends more or less on chance, and in different material

the number of strains showing the one or the other varia-

tion may be expected to differ. It has been maintained

that in strains which have been inbred for a long period

of time and in which a decrease in fertility occurred as

the result of the inbreeding, development of cancers takes:

the place of the lost fertility. In inbreeding cancer re-

places reproduction, as it has been expressed by Maud

Slye. In our material such a substitution did not take-

place; in inbreeding mice vanishing fertility was not re-

placed by the development of cancer under ordinary

conditions. Inbreeding does not lead to an increased

cancer rate.

3. If we cross strains with a similar tumor rate, the

offspring inherits the tumor rate common to both par-

ents; if both parents differ in tumor rate, the tumor rate

of the offspring is on the whole intermediate between

those of the parents. But all degrees of intermediacy

are observed. In our material the number of strains in:

which the rate of the parent with the higher tumor in-

eidence ‘dominated was on the whole greater than the:

contrary one.

We selected for our hybridizati especially strains:

which differed markedly in their tumor rate and other

characteristics and which had been followed over long

periods of time and had been found consistent in their

behavior. The English as a representative of a high

tumor rate strain and the Cream as a representative of a

low tumor rate strain were especially suitable for this

purpose. Inthe majority of cases we selected few individ-

uals for hybridization, either one male and one female or

one male and several females. We followed the offspring

through several generations. The near relatives of the

516 THE AMERICAN NATURALIST [Vou. LV

individuals used for hybridization were observed as to

-their tumor incidence and generally found to behave in

a way characteristic of their strain.

Sometimes we hybridized sisters with the same male,

or we used consecutively the same male with females

from strains which differed much in their tumor rate.

The results in the hybrids could usually be foreseen from

the known tumor rate and tumor age of the parent up

to a certain point of variability. The cases in which the

strains used for hybridization had a similar tumor rate

could be considered as controls. Here a similar tumor

incidence ought to have appeared in the offspring and

this is what usually occurred.

As we stated above, the results of hybridizations are

typically intermediate, but the rates and tumor ages of

the crosses may in some cases approach the parent with

the higher rate, in other cases the parent with the lower

rate.

We shall cite two examples, where the offspring re-

sembled the parent with the higher tumor rate. (1) A

son of a tumor mouse No. 240, belonging to thé strain

8 + German, was mated to a White Cream female. 8 +

German was a strain fairly rich in tumors and the par-

ticular family used had a tumor rate of 43 per cent. The

tumors appeared early in life. In the White Cream used

in this case tumors were extremely rare and they ap-

peared late in life. Among the offspring 9 female mice

lived long enough to be included in the records. Of these

9 mice, 5 died with tumors, 1 in the first and 4 in the

second age period. In this case the influence of the father

is undoubtedly very marked. In the Cream strain such |

a tumor rate was never observed even among isolated

families. The tumor age of the hybrid is, however, in this

case probably affected by the mother.

(2) In the second case which we wish to mention, an

English Sable male belonging to the fourth generation of

English Sable, who have normally a very high tumor

rate, was mated to 3 females belonging to the substrain

No. 641] INHERITANCE OF CANCER IN MICE 517

Cream Y. In the substrain Cream Y the tumor rate had

been zero. Four generations of the offspring were ob-

served comprising altogether 68 female mice which

reached an age sufficient for inclusion in our records,

Thirty-six of these mice, that is, 53 per cent., died with

tumors, 11 of these in the first age period. This is a

record which comes near that of the English Sable.

These as well as numerous other experiments seem to us

more in harmony with the conclusion that multiple factors

underlie the hereditary predisposition to mammary cancer

in mice than the view of Maud Slye who maintains

that the factor for mammary cancer in mice is a recessive

monohybrid.

(4) The age at which tumors appear is just as char-

acteristic of individual strains as the tumor rate. The

tumor age is also transmitted by heredity. In some

strains tumors appear relatively early, the percentage of

tumors appearing in the first age period of life comprising

the first twelve months is considerably greater than in

others; and this characteristic is on the whole just as con-

stant in the strains as the tumor rate as a whole.

We can distinguish two factors in the inheritance of

the tumor age: (a) In general in the strains with the

higher tumor rates the tumors appear at an earlier period

of life than in the lower tumor rate strains. This comes

out very clearly when we divide all the strains into three

classes, those with a tumor incidence above 40 per cent.,

the high tumor rate strains; those with a tumor incidence

between 20 per cent. and 40 per cent., the medium tumor

rate strains, and those with a tumor incidence below 20

per cent., the low tumor rate strains.

If we determine in each class the percentage of tumors

appearing in the different age periods, we find that the

tumors appear the earlier in life the higher the tumor

incidence and the difference between the different classes

is quite marked.

There is in our case a definite relation between the fac-

tors, tumor age and tumor rate. We can interpret this

518 THE AMERICAN NATURALIST [Vor. LV

relation by assuming that a certain average quantity is

inherited in the individuals of different strains which

determines the intensity in the tendency towards the de-

velopment of tumors. This intensity may depend on the

average number and character of multiple factors favor-

ing tumor growth which is characteristic of a strain. A

special kind of factors or a larger number of factors

causes both a higher incidence and an earlier appearance

of cancer in a certain strain.

(b) In addition to this intensity which is a akaradtar

istic of the strains in general, there is a peculiar tumor

age in certain strains which is independent of the tumor

rate. Strains with a similar tumor rate may differ in

their tumor age, and in hybrids tumor rate and tumor age

may be inherited separately in the offspring. Thus two

of our high tumor rate strains formed by the crossing

of the same male (European 151) with two sisters (first

and second daughter of No. 10, respectively) with tumor

rates of 72 per cent. and 54.5 per cent., respectively, have

relatively late tumors, in both only 15 per cent. of the

tumors appearing in the first age period. In the Cream-

English hybrids the tumor rates in two strains were sim-

ilar and approximately intermediate, but in one of them

the tumor age approached the late one of the cream par-

ent; in the other it was nearer the early one of the Eglish

parent.

We may therefore assume that in addition to the sum

total of multiple factors which determine at the same

time age and tumor rate, there are special factors which

determine tumor age.

5. In general the cancer rate in mice is not a sex- -linked

character. Either the cancer rate and age of the father

or mother strain may predominate in the cancer rate of

the offspring. This fact does not, however, exclude the

possibility that in certain cases a sex-linked factor may

enter as one of the multiple factors which in all probabil-

ity determine the inheritance of cancer. Certain of our

observations suggest such a possibility. We found, for

No. 641] INHERITANCE OF CANCER IN MICE 519

instance, that in the Cream-English hybrids the mother

strain was considerably more often predominant than the

father strain; in addition we found that in reversing a

cross different results were obtained in accordance with

the difference in the tumor rate of the mother strain. It

is, however, possible that these occurrences are chance

phenomena and we offered this interpretation merely as

a suggestion.

6. Our investigations make it possible to express in a

quantitatively definite manner the hereditary tendency

to cancer in individual strains of mice, the figures vary-

ing in different strains between zero and 100. This he-

reditary tendency is, however, not a simple quantity, but

composite, because

(a) The hereditary disposition to cancer is probably

due to the cooperation of multiple factors. The results

of -hybridization, which essentially were of an inter-

mediate character, the fact that all kinds of intergrades

between father and mother strain exist and that all pos-

sible variations in the hereditary tendency to cancer exist

in different strains, and that the hereditary tendency

determining the cancer age is not entirely identical with

the tendency expressed in the cancer rate, very strongly

suggest this conclusion. Variations in the number and

character of the multiple factors in the different individ-

uals may be responsible for the variations in intensity

which determine the tendency to cancer in individuals, and

different strains may differ as to the mean in the distribu-

tion of the factors among the individuals belonging to the-

strain. Thus we may assume that the strains English,

84 + 328, European + I daughter of No. 10 have on the

average a greater number of factors than the individuals

of the strain Cream and German + 8 and many others;

and it would be conceivable that in many cases a

tumor mouse belonging to the strain English differs

from a tumor mouse in the strain Cream, the former

often exceeding the minimum of factors necessary for

the production of tumors.

520 THE AMERICAN NATURALIST [Vou. LV

As far as the ordinary mammary cancer of the mouse

is concerned, no definite proof has so far been brought

forward to support the view that the hereditary tendency

to cancer is due to the presence of a simple recessive

factor.

(b) There is hidden in the figure expressing the tend-

ency towards the development of cancer a second factor

which is variable; namely, the activity of the ovary. In

all the strains the realization of the hereditary tendency

to cancer presupposes the activity of the internal secre-

tion of the ovary. Without this cooperation no cancer

can originate. With the full activity of this factor the

hereditarily transmitted character for intensity of can-

cerous tendency determines the upper limit of the cancer

rate. Again the intensity of this ovarian factor can be

expressed in a quantitative manner, the quantity in this

case representing the time during which the ovarian in-

ternal secretion had a chance to act. If through castra-

tion in the early stages of adult, sexually mature life, at

the age of three to four months, this internal ovarian se-

cretion is eliminated in mice, mammary cancer is prac-

tically prevented from appearing even in normally high

tumor rate strains. The longer the ovarian function has

a chance to act, the more the cancer rate increases up to

the range which is given in the figure for the hereditary

tendency to cancer. While we can thus experimentally

lower the cancer rate of any strain, we do not so far knuw

of a method which would permit us to raise the cancer rate

above this point. The latter is almost reached if castra-

tion occurs at the age of eight to ten months. Suspen-

sion of breeding also diminishes somewhat the cancer

rate in the great majority of the cases, but to a very much

less extent than the exclusion of the internal secretion of

the ovary, which latter is the true realizing factor, the

cooperation of which is necessary. In one strain in which

through segregation of the female mice breeding had

been prevented the cancer rate was even higher in the non-

breeding than in the breeding mice (9). Injury to the

mamilla by the suckling young which Maud Slye believed

No. 641] INHERITANCE OF CANCER IN MICE 521

.to be the external stimulus leading to the development

of cancer in mice can therefore not be an important factor

in the causation of mammary cancer in this species of

animals. On the other hand, our demonstration of a cer-

tain influence of breeding on the cancer rate in mice adds

another, though minor, factor to the internal secretion

of the ovary, which latter represents, as we stated above,

the principal realizing factor. Secondary realizing fac-

tors may therefore be added to this primary factor.

In principle, conditions are probably similar to what

we determined in the case of the typical mammary cancer

of the mouse in all other kinds of cancer. But we have to

assume that the internal secretion of the ovary is sub-

stituted in other cases by other variable factors, which

may be either internal secretions of a different kind or

external stimulations. The latter play, as is well known,

a very important rôle in the origin of cancer. They

represent in addition a quantity which can be increased

at will in contradistinction to the internal secretions and

other inner factors. Thus through the use of external

stimulation it may be possible to increase at will the can-

cer rate in certain kinds of cancers; in this way the he-

reditarily fixed intensity may become entirely obscured.

Yet it can not be doubted that after all this factor is

present even in these latter kinds of cancer, the best repre-

sentative of which is perhaps the Roentgen ray cancer in

man.

7. Thus it.has become possible to express in a quantita-

tive way the tendency to a disease, cancer. This tendency

is due to the interaction of two main factors, both in-

ternal, the one hereditarily fixed and the other accessible

to experimental variation. Both factors combined are

- more than the predisposition to cancer; they are in the

ease of this particular kind of cancer its essential cause.

There may be, as we have seen, other factors superim-

posed upon these primary factors, like the effect of preg-

nancy; but they are not necessary, and the first two fac-

tors suffice for the development of mammary eancer in

522 THE AMERICAN NATURALIST [Vor. LV

the variable numbers which are characteristic of the dif-

ferent strains of mice.

8. Is it possible to associate the hereditary. tendency

to cancer with the other factors characteristic of par-

ticular individual mice or of strains of mice? We found

in certain cases that from main strains substrains could

be detached which differed from the main strain not only

in color, but also in the tumor rate; the most noteworthy

cases of this kind are the English Silver and Silver Fawn

substrains, detached from the main English strain at an

early period of inbreeding. In this case the tumor rates

differed in a very pronounced manner from that of the

main strain. But the connection between color and can-

cer rate or age is in this ċase, as in some other cases, an

accidental linkage. There is no real causal connection be-

tween the color and the factors that determine cancer.

It is apparently similar in the case of other characters |

such as vigor, prolificity, size and rapidity of growth.

We find strains of all kinds among the high as well as

the medium and low rate tumor mice. This comes out

quite clearly in the case of the various English-Cream

hybrids. Here the tumor rate and age may be quite sim-

ilar, namely, intermediate in different crosses, and yet

some of these strains may be vigorous, others frail; some

prolific, others poor breeders. In crosses certain char-

acteristics, such as wildness or tameness, vigor and re-

sistance to disease, or frailty, prolificity or the opposite,

are inherited, just as the cancer rate and the cancer age;

but these characters may be distributed among the hy-

brids independently of the predisposition to cancer.

However, it so happens that some of the most prominent

low rate tumor strains are poor breeders, slowly growing,

although vigorous mice, while some of the high rate

tumor strains are prolific, more rapidly growing; but

this relation does not seem to hold good in all cases and

may therefore not be causal. Quite recently T. B.

Robertson observed among his mice that especially the

rapidly growing individuals were prone to become cancer-

No. 641] INHERITANCE OF CANCER IN MICE 523

ous and he believes that a causal connection between the

developmental rate and tendency to cancer exists.

-9. There may, however, possibly be an exception to this

independence of the hereditary transmission of the tend- -

ency to cancer. As we have stated, we arranged our

various strains of mice in three groups, in strains with a

high, with a medium and low tumor incidence, and we

found that in these three groups the cancer age varies

pari passu with the decrease in cancer rate. If we now

determine in these same three groups on the same per-

centage basis the age of death from all other causes

taken together except cancer, we find the differences be-

tween the three groups considerably less than if we com-

‘pare the percentages of the cancer age. There is, how-

ever, a distinct difference. In the group of the high can-

cer rate strains the age of death from all other causes but

cancer is decidedly earlier than in the medium and lọw

rate cancer strains. The difference between the medium

and low cancer rate strains is very slight, very much less

than that between the high and medium rate tumor

strains, but this slight difference is again in the same

direction. This relationship between cancer rate and age

of death from other diseases may be explained in two

ways: (a) We may assume that the mice dying from

cancer are the strongest, most resistant individuals of

the family or strain and those which are left back are

therefore relatively less resistant to disease; the higher

the cancer rate in a strain, the less resistant are the mice

not dying from cancer, and the earlier, therefore, their age

>f death from other causes. Or (b) the majority of the

strains in whiçh the cancer rate was high happened to

be less resistant strains and therefore the average age

of death from other diseases is earlier. The average

difference between the medium and low rate tumor

strains, as far as general power of resistance is con-

cerned, happened to be very small. Although it is per-

haps impossible to decide definitely between these al-

ternatives, we believe the second one to be much more

As THE AMERICAN NATURALIST [Vou. LV

probable. If the first alternative were correct, we should

expect to find a decided difference also between the age of

death from other causes than cancer in the medium and

low cancer rate strains. Here the difference is almost

negligible. Furthermore, there are some strains with a

very high tumor rate, but in which the rate of death from

cancer in the first age period is relatively small. In those

strains the resistant individuals would therefore be

spared by cancer in the first age period; thus the re-

sistant individuals would not be eliminated and the age of

death from other causes should accordingly be late in

these strains. Actually we find in such high tumor rate

strains an early age of death from other causes. We

may therefore conclude that in the material on which we

base our conclusions the large majority of the high rate

tumor strains were strains with a low general resistance

to disease. While, as we have stated above, a high or

low degree of resistance may be associated with either

a high, a medium or a low cancer rate, this association

of a low degree of resistance with a high cancer rate

in a prepondering number of strains may possibly be

more than a coincidence. Maud Slye states that cancer

strains are the strongest strains, a conclusion at variance

with our experience.

10. The tendeney to die from other causes than cancer

at a certain period of life, the resistance to-disease in

general, is also hereditarily transmitted, but as we have

stated above it varies among different groups of strains

much less than the predisposition to cancer. This should

be expected if we assume that there exists besides a

general power of resistance a special resistance or pre-

disposition to individual diseases, and that the latter

may vary among different strains and may thus to a

certain extent balance each other in various strains.

Again the tendency to die from other diseases than

cancer at a certain period of life depends upon the co-

operation of the generative organs; but while in the dis-

position to mammary cancer the internal secretion of

No. 641] INHERITANCE OF CANCER IN MICE 525

the ovary is the main factor and suspension of breeding

plays only a subordinate rôle, in the case of resistance

to other death producing conditions, the suspension of

breeding seems to be the main factor and the elimination

of the ovarian function only a subsidiary factor which

merely acts through the suspension of breeding which it

calls forth or in which at least the suspension of breeding

is by far the more significant factor. We found that the

differences in the tendency to die from other causes than

cancer which we observe normally between different

strains of mice are entirely or at least to a great extent

- eliminated in mice which are prevented from breeding.

All those strains in which breeding is prevented become

long lived. If a difference in the duration of life should

still exist between different non-breeding strains, it must

be very much smaller than that between breeding strains.

Furthermore, the difference in the longevity between non-

breeding mice and castrated mice is likewise very small

and this is the reason why we conclude that castration

prolongs the life of mice mainly through its effect on

breeding. As far as the cancer rate is concerned, on the

other hand, we have shown that castration at an early age

is much more effective than prevention of breeding.

11. In man it has been observed by several authors

that in older individuals suffering from cancer a mul-

tiplicity of slight malformations, often due apparently to

a misplacement of embryonal tissue, or a multiplicity of

benign, or rarely even of malignant, tumors could be ob-

served. Similar observations were made more recently

by Goodpasture in the case of old dogs. It is usual to at-

tribute these findings either to an inherited tendency to

tumor formation in general in which imperfections in

embryonal development play a certain part, and in which, .

as a result of this general tendency, various kinds of

tumors develop in the same individual in its old age, or

by some authors emphasis is laid on the importance of

old age as such in the etiology of cancer; old age is sup-

526 ' THE AMERICAN NATURALIST [Vou. LV

posed to bring about a multiplicity of tumors or cor-

responding malformations.

In a similar manner Maud Slye states tioro is in mice

inherited a general tendency to cancer. In hybrid strains,

according to this author, this general tendency finds ex-

pression in the first hybrid generations in a tendency to

develop sarcoma, while in subsequent generations more

specialized tissues are affected which develop into car-

cinoma, and in still later generations multiple tumors are

prone to appear.

We have not been able to observe such a cycle in our

strains of mice. We had uniformly in all generations to

deal with mammary carcinoma and in many autopsies

which we made of tumor mice we failed to observe other

kinds of tumors. This does not exclude the possibility

or even probability that occasionally lesions may have

been present in other mice which were tumors of a differ-

ent kind. We deseribed, for instance, a beginning squa-

mous cell carcinoma in a mouse afflicted with a mammary

cancer about 10 years ago; but on the whole such occur-

rences were rare and they could not be interpreted as due

to the inheritance of a general tendency to cancer; in each

case external factors would then at least partially deter-

mine which particular expression this general tendency

should find.

In our strains there was inherited: essentially, not a

general tendency to tumor formation, but a specialized

tendency to cancer of the breast. This does not exclude

the possibility that in certain strains a tendency to the

development of another kind of tumors may have been

inherited side by side with the tendency to mammary

cancer. In favor of this conclusion we may cite the ex-

periences in cases of the so-called endemic occurrence of

cancer, as, for instance, the cancer of the inner canthus of

the eye in cattle observed by us in 1899, the cancer of the

scrotum in the rat observed by Hanau, our observations

of sarcoma of the thyroid gland in the rat. All these are

instances of ce inheritance of specific kinds of cancers.

No. 641] INHERITANCE OF CANCER IN MICE 527

The most striking confirmation of this view has in recent

years been funished by Miss Slye, who discovered certain

families of mice in which a tendency to special cancers,

as, for instance, cancer of the liver, was inherited. We

therefore conclude that inheritance to cancer consists in

general in a tendency to the inheritance of a particular

kind of cancer. This agrees also with the results of Miss

Stark, who found in Drosophila two specific kinds of in-

heritable, tumor like formations originating by mutation.

12. Our continued investigations have thus borne out

our earlier conclusion that the endemic occurrence of.

cancer among animals is due to this hereditary trans-

mission of the disposition to cancer. In addition, infec-

tion with certain metazoon parasites which act as an

external stimulus comparable to the action, for instance,

of Roentgen rays, may play a part in certain cases; but

even here the metazoon parasites seem to act on a basis

of hereditarily transmitted disposition. The observa-

tions of Fibiger, with which the recent findings of Roh-

denburg are in agreement, suggest this conclusion.

13. While these statements apply directly only to an-

imals, the evidence on hand makes it probable that, in

principle, conditions are similar in man; here also in all

probability one or more factors are hereditarily trans-

mitted which determine the intensity in the tendency to-

wards cancer development. In man this tendency has,

however, in many cases been more or less equalized among

different families as a result of long continued cross

breeding (10). Wherever this factor can be eliminated

as among different races which remained relatively pure

or among a very stationary population, as in certain

parts of Norway, the evidence points to the conclusion

that here too marked differences in the tendency towards

cancer exist in various strains and races (11). Even

among the ordinary population some occasional striking

findings very strongly suggest this view.

Furthermore, Davenport (12) has shown that the tend-

ency to neurofibromatosis is hereditarily transmitted as

528 | THE AMERICAN NATURALIST [Vou. LV

a dominant. Similarly, the tendency to certain other

tumors is undoubtedly inherited. The recent statistical

studies of C. C. Little make it very probable that an in-

herited predisposition to cancer plays a part in human

cancer in general (13).

As to the increase in the cancer rate which seems to be

so general an occurrence, we may suggest that, so far

as it is not due merely to improved diagnosis, it could be

referred to a greater frequency in the dominance of the

parent with a tendency to a higher tumor rate in the

offsprin

As we sitoa above, such a condition of dominance was

observed among our. hybrid strains.

REFERENCES

1, Leo pgi and George Jobson. Medicine, 1900. Archiv, f. Klin. Med.,

Vol. 70.

2. Leo Loeb. Centralblatt f. Bacter., 1904, Vol. 37, 2

3. Leo Loeb. University of Pe rinag iene Med. iat 1907 (March-

Apri

4. Leo b. Centralblatt f. allgem. Pathologie, 1911, XXVII, 993.

5. E. E. Tyzzer. Journ, Med. Research, 1907, ried 155. NS "Report Can-

cer Commission of Harvard University, 1909, 1

. J. A. Murray. IV. Scient. Report Imp. aa ae eae Fund, 1911,

114.

oie wax eae Medicine, XVII, London, 1913, Section III, Part I.

ABOL

T. rop and Leo Look. sahe c. Koc. Exp. Biol. and Med., 1913,

XI, 34 E. Jour. Exp, Med., 1915, XXII, 646-713; 1918, XXVIII,

475. Pro oc, Soc, Exp, Biol. ond Med, 1918, XV, 72. Journ. of Cancer

Research, 1919, IV, 137.

Maud Slye, K. F : Holm and H, G. Wells. J. Cancer Research, 1916, I,

eng 503; 1917, IL 1 at 1919, IV, 207; 1920, V, 53, 205; 1921,

57.

9. A, = C. Lathrop and Leo Loeb. Journ. of Cancer Research, 1916, I, 1.

Leo Loeb. Journ. Med. Research, 1919, XL, 477.

10, Leo Loeb. Am. Journ. Med. Sciences, 1920, Vol. 159, 781.

11. A..C, Garmann. Zeitsch. f. Krebsforschung., 1913, XII, sig

12. C. B. Davenport. Proc. Nat, Academy Sciences, 1918, IV, 213.

13. C. C. Little. II. Internat. Congress of Eugenics, PE York, 1921.

Abstracts of Scientific papers, p. 22.

DEVELOPMENT OF GONADS AND TRANSFOR-

MATION OF SEX IN THE FROG

DR. EMIL WITSCHI

LECTURER IN ZOOLOGY IN THE UNIVERSITY or BALE, SWITZERLAND

In the following paper I intend to give a summary of

the sexual problems in the frog, as they result from de-

seriptive and analytic research. Richard Hertwig’s suc-

cessful experiments on sex-determination have directed

the attention of investigators to this object. From 1911

to 1913 the writer was fortunate enough to have the oppor-

tunity of working in Hertwig’s laboratory and since then

the investigations have been continued at the zoological

laboratory of Basle University. The present paper is a

brief summary of the development of the sexual char-

acters. In a later publication we hope to explain the ex-

periments on sex-determination and to describe the

chromosome cycle.

In the number of this journal for July-August, 1920,

Dr. W. W. Swingle has published an interesting study

on ‘‘ Neoteny and the Sexual Problem,’’ and in the Feb-

ruary number of the Journal of Experimental Zoology

the same author has described in greater detail the

developmental history of the male gonad of Rana cates-

biana. In both communications Swingle has disputed the

correctness of the work of the Hertwig school. The fol-

lowing explanations may be regarded as a reply to

Swingle’s critical remarks.

A. THE Sexuat Guanps -

1. Morphology of the Undifferentiated Gland—The

gonad of larve 22 mm. in length, just before sex-differ-

entiation takes place, has the following structure. A

simple germinal epithelium encloses the central primitive

gonad cavity. Five to seven sex cords, budding from the

529

530° THE AMERICAN NATURALIST [Vou. LV

mesonephric blastema at regular distances, fill the cavity

as shown in Figs. 1 and 2.

Fies. 1. AND 2. Transverse sections through the two haai of undifferentiated

glands. Larvæ 22 mm, total length; 12 days old.

2. The Ovary—tThe ovary is formed by direct develop-

ment from the indifferent gonad. In consequence of a

rapid increase of the germ cells by mitotice divisions the

germinal epithelium thickens. In larve of 30-35 mm.

total length the germ cells begin to be arranged in ovo-

eysts. Aftera fewsimultaneous mitotic divisions the cells

of the cysts enter the maturation stages. As Fig. 3

shows, the first cysts are found in the distal part of the

gland. Towards the end of the larval period the for-

mation of the oldest ovocysts is disintegrated. Each ovo-

eytis having passed through the stages of pseudoreduc-

tion (synizesis, leptotene, pachytene) is surrounded by

follicle cells and now enters the second period of growth.

The nucleus increases considerably in volume and in the

growing protoplasm numerous vitellogen granules appear

(Fig. 7, sec. Oc.). Yolk is however not formed normally

before the third or fourth season. The sex cords are of

greater importance only in the male gonad, forming the

testicular interstitium and the rete apparatus—while in

the ovary each cord develops into an ovarial sac. As Fig.

1 shows, a little slit may appear already in the undiffer-

entiated gland. More such are found when ovocysts

No. 641] TRANSFORMATION OF SEX 53}

begin to be formed in the germinal epithelium (Fig. 3, c).

But during the stage of metamorphosis the slits unite

and form only one cavity: the secondary gonad cavity

or ovarial sac. In consequence there are in each ovary

ð to 7 sacs, forming the thin gonad endothelium (Fig. 7,

ly EY):

edited pp os a young ovary. C = secondary gonad cavity. O =

= follicle cell. Tadpole 38 mm. total length.

IG. 4. yeei ei section through the earliest developmental stage of the

testicle. Migration of the germ cells. sp = first germ cell im the sex cord

(spermatogonia). Tadpole 22 mm. total length; 12 days old.

FIG.

Characteristics: 1. Persisting peripheral germinal epi-

thelium. 2. First stages of ovocytes in the larval period.

3. Occurrence of a second growth period. 4. Presence

of ovarial sacs.

Therefore, precocious ripening of germ cells is not my

‘chief criterion of sex-differentiation,’’ as Swingle says.

On the contrary the other three are more characteristic,

because they are completely absent in the male line!

3. The Testis.—In larve 22 mm. in length great differ-

ences in the size of the sex cords are found, as seen from

a comparison of Figs. 1 and 2. Animals with the stout

ones are undergoing transformation into males.

The testis is not formed by direct development from the

indifferent gonad. Its development begins with a change

in the position of the germ layers. The peripheral ger-

minal epithelium having disintegrated, the germ cells

cross the primitive gonad cavity and enter the sex cords

(Figs. 4 and 5). The follicle cells migrate with the gonia

532 THE AMERICAN NATURALIST [Vou. LV

and only the simple peritoneal epithelium remains (Fig.

5).

Shortly after this migration the canaliculi seminalis

are formed, at first as irregular slits, but already during

metamorphosis they develop into characteristic radiating

tubes. At all times they remain connected with the sex

cords. These on their part give rise to the rete vasculo-

sum Halleri and the connections with the mesonephros

(Vasa efferentia testis).*

In the following years the testes grow very slowly until

the fourth season, when a rapid increase due to active

mitotic divisions of the spermatogonia begins. The sper-

matic tubes become convoluted; spermatocysts are

formed and in the month of July the maturation cycle

begins. In August and September the spermatozoa are

developed.

In only one case have I found an abortive presperma-

togenesis in a three-year-old grass frog (R. temporaria).

But as Champy has observed, similar degenerating mat-

uration cycles are frequently formed in water frogs (R.

esculenta) in the second year. According to Swingle

precocious ripening of male germ cells in the bull frog

(R. catesbeiana) already occurs in first year larve, and

ripe spermatozoa are formed in second year animals.

Characteristics: 1. Central germinal,layer. 2. First

ripening stages giving rise to functional spermatozoa in

the fourth season. 3. The maturation divisions directly

succeed the pseudoreductional stages. (There is no sec-

ond growth period.) 4: Rete apparatus and Vasa effer-

entia testis.

4. Hermaphrodism.—I\t is a strange and interesting

fact that the typical development of the testis as just

described is rarely observed. Under natural conditions

most individuals develop first ovaries which later on

are transformed into testes. During this transforma-

tion the gonads, showing the characteristics of both sexes,

are hermaphroditic; but without exception the female

1 Cf, Witsehi, 1914.

No. 641] TRANSFORMATION OF SEX 533

characters become reduced and mostly disappear com-

pletely. In Fig. 6 we have a transverse section through

a larval hermaphroditic gland. As in the young ovary

(Fig. 3) the distal part contains ovocysts with ovocytes

in the synizesis stage, and in the sex cord is found the

second gonad cavity or ovarial sac. But the transforma-

Fic. 5. Cross-section through a young male gonad. Leaved germinal epithe-

ium: ll germ cells in the sex cord,

Fic. 6. Transverse section through a larval hermaphroditic gland. Female-

germinal epithelium; ovarial sac, At (i) migrating germ cells.

tion has already begun. The sex cord is more compact

than usual and its middle part is penetrated by immigra-

ting germ cells from the basal end of the germinal

epithelium. These germ cells after their entrance into

the sex cords are to be called spermatogonia.

After the metamorphosis the whole female germinal

epithelium undergoes degeneration (only the peritoneum

is preserved) while the central testis anlage develops into

a normal male gonad.

Sometimes great irregularities are observed. It occurs

occasionally that one gonad undergoes the transforma-

tion of sex earlier than the other in consequence of which

such animals pass through a stage of lateral hermaphro-

dism. Such cases have often been described and are, I

believe, of the greatest interest with regard to the devel-

opment of somatic sex characters. But even within the

same gonad differences can be found. Sometimes the

534 THE AMERICAN NATURALIST [Von LV

transformation begins at one pole and proceeds contin-

uously to the other. Fig. 7 shows a longitudinal section

of such a hermaphroditic gonad. The small testicular

part at the posterior end could easily be distinguished

macroscopically. The fifth sex cord is almost a nor-

mal male one. Numerous spermatogonia are scattered

throughout. ‘The slits, which are seen, represent the be-

ginnings of the spermatic tubes. At (a) a bridge of tissue

he)

Fic. 7. Longitudinal section through a ep taegutrnenay gonad, Transforma-

tion of sex hi Gok from the posterior to the anterior pole. I-V = the five

sex cords. sec. oc vocytes in the second adka period.

is seen, connecting the cord with what is left of the female

germinal epithelium. In the fourth cord the transforma-

tion has not progressed as far as in the fifth. The center

No. 641] TRANSFORMATION OF SEX 539

is occupied by the shrinking ovarial sac with a degenerat-

ing egg. The spermatogonia are not numerous but they

are already dispersed throughout the wall of the ovarial

sac. The germinal epithelium contains a degenerating

egg and an ovocyst. The third cord shows the beginning

of the process of transformation. It is greatly enlarged,

but only a few spermatogonia have migrated into its

basal part. Numerous degenerating eggs are scattered

over the distal tissue of the sex cord. The germinal epi-

thelium is disintegrating. The second and the first sex

cords show the typical proportions of an ovary: wide and

thin-walled ovarial sacs and well-developed germinal

epithelium. The ovogonia adhere to the peritoneum, the

larger eggs projecting inwards. Tbus the walls of the

sacs are folded on the outside. This preparation is taken

from a first year frog, several months after metomorpho-

sis. The transformation of sex in the grass frog often

occurs in the second year but likewise is sometimes found

in adult animals. Eggs scattered through the testicular

tissue have been frequently observed, but ignorance of

the development of the gonad has produced the belief that

they were derived from spermatogonies. Recently Levy

and Swingle claim that these ‘‘ so-called °’ eggs only

are enlarged ovocyte-like male germ cells; and Swingle

believes that ‘‘ the so-called sexually indifferent or sex-

ually intermediate forms of the Pfliiger-Hertwig school

are very probably male animals whose germ cells show

precocious ripening as far as the pachytene stage’’

(1920); and ‘‘as a consequence of this curtailment of

the maturation cycle to the early stages of the process,

without exception these writers (Witschi and others),

being unable to differentiate male from female, concluded

that all frog tadpoles first develop as females, then later

half of the female tadpoles must transform into

males. . - °" (1921).

It is evident that Swingle has misunderstood our previ-

ous communications. In Rana temporaria the ovocytes

do not always degenerate after the pachytene stages but

536 THE AMERICAN NATURALIST [Vou. LV

enter the second growth period (Fig. 7). It is difficult

to understand when Swingle says that R. Hertwig,

Kuschakewitsch and Witschi ‘‘ concluded that all frog

tadpoles first develop as females.’’ On the contrary I

have (1914) described the typical (or direct) develop-

ment of the testicle as it is to be observed in the alpine

Rana temporaria. Under optimal conditions of tempera-

_ ture in this variety of the grass frog already after the

twelfth day (larve 20 to 22 mm. total length) 50 per cent.

males are found; in such cultures transformation of sex

never occurs. ae

In his material Swingle has not seen this typical de-

velopment of the testicle. The described first and second

year males are in reality hermaphrodites. His photo-

graph 33, plate 4 (1921) does not show a transverse

section through a male but through the female gonad of

a first year tadpole, characterized by the ovarial sac

(secondary genital cavity) and the peripheral germinal

epithelium. Photograph 34 likewise is not taken from

a male gonad but from a hermaphroditic one. It shows

the same stage of transformation of an ovary into a

testicle as Fig. 45 in our publication (1914): representing

the gross structure of an ovary, but in the thickened wall

of the ovarial sac are several immigrated germ cells

(spermatogonia). Photograph 35 illustrates another

type of transformation, characterized by an excessive

proliferation of the sex cords, as is likewise described

in our publication (1914, Fig. 41).

The cytological facts deseribed in great detail by

Swingle will be discussed in another communication.

‘They do not alter our view of the significance of the

developmental changes.

If there should still remain any doubt in regard

to the correctness of my interpretation the following ac-

count may help to dispel it.

No. 641] TRANSFORMATION OF SEX 537

B. Tue Somatic Sex CHARACTERS

1. The Miillerian Duct.—The oviduct first appears as a

dense cell cord close to the lateral margin of the kidney.

Already in the first sumner the oviducts from the ostium

abdominale tube to the opening in the cloaca are com-

pletely formed. They are thin walled, contain a small

lumen, and run straight backwards. In the second year

they grow slowly and move away from the ureters. Ordi-

narily in the third year a large longitudinal growth

begins and the duct now becomes folded, as is well known

from the anatomy of adult frogs.

In males which show a typical development of the tes-

ticles, no Miillerian ducts of any significance are formed.

On the other hand, such animals as first develop ovaries

and later undergo the transformation of sex, also show

regular oviducts; and these continue to grow just up to

the time when the transformation of sex begins. This

parallelism in the behavior of the Miillerian ducts and the

gonads furnishes definite proof that the “‘ eggs” and

“* ovocytes,’’ described by the writer, are in fact really

eggs and ovocytes and that the transformation of sex

is a well-established fact.

After the transformation of sex, when the ovocytes have

disappeared, the Miillerian ducts begin to shrink, but they

do not disappear completely, and such reduced oviducts of

various sizes are often found in adult male frogs.

Regarding the question of the relations between so-

matic characters and the gonads, the lateral hermaphro-

dites furnish most interesting information. The Miiller-

ian ducts are always developed in correlation with the

gonad of the same side. Lateral hermaphrodites show

always well-developed oviducts on the ovarial side and

smaller ones on the testicular side. The correlation is

therefore independent of the action of hormones.

2. The Vesicula Seminalis and the Thumb Cushion.—

The male somatic characters appear chiefly in the second

year and always develop symmetrically, in lateral her-

538 <° THE AMERICAN NATURALIST [Von. LV

maphrodites, both sides depending from the first formed

testicle. They seem, however, not to be influenced by

internal secretions, as all experiments gave negative re-

sults. We will not enter here into this much discussed

problem, but refer the reader to our publication on her-

maphrodism,

REFERENCES

Champy. ee

“13. Cit pieces aimee’ des Batraciens, Arch. Zool. Exp., 52.

re oan pe

P und Physiologie der Geschlechtsbestimmung.

Hertwig, 1 p h.

Ueber den bree py Stand des Sexualproblems. Biol. Central-

blatt, 3

King, H. D.

708. The Oogenesis of Bufo

——. The Structure and Development of Bidder’s Organ in Bufo lent.

Journal Morph., 19

Swingle, W. W.

720. Neoteny and the Sexual Problem. Amer. NAT., 54.

721. The Germ Cells of Anurans. J. Exp. Zool., 32.

Witschi, E.

’14. Experimentelle Untersuchungen über die Entwicklungsgeschichte

der Keimdrüsen von Rana temporaria, Arch. mikr. Anat., 85.

721. Der Hermaphrodismus der Frösche. Arch. f. Fadail undone:

chanik, 49

THE RATE OF GROWTH FOLLOWING AN INITIAL

PERIOD OF SUPPRESSION *

DR. HOWARD 8S. REED

Tivision oF PLANT PHYSIOLOGY OF THE UNIVERSITY OF CALIFORNIA

CITRUS EXPERIMENT STATION

THE present paper attempts to discuss quantitative

aspects of the growth of animals which, though eventually

reaching the same approximate size, reached that size

at widely different ages. The data upon which the dis-

cussion rests have been drawn from the published articles

of several workers in the field of animal physiology and

deal with the growth, under varying conditions, of the

albino rat.

When the growth of an animal is suppressed for a long

time the capacity to grow persists, even beyond the

period at which growth ordinarily ceases in that species.

The studies of Osborne and Mendel have amply demon-

strated the existence of this capacity to grow and to

reach the weight characteristic of mature individuals of

their species. The growth impulse is something inherent

in the organism. The environment, while modifying the

amount of growth, has less influence upon the specific

character of the growth of organisms than has the es-

sential constitution of the living substance. A quantita-

tive study of the growth rate of organisms ought, there-

fore, to lead to considerations of a fundamental nature.

The nature of the growth rate in general is revealed by

the use of a few simple equations of the first order. They

show that growth proceeds at a rate similar to that of a

monomolecular reaction. Robertson’ and others have dis-

cussed growth in its relation to autocatalysis. In a recent

paper I have compared the equations of slowly and rap-

1 Paper No. 84 from the University of California Graduate School of Trop-

ical Agriculture and Citrus Experiment Station,

2 Robertson, T, B., Arch, Entwicklungsmech, d. Org., 37: 497-508. 1913,

539

540 THE AMERICAN NATURALIST ` [Vow. LV

idly growing apricot shoots. In each case the rate was

proportional to a function of the final length of the shoot.

The shoots which had a greater final length at the end

of the season grew more rapidly from the start than their

shorter neighbors, though the growing periods of the two

samples were the same. The particular point of interest

lay in the fact that the equations representing the growth

rates had the same value for the constant of the reaction,

differing only in the value of the constant expressing the

final length of the shoots.

The present paper attempts to supplement this work

by investigating the growth of organisms which reached

approximately mature size after being subjected to con-

ditions which suppressed growth in early life.

In the former case the difference between the two lots

was in their final size; in the present case the difference

between the two lots was in the time required to make

equivalent body weight.

A. Tar RATE or GrowTH OF Rats on ADEQUATE DIETS

The growth of the white rat has been so completely

studied by many investigators that no extended discus-

sion of the subject is required.

The rate of growth of rats varies slightly in different

lots, but in general it follows the course of a differential

equation. In later paragraphs I shall show that the

equations used are those which express an autocatalytic

reaction. The rate at which each sex grows is quite char-

acteristic. The females grow relatively faster in early

life than the males, come sooner to maturity, and weigh

less at maturity than the males.

The growth of a white rat in the first year comprises

two cycles. The first cycle, covering approximately 150

days, consists of a rapid increase in the weight and size

of the body. The second cycle, covering the remaining

200 days, consists of a thickening of the body and a

deposition of fat. The growth of rats in each of these

cycles may be expressed by the equation

No. 641] THE RATE OF GROWTH 541

log ya =K(t— t),

in which x represents the weight of the animals at time

t, a represents their weight at the end of the cycle, t, is

the time at which the weight, x, is one half a, and K is a

constant.

Although the quantitative relationships of this growth

rate have been ably discussed by Hatai,* it will be shown

subsequently that there are numerous reasons for using

the above-mentioned formulas for computing growth.

The computation of these and other growth rates studied

in this paper have been made with the aid ot tables pub-

lished by Robertson.*

Table I contains data on the growth of white rats in

the first year of life. The data for the rats were taken

from Donaldson’s tables 63 and 64.

The equations for the growth of the animals are as

follows: :

Males, first eyele, log 959. N =" .0187 (t — 73)

æ — 220

Males, second cycle, log mR) z = .0123 (t — 213)

Females, first cycle, log Ww = ==.0211 (t— 61)

— 170

Females, second cycle, log sec g = .0086 (t — 191).

It will be noticed that the calculation of the second

cycle involves a change of the axes of the coordinates

so that the new point of origin is near the point at which

the first cycle of growth ended. The first cycle of growth

‘in the female appears to terminate somewhat earlier than

that in the male and the value of K, the constant, was

greater in the growth curve of the female. These rela-

tions agree with the repeated observation that in early

life the female grows more rapidly than the male. In the

second cycle the female grows less rapidly than the male.

The close agreement between the observed and the cal-

n Donaldson, loc. cit

MaA e T. B., Univ. Calif. Publ. a ees 4: 211-228. 1915.

542 THE AMERICAN NATURALIST [Vonr. LV

9 TOO 200 000

Fie. 1. Curve of io baat for male white rat. ... cbserved weight ; ——,

calculated weight. The weight for days 0-150 was calculat from lo

[a/(228—a) ] = .0187 (t—73). The weight for days 150-365 was calculated

from log [ (a—220)/(280—a)] = .0123 (t—213). :

2 50

200

100

10 s

10 100 200 300 gon

Hao 2. Curve of eir for female white Th iei Popas weight

g Ceneu T

culated weight weight for days 0-124 calculated from log

P (170—r)] = 0211 ety The weight for ba piar was calculated

from log [ (#—166)/(226—«)] =.0086 (t—191).

No. 641] THE RATE OF GROWTH 543

culated. weights of the animals is shown by Table I and

by Figs. 1 and 2.

B. Tur GrowrTH or Rars RECOVERING FROM AN INITIAL

PERIOD OF SUPPRESSION

The published work of Osborne and Mendel contains

very convincing evidence that the white rat possesses an

inherent capacity for growth and that this capacity to

grow survives long periods of suppression due to in-

adequate nutrition. Rats whose growth had been sup-

pressed for over a year made prompt response and

quickly reached mature size when the inhibiting factors

were removed.

TABLE I

GROWTH OF ALBINO RATS DURING THEIR FIRST YEAR 5

| |

Males | | Females (Unmated)

Weight | | Weight

Age (Days) Age (Days) prae E

Ob- - b- Calcu-

served lated rved | lated

(Grams) | (Grams) (Grams) | (Grams)

First cycle First cycle

11 13 15 11 13 14

15 17 i7 15 18 16

21 Zt 22 21 23 21

31 32 32 29 31 30

` 40 42 44 40 44 45

49 57 60 49 58 61

61 82 85 61 78. 85

70 107 107 70 100 103

79 128 129 82 125 125

85 144 143 92 140 139

97 160 168 107 155 154

107 177 185 117 167 159

117 191 1 124 171 162

131 203 211 Second cycle

143 218 217 124 171 179

150 225 220 131 179 180

Second cycle 138 182 182

150 225 229 143 183 183

157 227 230 150 185 184

164 231 232 164 185 188

171 ra 296 234 178 192 192

178 | 239 236 192 196 196 +

185 ; 240 239 365 226 4

216 253 251

256 265 266

365 279 279

5 Data from tables 3 and 64 in: Donaldson, H. H. ‘The Rat.’’

delphia, 1915.

Phila-

544 THE AMERICAN NATURALIST [Vou. LV

A second feature of their results, which is no less note-

worthy than the first, is the rapidity of growth after

adequate diets were given. They show that the gains

made by the rats whose growth had been previously sup-

pressed were made in much less time than would be re-

quired for a rat on adequate diet to make the same gain

in weight. This inquiry is concerned only with certain

characteristics of the rate of growth following the initial

period of suppression and will not attempt any discussion

of the. nutritional aspects of the problem. The data dis-

cussed have been drawn from the work of Osborne and

Mendel. It has not been possible to obtain records of

enough individuals to give a statistically reliable aver-

age, yet the records employed are fortunately free from

extreme fluctuations and are satisfactory as far as in-

dividual records can go.’

The first case to be discussed is that of a male rat

(No. 1012) which at age 370 days had reached a body

weight of 127 grams, having been fed alternately ‘‘ gel-

atin food ”’ and ‘‘ milk food.’’* On the 368th day the ra-

tion was permanently changed to ‘‘ milk food plus mixed

food.’’ This change in diet was promptly followed by

rapid growth and the attainment of mature weight about

180 days later. It is evident from Osborne and Mendel’s

chart that the curve of ‘‘ resumed growth ’’ was steeper

than the normal curve of growth. This difference is

especially well marked during that portion of the time

in which there is an actual increase in body size and less

well marked during the time in which its increase is due

to the formation and deposition of fat.

A quantitative study of the resumed growth of this

animal shows the existence of two distinct cycles, each

of which is expressed by an equation of the type already

6 Osborne, T. B., and Mendel, L. B., Amer. Jour. Physiol., 40 : 16-20, 1916.

7 The writer is greatly indebted to Dr. Osborne and Professor Mendel not

only for their kindness in furnishing data, but for their criticism of this

manuser. ipt.

(oaas T. B., and Mendel, L. B., Jour. Biol. Chem., 18 : 95-107, 1914,

Chart

No. 641] THE RATE OF GROWTH 545

discussed. From the time at which an adequate diet was

supplied, until about the 432d day, the growth is ex-

pressed by the equation

This indicates a cycle of growth which was completed

at a body weight of 220 g. and which overlapped slightly

T o ——

2 (0)

200

360 420 480 540 600 |

; Fie. 3. Curve of growth E a male Dab rat following a suppression of 370

Gays.s. .isi. , observed weight; ——, leulated weight. The weight for days

2 was calculated preci log [e Ha O0—a#)] = .0193 (t—363). The weight

60—4

for days 433—600 was calculated from log [ (#—220)/(288—a)] = .0091 (#—475).

the second cycle of growth which was completed when a

body weight of 288 g. had been attained. The equation

representing the second cycle of growth is

‘log TE = = .0091 (t — 475).

The second cycle may be represented by a curve having

ordinate and abscissa axes originating at y—220 and

x — 432. The values are shown in Fig. 3. The agreement

between calculated and observed values is as good as

could be expected with only one animal in the sample.

The next case to be noted is that of a male rat

(No. 2161) which was stunted by a diet of inadequate

protein from age 38 days to age 248 days.® During this

9 Osborne, T. B., and Mendel, L. B., Jour. Biol..Chem., 23 : 439-454, 1915.

546 THE AMERICAN NATURALIST [Vou. LV

period the increase in body weight was from 53 to 73 g.

220 days after an adequate diet had been begun the rat

had grown to a weight of 300 g. The first cycle came to

an end at the 368th day, when a weight of 230 g. had been

reached, and is awe by the equation

log es = .0180 (t — 285).

The second cycle of growth is represented by the equa-

tion

— 225

log S = = .0150 (t — 408).

For a graphic representation of these values see Fig. 4.

324

200

(e)

100

50 ;

250 330 è 4410 4qo

Fie. 4. Curve of growth of a male white rat following a suppression of 248

ORG o observed weight: —-—-, calculated weight. ied weight for days

248-368 was calculated from log [#/(236—2)] = .0180 (t—- The weight fo

days 368-488 was calculated from log sp tip fete = .0150 (t—408).

The third case studied was that of a female white rat

(No. 2033) whose growth had been retarded by limiting

the daily quantity of food. Although the weight of this

animal was held below 60 g. for the remarkably long

period of 510 days, it had not lost the capacity for growth.

10 Osborne, T. B., and Mendel, L. B., Jour. Biol. Chem., 23 : 439-454, 1915.

Also Amer. Jour. Physiol., 40: 16-20, 1916

No. 641] THE RATE OF GROWTH 547

After 135 days on an adequate diet it weighed 222

The equations for the first and second cycles of growth

of this animal are

log Wes = .0260 (t — 519)

and

log ied = .0231 (t — 610).

For a graphic representation of these values see Fig. 5.

250

200)

ice

650

500 :

Fic. 5. Curve of growth of a female white rat following a suppression of 500

days. ...: , observed weight; ——, calculated weight. Er a for days

500-580 was calculated from log [#/(175—#)] = .0260 (t—519). The weight

for days 580—660 was calculated from log [ (#—170) esoe) = .0231 (#—610),

The constants of the different equations are grouped

for convenience in Table II. From them it appears

that male rats fed on adequate diet reached a weight

of 280 g. at the end of their first year, and the female rats

a weight of 230 g. in the same time. The females grow

relatively faster in the first cycle than the males and in

harmony with this property the values of K (the constant

of the reaction) in the first cycles were greater than

S THE AMERICAN NATURALIST [Vor. LV

those of the males. The values of K in the second cycles

were less than those in the first cycles. In the case of.

males on adequate diet the value of K in the second cycle

is about 66 per cent. of its value in the first, but in the

females the value in the second cycle is only 33 per cent.

of that in the first cycle.

The values of the constants are somewhat different in

the cases of rats subjected to an initial period of suppres-

sion. Rat No. 1012 reached approximately the same body

weight as a male rat fed continuously on adequate diet.

Its weight was 127 g. when full feeding began on the

368th day; the gain was, therefore, 153 g. in 212 days.

The male rats on adequate diet weighed 127 g. at the 77th

day; their gain of 153 g. was, therefore, made in 286 days.

The animal recovering from initial suppression, there-

fore, required considerably less time to make an increase

from 127 g. to 280 g., than adequately fed animals re-

quired to reach the same stage of development.

The case of No. 2033 (female) has special interest be-

cause of the remarkably long period of stunting. The

period of suppression started on the 39th day when the

rat weighed 53 g. and ended on the 510th day when it

weighed 57 g. The animal attained a weight of 222 g. in

135 days, following the resumption of full feeding. A fe-

male rat on adequate diet from the time of weaning would

require 295 days to increase from 57 g. to 222 g., or more

than twice the time to make the same gain. Other records

of the time required by stunted and by adequately fed an-

imals to attain a given weight are given by Osborne and

Mendel“ and in all cases the time was greatly reduced

in the case of animals recovering from stunting. These

authors likewise pointed out the broad biological signifi-

cance of this faster growth rate. This question of rate

is one of extreme interest in connection with the dynam-

ical aspects of growth

11 Osborne, T. B., and Mendel, L. B., Amer. Jour. Physiol., 40: 16-20,

1916.

No. 641] THE RATE OF GROWTH 549

TABLE II

COMPARISON OF CONSTANTS

3 A

First Cycle | Second Cycle

Duration of

EED both Cycles

nes K a | K

On adequate diet. Male........... 228 | .0187| 280 | .0123| 365 days

On adequate diet. Female......... 175 | .0234| 230 | .0076| 365 days

On restricted diets: |

M S MMO aa 220 0193 | 288 | -0091 263 days

INGO SIOL MME n we. 236 | .0180 | 300 |» 150 286 days

No 2033. Fomio e050 24 aa iid | -0260 | 230 | .0231 162 days

Í i asa

C. Rates or GROWTH as COMPUTED FROM VALUES OF da:/dt

It is unnecessary to dwell upon the prime importance

of the study of rates in physiological investigations. We

are concerned not only with what the organism is, but

how it came to be what it is. As soon as we begin to

study the problem of development, we encounter the ques-

tion of rates. No better means of studying the rate of

change in a system has yet been found than the differ-

ential calculus.

The differential equation representing the rate of auto-

catalysis is

= = ks (a — z).

a, « and t represent the same values as before, but

= K/a. We may proceed, therefore, to examine the

derivatives. of the equations used above to express the

sizes of the animals at various time intervals. The values

obtained for the growth of males and females are shown

in Figs. 6 and 7 in comparison with the observed weekly

increases of the animals which have been studied. The

computed values were obtained from the tables published

by Robertson” which give the values of (1/Ka).(dx/dt)

for corresponding values of K(t—t,).

The rate of growth of male rats on adequate diet was

computed for each cycle from the figures in Table I. The

12 Robertson, T. B., Univ. Calif. Publ. Physiol., 4: 211-228, 1915.

550 THE AMERICAN NATURALIST [Vou. LV

rates for each cycle are shown in Fig. 6. The growth rate

at birth had an appreciable value, as one might expect

2.5004 a

Aor

2.000+ +

= q

i" +

" i)

i il

. 4 >

| 500+ 7

4 `

dx j :

dt y `

' ;

j ;

1.000}

+ a

4 r

fi oO ©

500+ /'

H ro)

A \ Ann

Pian) fo} a

- v S

ae N © ze

a A : oe

ae: N NNS

owe alse > ~ È

0 ——_ a b25 -

100 200 300 400

Days o

Fic. 6. airia chee of male white rats. ..... rved increments of

rats on adequate calculated values ies Ni rats on adequate

diets ; ech values’ p dæ/åt for rat aria from a suppression of

370 days

From birth to the 73d diy the rate rose rapidly to its

maximum, then fell to zero value of dx/dt at about the

215th day.

No. 641] THE RATE OF GROWTH 551

Before the first growth cycle was ended the second cycle

had begun. The computed values show that this second

cycle began about the 80th day of postnatal life, reached

2.500+

100 200 300 400

Days

Fic. Growth rates of female white rats..... rved incre: sia

of hats on pico diet ; ——, calculated ana of dw/dt ee ale on adequat

t ; ——— calculated values of da/dt for rat recovering from a suppression of

500 days

a maximum about the 213th day and ceased shortly before

the 400th day. Since the value of kx(a—x) approaches

zero when # is very small and again when « is nearly

552 THE AMERICAN NATURALIST [Von. LV

as large as a, it is possible to trace the curve in both

directions until it reaches zero values of dx/dt. The ac-

tual values of da/dt in the second cycle are much smaller

than those in the first cycle.

Attention must be drawn to the way in which these two

curves overlap. It will be noted that the second cycle

of growth began shortly after the rate of the first cycle

began to decline and that the first cycle continued to

about the time of maximum rate in the second cycle. The

true curve of the growth rate of these animals is, there-

fore, the arithmetical sum of the values of dx/dt for the

various values of ¢. This is shown in Fig. 6 and agrees

very well with the observed weekly increments.

The curves in Fig. 6 were obtained by plotting da/dt

as ordinate and ¢ as abscissa.

If we let dx/dt—z, we may write

gz — kz(a— s),

which when differentiated becomes

a = ak — 2ka,

If the right-hand member of this equation be equated to

zero it will give the values of x for which ¢ is either max-

imum or minimum.

Let ak — 2ka = 0; then

2ka = ak,

a= a/2,

Therefore the rate, z, is either maximum or minimum

when x—a/2. To find whether z is maximum or min-

imum it is only necessary to get the second differential

of the above equation

de = — 2k.

Since this quantity has a negative sign 2 (—da/dt) is a -

maximum when v= 4/2. In other words the rate of

growth is a maximum when the cycle has reached a stage

at which the weight of the animal is half the weight it-

attains at the end of that cycle. When x<(a/2) the curve

rises and when g> (a/2) the curve falls.

No. 641] THE RATE OF GROWTH 558

The rate of growth of an animal recovering from an

initial period of suppression may be studied in compar-

ison with that of a rat on adequate diet. Take the case

of rat No. 1012 (male). The computed values of dæ/dt

have been plotted in Fig. 6. The time at which dx/dt

was a maximum in the first cycle was made to coincide

with the maximum for the first cycle of the rats on ad-

equate diet. This arrangement was adopted to facilitate

comparison. |

The curve for the first cycle as plotted in Fig. 6 is a

very fair duplicate of the curve for the same cycle of

growth for the rats on adequate diets. There is more

difference in the case of the second cycle. The curve for

rat No. 1012 has a maximum which is lower and occurs

somewhat nearer the dx/dt axis than that of the other

class of animals. In other words, the increased weight

due to formation of fat in this animal began relatively

earlier than in animals on adequate diets. In the main,

however, there are no striking differences between the

relative growth rates in the two cases, except that their

maxima are nearer together.

The same sort of computations have been made for the

growth of female rats and they are shown graphically in

Fig. 7. An extended discussion of them is unnecessary,

as it would be in many respects a mere repetition of what

has been said. A comparison of the curves of female rats

on adequate diet with those of rat No. 2033 shows (a)

that the rate of resumed growth was faster in the second

cycle, (b) that the second cycle was of shorter duration,

and (c) that the maximum of the second cycle lay closer

to that of the first cycle than in the case of rats on ad-

equate diet.

In view of the fact that rats recovering from initial

suppression reach mature weight more quickly than an-

imals fed on adequate diets, it is somewhat surprising to

find such a close similarity in the values of da/dt for the

same time intervals. One might expect that the curves

for the recovering animals should be higher and steeper.

554 THE AMERICAN NATURALIST [Vou. LV

I believe that the reason for the quicker growth in the

recovering animals lies, not in a faster growth rate in the

cycle, but in the shorter time between the maxima of the

two cycles. In other words, the final weight is reached

more quickly because the second cycle of growth com-

mences relatively earlier and is added to the first cycle.

Unfortunately, the weights of recovering animals were

not taken at sufficiently frequent intervals to afford data

upon the actual rates of growth in this class.

If we assume that growth is the result of a catalyst

acting upon a substrate, it seems that we have a key to

the explanation of what is observed. The catalyst of the

first cycle was produced in the pre-natal stages. Al-

though there was no appropriate substrate available in

the starved animals, the catalyst did not disappear.

When an appropriate substrate was given, this catalyst

acted upon it, producing a cycle of growth essentially

equivalent to that shown by animals fed on adequate

diets. This may mean that the catalyst persisted unim-

paired until it was destroyed in the course of the reaction.

The catalyst responsible for the second cycle likewise ap-

_ peared and induced the formation of fat. If the second

catalyst is in some way dependent upon a time factor for

its formation (or activation) it is plain that it should

show its activity relatively earlier in the case of animals

recovering from a long period of initial suppression, be-

cause of a quasi cumulative age effect. The effect of this

would be what we have seen to happen, viz., a crowding

of the cycles nearer together.

The writer is fully aware of the hazards encountered

in attempting to represent so complex a reaction as

growth by a simple formula. The phenomena of growth —

appear, however, to be coordinated into a single self-con-

sistent process, in which many chemical and physical

factors are combined. The possibility of expressing

growth by a simple formula showing that an increase in

mass is definitely related with a function of time ought to

lead to considerations of a fundamental natnre.

No. 641] THE RATE OF GROWTH 555

D. Summary

1. The growth of white rats during the first year shows.

two cycles, and each cycle follows the course of an auto-

catalytic reaction. The first cycle covers the period in

which the skeleton rapidly increases in size; the second

covers the period in which there is a production and depo-

sition of fat.

2. The equations for the growth of the two sexes differ

in their constants, but each expresses the course of an

autocatalytic reaction.

3. The earlier period of stunting did not prevent the

animals from attaining the full weight characteristic of

their sex after having an adequate diet. An equivalent

gain in weight was made more quickly in the animals re-

covering from suppression than in animals on adequate

diet.

In other words the animals grew somewhat more rap-

idly during their period of recovery.

4. The growth of white rats recovering from a long

period of suppression follows the curve of autocatalysis,

though a portion of the first cycle has been run during

_ the long period of suppression.

5. The differential equations expressing the growth

rates show that the two cycles overlap to some extent.

The sums of the overlapping values approximate closely

the observed increments. The second cycle of growth of

rats recovering from starvation began and reached its

maximum relatively earlier than in the case of rats on

adequate diet.

6. There is evidence for the idea that each cycle of

growth in this case had its specific catalyst and that the

potential activity of the catalyst was not impaired by

long periods of inadequate nutrition.

556 THE AMERICAN NATURALIST [Von. LV

SHORTER ARTICLES AND DISCUSSION

ESTIMATING THE NUMBER OF GENETIC FACTORS

CONCERNED IN BLENDING INHERITANCE

In Science for July 29, 1921, Dr. W. E. Castle develops a

“‘method of estimating the number of genetic factors concerned

in cases of blending inheritance,’’ which appears so simple,

so attractively cogent, and so usable, that it is to be feared that

the erroneous assumptions upon which it is based may be given

less consideration than they merit.

out seven years ago,’ I took some pains to point out certain

fallacies which had crept into genetical literature through the

tacit (or express) assumption that the several factors affecting

the size of an organism or its parts, or the intensity of develop-

ment of any character, are similar to each other in kind and

equal in effectiveness. It has seemed to me that since that time

there has been marked improvement in the literature dealing

with this particular phase of genetical phenomena—whether in

response to my paper or through independent following out of

the simple logic of the case, it matters not. It comes as a dis-

tinct shock, therefore, to see the sudden reversion in Dr. Castle’s

paper to a supposedly outgrown and abandoned conception.

Dr. Castle has probably forgotten the bearing of my paper,

though he commended it very highly in a letter written at the

time of its publication; for so far as I can now recall he has

never referred. to it in any of his frequent papers, published

since that time, on subjects involving the multiple factor hypoth-

esis as an explanation of blended inheritance and the modifi-

cation of a so-called ‘‘unit-character.’’ This omission has been

the more interesting because my paper even gave the precise

interpretation of his hooded-rat case, which has been very lately

espoused by him.”

Referring to the particular question now under discussion,

my paper said:

Attempts to determine how many plural determiners for any

quantitative character are involved in a particular cross are as yet

premature. Such attempts are based on the unproven hypothesis

1‘ Duplicate Genes for Capsule-form in Bursa bursa-pastoris,’’ Zeitschr.

f. indukt, Abstamm. u. Vererb., 12: 97-149. 1914,

2 Castle, W. E., ‘* Piebald Rats and the Theory of Genes,’’ Proc, Nation.

Acad, Sci. [U. 8S. Amer.], 5: 126-130, 1 fig., April, 1919,

No. 641] SHORTER ARTICLES AND DISCUSSION 557

that the range of variability in F, equals the combined ranges of P,

and F, generations, and the unwarranted assumption that the dif-

ferent plural determiners are essentially equal in effect.

What was premature seven years ago might be conceivably

no longer premature, of course, but since the basis upon which

Dr. Castle now proposes to estimate the number of such plural

factors in specific crosses involves the same fallacies which made

previous attempts untenable, it is needful to reiterate that noth-

ing in the evidence justifies the belief that this new plan will give

any sort of approximation to the actual facts. Nevertheless, the

method proposed by Castle may be expected to have a certain

amount of interest as representing a limiting case.

It is absurd to suppose that the height of a man will be as

much affected, severally, by the factors which increase the thick-

ness of the scalp, as by those which affect the length of the long

bones of the legs; or that factors which produce changes in the

number of internodes of a plant will generally add severally the

same increment to the stature of the plant as will factors which

increase the length of some or all of the internodes. Castle rec-

ognizes this weakness, but seeks to minimize its importance and

declares that ‘‘no other assumption will permit of a general

treatment of blending inheritance.’’ He means, of course, merely

that on no other basis can a generalized mathematical scheme be

_ developed such as that which he has here presented.

But even if we allow such assumption of equivalence of ge-

netic factors to pass on the ground of mathematical expediency,

there are several other conditions involved in Castle’s scheme

which are equally unwarranted and which will profoundly affect

the validity of the conclusions arrived at.

For example, the two strains mated together are supposed to

stand at the two extremes of the total potential genetic vari-

ability in their progeny and all of the determiners are assumed

to be additive in their effect, so that if we let the factors be repre-

sented in the usual manner by letters, ee lesser parent must have

the formula X Xaabbecddeeff .. hile the larger parent oc-

cupies the other extreme YYAA BBCCDDE EFE.. Cert neea

scarcely be pointed out that while such a situation might be

realized in some specific case it could not be generally true, and

the greater the number of factors involved in any specific cross,

the less likely would it be that the larger parent would possess

them all and the lesser parent none. In my paper, referred to

above, I said in regard to this point (p. 132) :

558 THE AMERICAN NATURALIST [Vor. LV

Nilsson—Ehle (1911) has described a case in which the range of

variation in the length of heads of wheat in the F, considerably ex-

ceeded the combined ranges of the two parents. Hayes (1912) has

found a similar case in the number of leaves in tobacco, and Emerson

and East (1913) have seen the same phenomenon in the length of

internode and total length of stalks in maize. It seems probable that

such transgressive variation may be the rule rather than the excep-

tion when very complex characters are investigated; for it is hardly

to be expected that a large number of plural determiners affecting

such a character shall all act in the same direction or that the parent

having the highest development of the given character’ shall generally

contain all the genes which the other chosen parent possesses. When-

ever such transgressive variability is producible by the genotypic 1 re-

combinations of parental characters, the frequency with which F,

dividuals simulate either parent gives no clue to the total CHEN a

plural determiners which have been brought together, with respect

to any character under consideration.

We might even derive a mathematical expression for the prob-

ability that the parents would stand at the extremes of total

genetic variability, by assuming that the two parental types are

taken at random. This would be perhaps a fair assumption, since

the number of fàctors can not be determined by inspection.

There would then be, if we let n represent the number of factors

involved in the cross, n!/2 ways in which the event in question

can not happen and only one way in which it can happen ; hence

the probability that all the factors would be present in the larger '

parent and absent in the lesser parent would be as 1 to n!/2.

In the specific case of East and Emerson’s corn, cited by Castle

as having about 15 factorial differences, (n= 15), there would

be, therefore, 633,477,184,000 chances to one against all of these

size-modifying factors being present in the larger parent and

absent in the smaller parent. When we let n==50 or 150, to

agree with the numbers indicated for the rabbit crosses, the

chances become practically infinitesimal and we must fairly

conclude that it has never happened, and never will happen,

that a cross involving so many independent size differences has

been made, or will be made, between individuals standing at the

opposite extremes of the total potential genetic variability.

The remark made in my paper on duplicate genes, regarding

the inadequacy of the frequency with which parental types are

duplicated in the F, as an indication of the number of factors

involved, applies equally well to the validity of conclusions

awn from changes in the * standard deviations. The

No. 641] SHORTER ARTICLES AND DISCUSSION 559

correctness of this conclusion will be made sufficiently obvious

by consideration of a simple illustrative case: Let us assume

that there are involved in a given pair-mating the six duplicate

size-factors, AABBCCDDEEFF, and their absences or corre-

sponding recessives, aabbccddeeff. The relation of the F, 'and F,

variability will be exactly the same no matter which of the four

following types of mating is made: (a) aabbccddeeff X

AABBCCDDEEFF (difference between parents, 12 units) ; (b)

AAbbccddeeff X aaBBCCDDEEFF (parental difference, 8

units); (c) AABBccddeeff X aabbCCDDEEFF (parental dif-

ference, 4 units) or (d) AABBCCddeeff X aabbccDDEEFF

(parental difference, zero). Neither Castle’s original scheme

nor Wright’s suggested modification® of it can be true, at one

and the same time, for more than one of these types of mating

and they have been specifically designed only for the first-men-

tioned type (a) ; but as we have seen in a previous paragraph, if

the number of factors involved is large, the number of matings

of type (a) is almost infinitesimal in comparison with the num-

ber of matings of the other types, (b) + (c) + (d) + (e)

+... ton terms.

It is not necessary, however, to suppose that all factors which

affect a blending character are additive. Indeed, it is quite cer-

tain that they are not, and that some factors act in a negative

direction and others in a positive direction. Inhibiting and

depressing factors have been fully demonstrated, as may be ex-

emplified by such classic cases as the inhibitors of horn produc-

tion in cattle and sheep; the inhibitor of indeterminate growth

in the tail of the Japanese long-tailed’ fowl; the condensation

factor characteristic of the compactum type of wheat; and a

series of dominant depressing factors which Davenport’s* studies

have made probable in the case of human statures. The fact that

Castle himself was able to make progress in the minus direction

in his selection experiments suggests the probable accumulation

of factors acting in a negative direction, though in this case the

evidence is not decisive, as the same effect would have been se-

cured by the gradual elimination of factors acting in a positive

direction. It seems to me that the combined action of plus-act-

3 Castle, W. E., ‘‘ An Improved Method of Estimating the Number of

Genetie Factors Concerned i ‘in Cases of Blending Inheritance,’’ Science, 44:

223, Sept. 9, 1

4 Davenport, "i B., ‘‘ Inheritance of Stature,’’ Genetics, 2: 313-389.

July, 1917.

560 THE AMERICAN NATURALIST [Vou. LV

ing and minus-acting growth factors gives a very true, though

somewhat formal conception of the general situation in all organ-

ized beings ;—the interplay of growth-promoting and growth-

inhibiting factors may be thought of, in a figurative sense, as

forming, within the limits. of fluctuating variation, a sort of

elastic ‘‘mold’’ into which any organism, whether plant or

animal, develops, and which gives it its wonderful specificity of

form and size.

But on the basis of this conception of factors, acting, some

in a positive and some in a negative direction, the combined

action of the negative group exactly balancing the combined ac-

tion of the positive group, and jointly determining the mean

size or the average condition with respect to any blending char-

acter which may be under consideration, it becomes unnecessary

to assume the absence of dominance. I have been teaching my

students for the past six years that the postulation of lack of

dominance which has always been made the basis of the multiple-

factor interpretation of inheritance of size or of other blending

characters is wholly unnecessary and that those who have dis-

cussed this type of inheritance have been led to place an alto-

gether unnatural and unwarranted stress on the occasional oc-

- currence of incomplete donrinance in other cases.

But whether any or all of the size-factors are dominant or

not materially affects the amount of change which they effect

in the value of the F, standard deviation, and must correspond-

ingly change the estimate of the number of factors involved when

that estimate is based on the value of these F, standard devia-

tions. As a simple example, I may cite the hypothetical illus-

tration given in my 1914 paper (p. 129), referred to above:

Thus, if a plant possessing a partial inhibitor or reducer of inter-

node-number be crossed with another plant having a stimulator for

internode length, all the other genes being the same in the two cases,

the height of the F, plants would be intermediate between the heights

of the parents, with variability due alone to fluctuation, as it is in the

homozygous parents. The F, would show increased variability, and

this increase would uppear greater if the two differentiating genes

were dominant, than if dominance were absent.®

To illustrate this fact further, let us assume that the six

size-modifying factors which differentiate two mates are

AABBCC, acting in the plus-direction, and DDEEFF, acting

5 Not italicized in the original.

No. 641] SHORTER ARTICLES AND DISCUSSION 561

in the minus-direction, and that the effect of each factor pair

is represented by 2 units, thus making up the total of 12 units

arbitrarily chosen by Castle. If the various permutations are

worked out, the F, series of frequencies is found to be as follows:

TABLE I

DISTRIBUTION INTO SIZE-CLASSES priya FROM THE INTERPLAY OF SIx

flan ge 3 POSITIVE AND 3 NEGATIVE, WHEN DOMINANCE Is WANTING

AND WHEN Sena Is COMPLETE

Class Values and Frequencies

Condition

as to ) | | | |

Dominance —6 -|4 —3|—2|-1| 0 11121341516

| | | |

dominance............ 2 | 66 220 495 792 0st 66 12)

Dominance complete...... 27 270 981 [1.5 1,540 981 2 |

These frequencies are exhibited graphically in Fig. 1, the curve

produced by dominant factors being reduced to the same area

as the curve of no dominance by dividing each frequency by 2.

It is thus seen clearly that whether dominance is present or ab-

sent, the resulting curve is of the same general type.

7 +

Fic. 4, “The Mepa produced by six equivalent factors, three Tessere in plus

direction and t in minus erg The unbroken curve shows the result

when dominance nae wanting. Dotted curve shows the effect of errai dom-

inance in all six factors

The standard deviation of the curve with no dominance is

\/3 = 14.43 per cent., as Castle has stated, but when all of the

562 THE AMERICAN NATURALIST [Vou. LV

six factors are completely dominant the standard deviation is

v4.5, or 17.67 per cent., which is the percentage corresponding

to 4 factors as given in Castle’s Table II. In other words,

six factors with dominance produce as great an increase in the

F, variability as four factors would produce if none of them

exhibited dominance.

It can be shown in the same way, by means of a simple test

case, that Castle is entirely in error in saying that ‘‘if one factor

really has an influence greatly superior to that of other factors

in a case of blending inheritance, this will be seen in the pro-

duction of asymmetrical or multimodal variation polygons in F,

and F,.’’ The consequence he mentions would follow only in

ease the factor having the greater influence happened to be domi-

nant, for in the absence of dominance each factor enters into

combination with every other factor in al : 2 : 1 series of inten-

sities which gives the probable-error type of distributions. Thus,

if we assume that four factor pairs AABBCCDD differentiate

two chosen mates with respect to some blending character, and

that three of these factor pairs, AABBCC, each acts in a plus

direction with a value of 2 units, and the fourth factor pair,

DD, acts in a minus direction with a value of 6 units, the F, will

be intermediate as before while F, will give the series of fre-

quencies shown in Table IT.

TABLE II

DISTRIBUTION INTO SIZE-CLASSES RESULTING FROM THE INTERACTION OF

ACTOR PAIRS ACTING IN THE PLUS DIRECTION, EACH ADDING

Two UNITS, AND A FOURTH Factor PAIR ACTING IN THE

MINUS DIRECTION AND SUBTRACTING 6 UNITS

Class Values and Frequencies

Condition as to

Dominance | OH

6 —5|—4|—3|-2)-1 0} 1| 2) 3/ 4/5) 6

No dominance............. 1 | 6 |15|22|27/36|42|36|27/22|15) 6| 1

Dominance complete........ | 3 | 27 81 82 9 27 27

It will be noted that while the curve of no dominance in this

ease is not quite a typical probable-error curve, it is nevertheless

perfectly symmetrical and the deviations from the typical curve

are such that they would never be detected when the several

classes are modified by concurrent fluctuations from environmen-

tal causes.

No. 641] SHORTER ARTICLES AND DISCUSSION 563

minus direction, with an ect

shows no dominance in any factor; Aatted

eurve shows effect of complete oiia a i four factors

The standard deviation of this series of values is V6, or the

same as would be produced by 3 factors if all were of equal value,

as shown in Castle’s table. This curve is shown graphically in

Fig. 2 together with the corresponding curve for factors weighted

in the same manner, but exhibiting dominance. Such an irreg-

ular, multimodal curve as the latter would be easily detected if

there were no fluctuating variations, but Castle apparently over-

looks the effectiveness of fluctuating variations in obscuring

the details of the underlying curve of genotypic variability.

When he interpolates between the standard sizes of two breeds

of rabbits 50 to 150 centers of genetic stability, he should take

into account that the fluctuations about each of these centers will

be sufficient to scatter the individuals which belong in any one

class over quite a considerable number of other classes, thus

filling up the gaps and hiding a great deal of putative multi-

modality in the genotypic curve, which would result from ine-

qualities in the relative effectiveness of the several plural factors

involved in any case of blending inheritance. Castle apparently

thinks that this confusing concurrence of genetic and fluctuating

variation could be dissolved to a certain extent by rearing ‘‘ade-

quate numbers’’ in F, and F,, but it must not be forgotten that

the masking effect of fluctuations advances pari passu with the

increase in the size of the population. Only the irregularities due

564 THE AMERICAN NATURALIST [Vou. LV

to too small size of the random sample could be eliminated by the

rearing of larger numbers; the effects of the interplay of numer-

ous environmental factors could not be thus eliminated.

rom the foregoing considerations it must be clear that. Castle

is altogether too sanguine as to the value of his method when

he says:

It is perhaps not to be expected that results more than approxi-

mately correct would be given by this method, unless rey large

numbers of both F, and F, individuals have been studie

I believe the cosain is justified that even a ‘dels large’’

number of individuals of the F, and F, could not be expected

to give correct estimates of the number of factors interacting

in any case. As between the method of Castle in estimating the

number of hypothetical duplicate factors operating in any case

on the basis of the change they produce in F, standard devia-

tions, and the method of Punnett, of formulating a genotypic

situation on the basis of a small number of definitely weighted

factors, I am convinced that the latter method is much to be

peers even though it does not lend itself readily to a

— treatment of blending inheritance.’’

GEORGE H. SHULL

PRINCETON UNIVERSITY

Dr. SHULL has kindly sent me his manuscript in advance of

its publication and generously asks me to comment on it, an

invitation which I gladly accept.

The difficulties which he thinks might be encountered in apply-

ing the method, which I have suggested, for estimating the num-

ber of genetic factors involved in cases of blending inheritance,

seem to be essentially these.

1. The possible unequal influence of the several factors which

are responsible for a case of blending inheritance, making it

difficult to estimate their number from the total effect observed.

2. The possibility that some of the factors may be dominant

in character and others not.

. The possibility that some may be positive in action and

others negative or inhibitive.

I think that these difficulties arise chiefly from the attempt of

Dr. Shull to extend the application of the method beyond the

field for which it was proposed. Cases of dominance or of trans-

6 Italics are mine,

No. 641] SHORTER ARTICLES AND CORRESPONDENCE 565

gressive variation in F, would not come within the scope of

blending inheritance as I defined it, when proposing the method,

but only cases in which both F, and F, are intermediate between

the parent races. But Dr. Shull maintains that some dominant

factors may be involved even where dominant characters are not

in evidence. This may be admitted as a possibility even though

we have no evidence for it. What of it? Dr. Shull formulates

a test case, the strongest one imaginable, in which all (six)

factors affecting a size character are ‘‘completely dominant’’

and finds that in this case the method which I suggested would

indicate a smaller number of factors than the true one, or four

instead of six. If this is the maximum error to be anticipated

when all factors are completely dominant, and it is really doubt-

ful whether any factors are dominant in the case of blending

characters, the possibility need not give us great concern.

Further, if all factors involved in producing a character are

‘“completely dominant,’’ how can the character itself keep from

being dominant? And if it is dominant, the case will be auto-

matically removed from the field of blending inheritance. Shull

seeks to avoid the difficulty in his hypothetical case by putting

three dominant factors in one parent race and three in the other,

but this arrangement, by recombination of factors, in F, would.

result in segregation of the parental types or in transgressive:

variation, either of which events would remove the case from

the category of blending inheritance as I have defined it.

The supposed difficulty, that some factors involved in the pro-

duction of blending characters may be positive in action while

others are negative, is purely formal. With three positive and

three negative factors, in his hypothetical case, Shull comes out

with identically the same distribution in thirteen size classes.

that I calculated for the same number of factors all positive,

both of us assuming no dominance to occur.

It remains to deal with the first-mentioned difficulty, the

possible unequal influence of the several factors assumed to

oceur in blending inheritance. I had anticipated this difficulty

in my first paper, but had assumed that, ‘‘if one factor really

has an influence greatly superior to that of other factors in a case

of blending inheritance, this will be seen in the production of

asymmetrical or multi-modal variation polygons in F, and F,.’”*

Shull challenges this statement and artfully constructs a case to

1 I should have limited the statement to F,.

566 THE AMERICAN NATURALIST [Vou. LV

disprove it in which the influence of one factor exactly equals

and negatives that of three other facters. In this case he finds

that the variation curve is symmetrical when no dominance

occurs, but asymmetrical and bimodal otherwise. Had the hypo- |

thetical factors been less carefully weighted by Shull, he would

not so easily produce a symmetrical F, curve. Consider an

actual case in which a single factor of superior influence occurs

and yet in which there is no dominance, that of the blue Andalu-

sian fowl. Black mated with splashed white produces blue in F,,

an apparent blending. Yet F, falls, as every one admits, into

three distinct classes notwithstanding the occurrence of one or

more modifying or inhibiting factors affecting the result in a

minor degree (Lippincott). ‘‘Fluctuating variation’’ does not

here obscure segregation, as Shull assumes would be true in

hypothetical cases of blending inheritance in which factors of

very unequal influence occur, even when large numbers are

studied. Now I should not class the case of the Andalusian fowl

as blending inheritance, but I think it may serve to show that

Shull’s objection is not well founded, in accordance with which

he assumes that a factor of major influence will not be readily

detected, even when it is operating in conjunction with minor

modifying factors.

W. E. CASTLE

To THe Eprror or THe AmeRICAN Naturauist: The forego-

ing article and rejoinder are submitted for publication in their

original form. I do not think that Castle’s ‘‘ comment ’”’ ade-

quately meets the several difficulties which I have pointed out,

and he presents no considerations which seem to me to warrant

a modification of the statements I have made. Others who are

interested in the topic under discussion may be depended upon to

recognize the validity or non-validity of any of the propositions

made by either of us. It should be said, however, that there was

nothing ‘‘artful’’ in my choice of an illustrative case to show

that inequalities in the effectiveness of the several genes do not

necessarily produce asymmetrical and multi-modal variation-

polygons when dominance is not present. The case could have

been made still more striking by using a larger number of factors,

but the additional labor required did not seem necessary. Ob-

viously Castle has not tried out cases in which the factors are

weighted differently from the weightings I assigned to them,

No. 641] SHORTER ARTICLES AND CORRESPONDENCE 567

or he would not say that in that case I ‘‘would not so easily

produce a symmetrical F, curve.’’

That he has not assimilated the significance of the inter- play

of pias acting and minus-acting factors is shown by his ques-

tion

It all factors involved in producing a character are “ completely

dominant,” how can the character itself keep from being completely

dominant ?

The curves in my Fig. 1 are an adequate answer to this question.

Finally, it seems hardly necessary to point out the inadequacy

of the case of the Blue Andalusian fowl, as a proof that inequal-

ities in genotypic variation are not masked by fluctations when

the amount of fluctuating variation is large in comparison with

the distances between the centers of genetic stability which are

determined by the numerous factor combinations putatively in-

volved in the several examples cited in Castle’s original paper.

E0. H. SHULL

“

GENETIC TERMINOLOGY

THE genetic terms recently proposed by G. H. Shull! seem

to supply a real need. Their general use would certainly tend

to reduce both the danger of ambiguity and the need for cum-

bersome descriptive phrases. Probably a few additional numer-

ical terms, such as dizygous and trizygous,* would also be useful.

Some existing equivalents have an obvious disadvantage with

respect to compounding; the compound trichromosomal, for ex-

ample, could not replace trizygous (‘‘dependent on three pairs.

of chromosomes’’). Monozygous and pleiozygous, although often

interchangeable with linked and unlinked, should be useful;

the latter, for example, to characterize genes that are located in

several pairs of chromosomes but are not all necessarily unlinked

with each other. These words are so interrelated among them-

selves, and so closely related to terms in general use, that all

their advantages can be realized with little effort.

Shull suggests that it is time ‘‘to abandon the use of ‘Mende-

lian’ and ‘non-Mendelian’ as definite categories, and to adopt

1 Shull, George H., ‘‘ Mendelian or Non-Mendelian? ’’ Science, N. S., 54:

213-216. Sept. 9, 19 5 i

2 For the Greek numeral prefixes see Blakeslee, Albert F., ‘‘Types of

Mutations and their Possible Significance in Evolution,’’ Am. NATURALIST,

55: 254-267. 1921.

568 THE AMERICAN NATURALIST [Vou. LV

other terms which will have greater precision of meaning.’’

Let us accept his timely proposal, which obviously applies —

especially to the more technical and precise terminology of genet-

ics. With his new terms available, we may safely relegate the

older ones, aside from historical references, to the more popular

language of science.

Shull uses the older words to illustrate the application of his

proposed terminology, but he does not specifically discuss their

future delimitation in case they still retain a certain usefulness.

Their future, I believe, deserves consideration. It seems certain

that they will remain familiar words because of their historical

value, in relation both to Mendel’s work and to its earlier ex-

tension. Doubtless they will long be especially useful in the

more popular presentation of genetic topics, to obviate burden-

some use of more precise but more formidable expressions.

Historically, it is plain that the meaning of Mendelian has very

largely kept pace with the widening conception of the funda-

mental applicability of Mendel’s theory, although often, as Shull

states, with the addition of qualifying expressions. When this

widening process reaches the farthest point of practical useful-

ness, it leads to a broad definition of Mendelism which, I believe,

deserves general acceptance.* It furnishes, for example, a con-

venient and familiar popular equivalent of zeuxis for the char-

acterization of ‘‘chromosomal heredity,’’ at least so far as the

inheritance phenomena of sexual reproduction are concerned.

other senses of Mendelian seem to require more technical

detail in definition, or to be otherwise less useful for the purpose

in question.

Again, the most significant conflicts of ‘“Mendelism” with its

critics have raged along a line of demarkation essentially corre-

sponding to the broader definition. ‘‘Mendelians’’ once en-

countered frequent denials of the completeness and the gener-

ality of segregation, and frequent assertions that new somatic

ratios implied other modes of inheritance of equal significance

with Mendel’s. The triumph of the chromosome theory has

been the definitive establishment of the fundamental significance

of ‘‘Mendelian heredity.’’

Finally, the broadest definition is fully justified logically, al-

though it may not be superior in this respect to some other

delimitations of the term. It may be held with good reason that

3 Not forgetting, of course, that the older usage varies.

No. 641] SHORTER ARTICLES AND CORRESPONDENCE 569

even linkage represents an addition to Mendel’s genetic theory,

rather than an exception to it. His scheme of independent sepa-

ration and recombination of potentialities at gametogenesis is

still adequate for the innumerable cases resembling his. Further,

all genetic factors belong to theory rather than to observed fact,

as do atoms and molecules. A gene is a supposed reality; it is

something which many geneticists now assume, on the basis of

evidence which they consider essentially conclusive, to be an

actual part of a chromosome. The idea of lethal genes, there-

fore, or even that of the gene as a part of a chromosome, just as

truly constitutes an addition to Mendel’s genetic theory as a

whole, as does the explanation of linkage ratios. The difference’

is, from this viewpoint, one of degree rather than of kind. If we

admit some added hypotheses as Mendelian, why should we neces-

sarily exclude any others which plainly relate to the same unified

nuclear mechanism ?

Even if-we hold, as we may, that Mendel’s theory has been

revised as well as extended, its most fundamental. feature, by

present standards, is left unchanged. What, from our present

viewpoint, is Mendel’s most fundamental genetie conception?

Is it not that of a genetic shuffling, a segregation and recom-

bination, of genetic units which maintain their individuality

throughout the processes of reproduction and of development ?*

Most usefully and even most commonly, it seems to me, Mendel-

ism signifies the general type or mode of inheritance whose most

fundamental principle of character distribution was discovered

by Mendel; and this is ‘‘zeuxis,’’ or chromosomal heredity, in

sexual reproduction. This delimitation of Mendelism seems to

me fully as good logically, better in accord with history, and

much more promising of future usefulness in the field where

the term is still needed, than any of the less inclusive senses in

which it has been employed.

Dr. Shull says in correspondence, ‘‘I can see no objection to

the general non-technical use of the words ‘Mendelian’ and

‘Mendelism’ in just the sense which you propose.” And I þe-

lieve that Morgan, East, Jones and Wright are far from being

alone when they positively favor the broader definition.

4 Bateson, W., ‘‘Mendel’s Principles of Heredity.’’ 1909. Cambridge

Univ. Press. (See p. 13.)

T. H., Sturtevant, A. H., Muller, H. J., and Bridges, C. B.,

‘t The Mechanism of Mendelian Heredity.’’ 1915. New York, Henry

Holt & Co. (See p. 1.)

570 THE AMERICAN NATURALIST [Vou. LV

Let us, then, take the course which is obviously. more useful,

and also honor the memory of the great pioneer of genetics, by

applying his name to his great idea in all its later ramifications.

But—wherever newer and more precise terms will better pro-

mote the science of genetics, let us be ready to use them. Shull’s

recent contribution to genetic terminology promises considerable

and lasting usefulness.

Howarp B. Frost

UNIVERSITY OF CALIFORNIA

In connection with Dr. Shull’s' interesting and important

proposals concerning genetic nomenclature, attention should

be called to a situation which neither these proposals nor the

terminology in general use recognize. It is customary to refer

to individuals carrying single X or Z chromosomes, as being

heterozygous for sex-linked genes. For some time this has

seemed to be ill-advised to the writer.

The situation prevailing in an XX or ZZ individual hetero-

zygous for a sex-linked gene clearly differs from that of an XY

(or XO), or a ZW individual in the vast majority of cases,

though Schmidt’s work on Lebistes reticulatus, to which Dr.

Castle? recently called the attention of American workers, pos-

sibly indicates that for XY individuals it does not necessarily

always differ. In the one case, usually there is a demonstrable

allelomorph, not infrequently competitive enough in its expres-

sion to produce more or less of an intermediacy between the

two homozygous forms. In the other, usually there is not. ’

The term heterozygous, as much as homozygous, indicates an

allelomorphie pair, yet in XY and ZW individuals, with the

one possible exception noted, a pair of sex-linked genes has not

been demonstrated, and is clearly impossible for XO individ-

uals. To all appearances the sex-linked genes in such indi-

viduals are without synaptic mates. They are therefore sim-

plex but not heterozygous. :

In order to recognize this situation, and in a measure at

seribe it without using presence and absence terminology, and

in harmony with the terms proposed by Dr. Shull, I should

like to suggest the noun hemizeuxis (a half yoking) and the

corresponding adjective hemizygous (half yoked). Should such

1 Shull, Geo. H., 1921, Science, N. S., 54: 213-216.

2 Castle, W. E., 1921, Science, N. S., 53: 339-342,

No. 641] SHORTER ARTICLES AND CORRESPONDENCE 571

a suggestion prove acceptable there would be the three adjec-

tive series: homozygous, heterozygous, and hemizygous, refer-

ring to the three possible conditions with respect to any single

gene, namely, ‘‘like mates,’’ ‘* differing mates,’’ and ‘‘no mate.’’

The term might also be used in cases where a non-deficient

chromosome is paired with a deficient one. The deficient indi-

vidual would be hemizygous for the genes at those loci of the

non-deficient chromosome which were involved in the deficiency

of its mate.

WILLIAM A. LIPPINCOTT

KANSAS AGRICULTURAL EXPERIMENT STATION

CROSS-OVER VALUES IN THE FRUIT-FLY, DROSO-

PHILA AMPELOPHILA, WHEN THE LINKED

FACTORS ENTER IN DIFFERENT WAYS

Tue factors for bar, round, red and white eye are sex-linked,

being located in the X-chromosome, as shown in inheritance.

For example, when a bar-eyed male is mated to a normal female

all the resulting male offspring (F1) are normal and all of the

females bar eyed, as bar is dominant; that is, factors that occur

in the X or first chromosome have a “‘ criss cross °’ mode of in-

heritance following the distribution of the X-chromosomes, there

being but one X-chromosome in the males and two in the females.

The relative positions of the linear series of factors in the X-

chromosome of Drosophila have been determined by Morgan and

Bridges (237 Carnegie Institution). Factors that lie near to-

gether in the chromosome are more likely to be transmitted in |

the same combinations to the gametes than those that lie far

apart; that is, the strength of linkage depends on the distance

apart of the factors. The failure to transmit the same combina-

tions of factors that enter from the parents to all the offspring

is due to a crossing-over of some of the factors. For example,

a red bar-eyed fly, mated to a white round-eyed fly, give in the

second generation (F2) white bar and red round-eyed flies, as

well as flies like the original parents; that is, there has occurred

a recombination of the factors due to crossing-over.

Are the cross-over values the same when the linked factors

enter in different ways? My experiments performed in the

University of Chicago Laboratories give data relative to this

question. Using the presence and absence hypothesis, let (B)

represent the factor for bar and (R) the factor for red, then let

572 THE AMERICAN NATURALIST [Vou. LV

(b) represent the absence of the factor for bar and (r) the

absence of the factor for producing red-eyed flies. The factorial

composition, then, of the red bar-eyed flies is (BR) and for

white round-eyed flies (br). As both the factors for bar- and

red-eyed flies are sex linked—that is, they occur in the X-chromo- .

some, there being but one in the male and two in the female, as

above mentioned. Then-the male red bar-eyed flies are repre-

sented by (BR) and the female white round-eyed flies by (br—

br).

Matings were made as follows:

(2 pairs) Red bar-eyed males (BR). White eyed females (br—br) F1

171 White round-eyed males (br). 184 Red bar females (BR—br). |

| F2 Males. F2 Females.

| Red | White | Red | White ‘

Bar. Bar. Red. | White. Bar. Bar. Red. White.

self ERSTE Be Me NORGE Ipod saat NRE at a ———

(27.1) | 50 37 $o0°5 BO 49 31 42 34

oe ae ee 17 30 31 38 25 27 40

“3 39 32 29 42 40 a 33 30

bee! 26 15 27 17 28 12 29 17

ee 53 | 26 55 41 51 27 35 29

rejo aaco n 23 . 26 43 17 27 30

| | 138 207

226 | 196 249 | 133 | 193 | 180

(BR) (Br) | (Rb) (b) |(BR br)| (Br br) (Rb br) (br br)

As white bar-eyed flies and red-eyed flies are the cross-over

classes, then the percentage of crossing-over is equal to 660,

divided by 1,522, and the quotient, multiplied by 100, giving

43.3 per cent.

I next extracted pure lines of white bar-eyed flies and red-

eyed flies, using the white bar-eyed males and red round-eyed

females in the matings. Their factorial composition being (Br)

for the males and (Rb—Rb) for the females.

F1 99 red-eyed males (Rb) and 105 red bar-eyed females (Rb—Br) F2

(from about thirty matings from F1).

Males. Females.

Red. White. Red Bar. ` Red.

Red White

Bar. Bar.

1127 1265 1456 | 1045 2658 2606

(BR) (Br) (Rb) | (br) | (BRorelse Rb) | (Rb or else Rb)

| (Br Rb) (br Rb)

No. 641] SHORTER ARTICLES AND CORRESPONDENCE 513

The cross-over classes are the red-bar and white-eyed male

flies. The percentage of crossing-over being 2,172, divided by

4,893, and the quotient, multiplied by 100, which gives 44.4 per

cent. The difference, then, in the cross-over values when the

linked factors entered in different ways was but 1.1 per cent.,

which does not seem to be a signicant difference.

J. D. Ives.

ON COUNTING CHROMOSOMES IN POLLEN-MOTHER

THE genetic study of hybrids between species with different

chromosome numbers and of certain mutants requires the count-

ing of many chromosome groups and raises the question of the

best technique for the purpose. The staining qualities of aceto-

carmine, which has long been used for preliminary work, espe-

cially by zoologists, are considerably improved by a trace of

ferric salt. (Bolles Lee, in his well-known manual, gives for-

mule for iron carmine; but this has no advantage for sections

over iron hematoxylin: )

Iron Aceto-carmine 1.—Ordinary aceto-carmine is prepared

by heating a 45 per cent. solution of glacial acetic acid to boiling

with excess of powdered carmine, cooling and filtering. The

young anthers are teased out with steel blades or needles in a

drop of this until it changés slightly toward bluish red. An

excess of iron spoils the preparation. Anther remains are re-

moved, and a large thin coverglass (22 by 50 mm.) applied,

using the minimum of liquid: The edges are sealed with vase-

line. The preparation, if there is no excess of iron, may improve

for a day or two.

Tron Aceto-carmine 2.—To a quantity of aceto-carmine a trace

of. a solution of ferric hydrate dissolved in 45 per cent. acetic

acid is added until the liquid becomes bluish red, but no visible

precipitate forms. An equal amount of ordinary aceto-carmine

is then added. The anthers are teased out with nickel instru-

ments. If the stain is too dark, more aceto-carmine is to be

supplied. It may be diluted with 45 per cent. acetic.

Tron Aceto-carmine 3.—Anthers at the right stage are put into

a mixture of 1 part of glacial acetic acid to 9 parts of absolute

alcohol, to which sufficient solution of ferric hydrate in 45 per

cent. acetic has been added to color the liquid brown (the amount

574 THE AMERICAN NATURALIST [Vor. LV

varies with different objects). After some days or weeks the

anthers are teased out in ordinary aceto-carmine, avoiding the

use of steel instruments.

The chromosomes are usually ‘ati „accurately counted in the

metaphase of the second division, in dicotyledons. When the

preparation is a day or two old, the cytoplasm has swollen; and

a slight tap on the thin coverglass above any particular cell will

usually free the cytoplasm from the cell wall, and another tap

flatten it out with its contained chromosomes.

Satisfactory results have been obtained by these methods dur-

ing the past year with Datura, Canna, Antirrhinum, Linaria,

Brassica, Dahlia, Secale, Asparagus, Matthiola, Phaseolus, Stizo-

lobium, Tradescantia, Hemerocallis, Iris, Gladiolus, Zea and

Portulaca. The methods failed with Cnothera and Rhodo-

dendron.

The second of the above methods will probably be of the

widest applicability. The preparations will keep for a week or

more, if an excess of stain and of iron are avoided. The method

is quicker for counting chromosomes than staining sections with

iron hematoxylin, and in favorable cases the results may be

more certain. Thus in good preparations of Datura over a

thousand pollen-mother cells are scattered singly on one slide,

many of them showing the metaphase of the second division, and

some having both plates in one plane, with the chromosomes well

spaced and stained a deep bluish red, while the cytoplasm is

unstained. It takes certainly a modicum of patience to acquire

skill with this, as with most microscopical methods.

JOHN BELLING.

STATION FOR EXPERIMENTAL EVOLUTION,

CoLD Spring HARBOR, N. Y.

INDEX

NAMES OF CONTRIBUTORS ARE PRINTED IN SMALL CAPITALS

ALTENBURG, A., Interference in Pri-

, American Follicu-

linas, 347.

Antherophagus ochraceus Mels in

the Nests of Bumble Bees, T. H.

FRISON,

Barley, a Species Cross in, W. S.

MALLOCH, 281.

Bassler, R. S., and F. Canu, Early

Tertiary Bryozoa, A. ROBERTSON,

69.

BATESON, W., Genetic Segregation, 5.

BELLING, J. , Counting Chromosomes,

573.

RLAKEEIME, A. F., Types of Muta-

tion, 254.

BRIDGES, C, Gametic es Ass

served Ratios in a

Bruges, C. T., Taxonomic Afinitie

and Food Habits i in By aa,

Pait L., Variation and He-

redity in Lupinus, 427.

Chiton, Homing Behavior in, W. J.

CROZIER 218;

C romosomes, Counting, J. BELL-

ING, 57

CLAUSEN, R, E., and T. H. Goop-

uclear Divi-

c Infusorian, R.

. Wu, 328.

irst Generation

Hybrids, 116.

CROZIER W. J., Homing Behavior

in Chiton, 27 6.

DETLEF J. A., Mutation in the

House “is e, 469.

Drosophila, Gametie and Observed

s, C. B. Br

Ratio IDGES, 51; Orange

y Color in, F E and M

Den eee ar-eye, Sexual Di-

morphism in Ny, 404;

his

Cross- ove Tiloe in, J. D. IvEs,

DurHAM, G. B., Inheritance of Belt-

ing Spotting in Cattle, 476.

Ears, Short, mgp in the Mouse,

C: J I 42

Egg- secretion, The “Duality of, 0,

Fasten, N., Parasitie Copepods,

449,

Fiddler Crabs, Intersexes in, M, J.

a BUN 80; Juvenile, Variation

T RGAN, 82.

Folliculinas, American, E. A, AN-

RE D:

FRISON, H, Ant gel coma

pe ER Mels” in the Nests of

e B e

Frost, B., atic Segregation,

461; Genetic EREA. 567.

Generic, Ts in Zoology, W. L.

MCATFE, 89.

par Segregation, BATESON, 5;

For Nomenclature, C.

poe gi 75: F tors in Blending

Inheritance, $ SHULL, 556,

566; i

E. CASTLE, 564; Termi-

vartita H. B. FROST, 567; W. Ai

LIPPINCOTT, 570.

GLASER, cha The Duality of Egg-

secreti , 36

Growth, Rate of, H. S. REED, 539.

GUYER, M. F Immune Sera, 97.

Hance, R., Chromosomenzah] von

Zea Mays L., Y. Kuwada, 268.

HANNA, G. D., Genital Organs of

Hermaphroditie Fur Seals,

HARTLEY, C. P., and H. 8. palea

Reproducing Power me i 184,

Hmen, R. B., Yolk Mass in Oviduet

373.

Unilateral Reactions

Generation, G. N.

Vigor of, D. F.

Hymenoptera, Taxonomic yeo

134,

Immune Sera, M. F. Guyer, 97.

Inheritance, of Blue in ‘Poultry,

W. A. LIPPINCOTT, 289; of Belt-

ing Spotting in Cattle R-

HAM, 476; of Caneer, bh. nee,

Interference in Primula Sinensis, A.

BURG, 78.

ae a Seeretions in Amphibians,

, 193,

Ives, D., Cross-over Values in

ews 571.

576

Jones, D. F. Re A, of First Genera- |

tion. Hy brids,

KJERSHOW-AGERSBORG, pA P hea Me-

ri ae na (Gould), 22

Kuw ea a a von

Zea Mass b R. Hance, 268.

, P. A., and S. WRIG

Aa = 'Guinea-pigs to goe

culosis,

Lite, “reat of, RAYMOND PEARL,

481.

Linkage in Mice, C. WACHTER, 412.

Lippincott, W. A., heritance eE

Poultry, 289; Genet

Terminology,

Tos ; Genetic Form and No-

menclature, 175

EB, L., Tihai of Cancer in

Mice

LYNCH, C

somal Mutation, 421.

THE AMERICAN NATURALIST

[Vor. LV

Plants, Body and Organ Size in, E,

W. SINNOTT, 385

Polyembryony in the Armadillo, C.

R. STOCKARD, 62.

RATHBUN, M. J., Intersexes in Fid-

i bs, 80.

dler Cra

| Reactions, Unilateral, of the Melano-

| of t

phores d in Fishes, 286

REED, T S., Rate of Growth, 539.

| REES, Neuromotor Appa-

ratus of Paramecium, 464.

NSHAW, R., pe and the Utili-

zation T Foo d, 73.

J., Short Ears an Auto-

McAteer, W. L., Generic Names in

Zoology, 89.

Maize Messy og, Power of, oe =

HARTLEY and H, S. GARRI

184.

MALLOCH af tag F, Species Cross

in Bar rley, 2 ;

. W. VON

S ara ar EP mae appa A,

PETRUNKEVITCH, 178, 477.

Morean, T. H., Variation in Juve-

i bs, 8

Neuromotor Apparatus of Parame- |

mie ak | WACHTER, C., Linkage in Mice, 412.

Fish, 165.

Peyra a W. L. MCATFE, 89;

C. C. LITTLE, 175.

Oseoop, W. H., Turkey for Experi- l

ment, 84.

PEARL, M f

Duration of Life, 481.

la nda ‘178, 477. —

| WIENER, L., Sucking

WITSCHL Gonads

ROBERTSON, A., Early Tertiary Bryo-

zoa, F. Canu and R. 8. Bassler, 69.

Genital

473.

Segregation, Genetic, w. cee SON,

Seals, Hermaphrodi = aN

SHU Genetic Factors in

Blending Inheritance, 556, 566.

Nnort, E. W., Bod and Organ

Size in Plants, 385.

Srockarp, C. R., Polyembryony in

the A

Sucking Fish, i WIENER, 165,

oer egy oy. Resistance of Guinea-

HT, 20.

Turkey tar Experiment, W. H. Os-

Goon, 84.

UHLEN , E., Internal Seeretions

in ppe dk 193.

Variation and mene in Lupinus,

L. BURLINGAME, 427.

culosis,

Yeast and the Utilization of Food,

R. R. ReNsHaw, 73.

Yolk Mass in the Oviduct of a Hen,

R. B. Hien, 373.

ZELENY, C.

a oo in

Bar-eye Drosoph